# Welcome to VOICEB.ai DOCS

Create a new AI voice channel for your business — one that assists before your team gets involved.

🧠 This documentation helps you design, deploy, and operate **Autonomous Sales Voice Agents (ASVA)** that handle inbound demand end-to-end: qualifying, assisting, routing, and closing — before humans get involved.

{% hint style="info" %}
**VoiceB tip:** VoiceB is not a voice interface. It's the **revenue execution layer for inbound sales** — the intelligence layer between ad spend and closed revenue — built on goal-driven prompting, orchestration, and real-time execution, all configured through **Agent Studio**.
{% endhint %}

#### Overview

**VoiceB.ai is an agent-first AI voice platform built for revenue outcomes, not conversations.**

Inbound sales is structurally broken:

* Forms are abandoned (60–70%)
* Call centers are expensive and inconsistent
* Chatbots don't convert
* Good demand reaches humans too late — or never

**ASVA fixes this.** VoiceB deploys Autonomous Sales Voice Agents that:

* Detect intent in real time
* Discard invalid or non-sales demand automatically
* Qualify and route valid leads with precision
* Close simple sales directly via voice
* Escalate to humans only when it adds value

All while preserving caller ID, meeting GDPR / EU AI Act requirements, and generating full audit trails.

VoiceB is built for **enterprise inbound environments** where scale, compliance, and conversion matter: telecom, energy, insurance, banking, automotive, and high-volume B2C/B2B.

***

#### The End-to-End Vision — From Ad Click to Closed Sale

VoiceB owns the entire lead journey. A lead arrives from a Meta or Google ad, expresses intent through a call, the website, or click-to-call, and VoiceB's agent qualifies and closes it autonomously — writing the confirmed transaction to your CRM and sending the sale-level conversion signal back to the ad platform. All within seconds, without human involvement.

The result is a closed loop: your campaigns optimize on **real revenue**, not form fills or clicks.

```
   Ad platforms          Contact point            VoiceB Agent             Revenue systems
   (Meta, Google)  ──▶   (call · WebRTC ·   ──▶   qualify · handle   ──▶   (CRM · ERP ·
                          click-to-call)           objections · close       billing)
        ▲                                                                        │
        └──────────────────  Conversion signal · confirmed sale  ───────────────┘
```

VoiceB does not manage your budgets or keywords. It supplies the sale-level conversion signal that makes campaign optimization accurate — so ad spend becomes a function of budget, not contact-center capacity.

***

#### What the Agent Does — Autonomy Levels (L1–L4)

Every VoiceB agent operates across four levels of autonomy. **All plans include all four** — clients expand autonomy within the same platform as trust and results grow, with no re-platforming.

* **L1 — Qualification:** filter invalid calls, detect intent, capture structured, validated data.
* **L2 — Intelligent Routing:** route sales-ready demand to the right human agent, with full conversation context.
* **L3 — Assisted Autonomous Sales:** the agent handles most of the conversation; humans step in only for edge cases.
* **L4 — Full Autonomous Sale:** end-to-end close with no human — data collection, identity validation, confirmation, CRM write-back, and post-call email / WhatsApp triggers.

> Read the full framework → Autonomy Levels (L1–L4)

***

#### Agent Studio — Build Agents by Talking to Them

[**Agent Studio**](https://docs.voiceb.ai/agent-studio) is VoiceB's conversational build interface. You don't write prompts or upload scripts — you describe your goal, and Agent Studio configures the agent's **Goal, Script, Objections, and Guardrails** for you.

* Build and refine agents in natural language — no prompt-writing
* Scrape a URL to generate the knowledge base automatically
* Query your calls and outcomes conversationally, and run analytics
* Review and edit every block before going live

<figure><img src="/files/DEVazozgrsUm5XZpkLCF" alt=""><figcaption></figcaption></figure>

> [Build your first agent → Getting Started in 20 Minutes](https://docs.voiceb.ai/product-guides/getting-started-with-voiceb.ai-in-20-minutes)

***

#### Video — Watch a platform walkthrough

{% embed url="<https://www.loom.com/share/42f9f906163b44f3bc1d320762887c8a>" %}

***

#### Use VoiceB to

* Qualify inbound demand before humans invest time
* Discard non-sales traffic automatically
* Route valid leads instantly to the right team
* Close simple sales autonomously with AI
* Assist users in complex journeys 24/7 (forms, applications, purchases)

***

#### What Makes VoiceB Different

🔑 **Agent-First Architecture**\
We don't build chatbots. We build autonomous agents designed to execute sales outcomes.

🎛️ [**Agent Studio**](https://docs.voiceb.ai/agent-studio)\
Build and refine agents conversationally — describe your goal and Agent Studio configures the Goal, Script, Objections, and Guardrails. No prompt-writing, no scripts to maintain by hand.

🎯 **Outcome Engine**\
Every call is classified into deterministic, auditable outcomes — Sold, Transferred, Invalid, Low Intent. This is what makes outcome-based pricing possible and reporting credible to enterprise CFOs.

🔁 **Closed-Loop Conversion Intelligence**\
Every confirmed outcome is sent back to your ad platforms (Meta, Google, and more) — closing the loop from ad click to closed sale, so campaigns optimize on real revenue instead of form fills.

⚖️ **Compliance-First Voice Infrastructure**\
Caller ID preserved, SIP-native transfers, GDPR and EU AI Act ready by design.

🛠️ **No-Code Deployment**\
One line of JavaScript or a VoIP number — live in hours, not weeks.

🗣️ **Human-Quality Voice at Scale**\
Multilingual, low latency, natural conversations optimized for conversion.

🔒 **Enterprise-Grade Auditability**\
Full transcripts, recordings, outcomes, and decision traces for every interaction.

***

#### Proven in Production

* [**Holaluz (energy)**](https://docs.voiceb.ai/blog/holaluz-sales-autonomy-project)**:** 30.4% of valid inbound calls closed autonomously; 78% of total inbound traffic auto-filtered before a human is involved.
* [**Primagas / SHV Energy (energy)**](https://docs.voiceb.ai/blog/business-case-shv-energy-primagas-spain)**:** +150% sales, +250% lead-handling capacity, −37% cost per acquisition, with every lead attended in under 4 seconds.

***

**What We Don't Do**

* ❌ Cold outbound calling to purchased lists
* ❌ Generic voice automation

**What We Do**

👉 **Make inbound sales autonomous, efficient, and measurable.**

***

### Get Started

We've put together some helpful guides to get you set up with our product quickly and easily. Book a demo now:

{% hint style="success" %}
**Book a demo** [**here**](https://cal.com/team/voiceb-demo/voiceb-demo)**.**
{% endhint %}

***

## Quick links

{% content-ref url="/pages/ZxCHTBbiYRC6UjplhyJG" %}
[What we do](/overview/what-we-do)
{% endcontent-ref %}

{% content-ref url="/pages/dE35iGFuyxUUCoobGyKZ" %}
[VOICEB Features](/overview/voiceb-features)
{% endcontent-ref %}

## Get Started

We've put together some helpful guides for you to get setup with our product quickly and easily.

{% content-ref url="/pages/89FB48EssGPRVyrLUUoQ" %}
[Getting set up](/help-center/getting-set-up)
{% endcontent-ref %}

{% content-ref url="/pages/YSRx2CJjisNYnlE1hv8p" %}
[Setting permissions](/help-center/getting-set-up/setting-permissions)
{% endcontent-ref %}

{% content-ref url="/pages/ccfJgpwhryOw84odo1uU" %}
[Inviting Members](/help-center/getting-set-up/inviting-members)
{% endcontent-ref %}


# What we do

Last update: January 9th, 2026.

VoiceB turns inbound voice into a **fully autonomous revenue channel**.

We build **Autonomous Sales Voice Agents (ASVA)** that qualify, sell, route, and close inbound demand — at scale, in real time, and with measurable outcomes.

> Conversion breaks at the moment of intent.\
> The inbound call-center model is structurally incompatible with modern, dynamic sales.

VoiceB is the operating system that fixes that.

***

### The Problem We Solve

Inbound voice is still the **highest-intent sales channel**.\
Yet most companies run it with a broken operating model:

* Queues and missed calls
* Limited business hours
* High agent churn and ramp-up time
* Inconsistent knowledge and performance
* No real-time feedback loop to marketing

The result: **lost revenue at the exact moment of intent**.

This isn’t a tooling problem.\
It’s an autonomy problem.

***

### Our Approach: Sales Autonomy by Design

VoiceB replaces human-first inbound sales with **AI-first autonomy**, and brings humans in only when they add value.

We don’t deploy chatbots, IVRs, or call center automation.

We deploy **Autonomous Sales Voice Agents** that operate across four levels of autonomy:

#### Level 1 — Qualification

Filter invalid calls, qualify intent, and capture structured lead data.\
→ Save 20–50% of call center hours immediately.

#### Level 2 — Intelligent Routing

Route the right lead to the right closer based on intent, product, value, and availability.\
→ Higher close rates without increasing headcount.

#### Level 3 — Assisted Autonomous Sales

AI handles the sales conversation end-to-end. Humans step in only for edge cases.\
→ Higher conversion with fewer agents.

#### Level 4 — Full Autonomous Sales

AI closes the sale autonomously (for supported and compliant use cases).\
→ 24/7 revenue generation at unlimited scale.

One platform.\
You choose the level of autonomy.

***

### What VoiceB Actually Does

VoiceB operates at the moment of inbound intent and:

* Answers inbound calls instantly (WebRTC, click-to-call, phone, VoIP)
* Qualifies and scores demand in real time
* Handles objections, explains pricing, and pre-sells
* Routes or escalates only when needed
* Closes sales autonomously where possible
* Pushes structured outcomes to CRM and marketing systems

Every interaction produces a **deterministic outcome**, not just a transcript.

***

### Built for Enterprise Inbound Sales

VoiceB is designed for **high-volume, high-intent inbound environments**, including:

* Telecom
* Energy & Utilities
* Insurance
* Banking
* Automotive

These are revenue-critical operations — not demos or experiments.

That’s why VoiceB includes:

* Full audit trails and outcome versioning
* Compliance-ready data collection
* Real-time performance analytics
* CRM and webhook integrations
* Multi-provider voice and LLM reliability

***

### Voice Quality + Sales Intelligence

Yes, VoiceB sounds human.

We integrate best-in-class voice technologies (including providers like ElevenLabs and others) to deliver natural, human-grade conversations.

But voice is not the product.\
**Sales autonomy is.**

With VoiceB, you get:

* Best-in-class voice realism
* Plus autonomous qualification, selling, and closing
* Plus enterprise governance and measurement

The best of both worlds.

***

### How We Measure Success

We don’t optimize for:

* Minutes
* Agents
* Seats
* Demos

We optimize for **conversion**.

Every VoiceB deployment is measured by:

* Conversion lift
* Qualified demand filtered
* Sales capacity unlocked
* Revenue impact

That’s also why our pricing is **outcome-based**.

***

### In Short

VoiceB turns inbound voice into a **scalable, autonomous revenue engine**.

> No IVRs.\
> No chatbots.\
> No headcount bottlenecks.

Just measurable conversion — at scale.never sleeps, never misses a call, and knows exactly when to qualify, route, or close.

{% hint style="info" %}
Book a call with us [here](https://voiceb.ai/solutions/appointment-booking).&#x20;
{% endhint %}

***


# VOICEB Features

VoiceB.ai is a complete Autonomous Sales Voice Agent (ASVA) platform built to qualify, sell, route, and close inbound leads — without human intervention.

By Alex Bisbe. Updated on March 11th, 2026.

This page covers the full feature set available today.

***

### Agent Studio (Beta)&#x20;

[Agent Studio](https://docs.voiceb.ai/agent-studio) is the core interface for building, managing, and iterating on your AI sales agents. It gives business teams direct control over every layer of agent behavior — no engineering required.

<figure><img src="/files/y3cjwkHp2zYG8G2iED6W" alt=""><figcaption></figcaption></figure>

#### Campaign & Agent Management

* **Create & Edit:** I can set up new **Campaigns** and **Agents** (Inbound, Outbound/Click-to-Call).
* **Scripting:** I can write and modify agent **Script Blocks** (Goal, Script, Objections, Guardrails) to fine-tune how your AI speaks.
* **Phone Numbers:** I can list available **DDIs** and assign them to your agents.
* **Configuration:** Update agent settings, languages, voices, and provider models (Bland AI / ElevenLabs).

<figure><img src="/files/BDpIp9KGRdYBKaqC0LMO" alt=""><figcaption></figcaption></figure>

#### Analytics & Insights

* **Performance Metrics:** specific data on call volume, average duration, and conversion rates.
* **Visualizations:** I can generate **interactive charts** (Bar, Pie, Line, Funnel) to visualize lead outcomes or agent performance.
* **Agent Comparison:** Compare how different agents perform side-by-side.

<figure><img src="/files/fo1tRxGNE2BbnDOKHRcc" alt=""><figcaption></figcaption></figure>

#### Call Data

* **Deep Dive:** specific call details, including **transcripts**, summaries, and lead results (e.g., "Sold", "Not Interested").
* **Reporting:** I can export call logs to **CSV or Excel** for you to download.
* **Tagging:** Create and manage tags to categorize calls automatically.

#### Knowledge Integration

* **Fetch URLs:** Give me a link to a product page or FAQ, and I can read it to help write better scripts or enrich your agent's knowledge base.

{% hint style="info" %}
Desiged for Zero Prompt Mantainace
{% endhint %}

Agent Studio is the foundation for treating your sales agent with the same rigor as a production system — configurable by the people closest to the commercial reality, not dependent on IT.

***

### Voice Channels

VoiceB supports three distinct inbound channels, each suited to a different deployment context.

#### Inbound Phone (PSTN)

Assign a dedicated phone number to your agent. Any call to that number is handled autonomously by the AI agent. This is the standard model for inbound call center replacement.

#### Click-to-Call API

Trigger an outbound call to a prospect the moment they submit a form or click a call button. Your system sends a request to the VoiceB API with the prospect's phone number, and the agent calls immediately. [→ API documentation](https://docs.voiceb.ai/api-click-to-call)

You can also pass **variables** alongside each Click-to-Call request to personalize the conversation from the first second — injecting the prospect's name, product interest, campaign source, or any other field directly into the agent's context. [→ Using Variables in Click-to-Call](https://docs.voiceb.ai/using-variables-in-click-to-call-to-personalize-conversations)

#### Web Widget — WebRTC VoIP (SuperVoice)

The VoiceB Web Widget runs exclusively on WebRTC and launches an AI voice conversation directly in the browser — no phone number, no app switching, no friction. The user clicks, speaks, and the agent responds instantly.

This is the backbone of **SuperVoice**: VoiceB's product for converting website visitors during their digital journey. The widget embeds on any page with a single script tag and is configurable for brand, language, legal links, and agent assignment.

<figure><img src="/files/0nqnsgXRFEiPKQHQuLNZ" alt=""><figcaption></figcaption></figure>

WebRTC is especially powerful for mobile users: the conversation runs in the background while the user continues interacting with the page.

[→ Web Widget & WebRTC documentation](https://docs.voiceb.ai/new-voiceb-web-widget-and-webrtc-voip-channel)

### Enterprise Telephony — SIP Trunk Integration

*Available for enterprise deployments*

VoiceB agents can run as native nodes inside corporate telephony networks through direct SIP trunk integration. Instead of receiving forwarded calls via external infrastructure, the agent lives inside the operator's IP network — handling calls without leaving the enterprise perimeter.

We have completed this integration with one of Europe's largest telecommunications operators. VoiceB agents run natively on their internal network with no external routing and full data residency compliance. The operator is now packaging this capability as a product for their own enterprise clients — turning a single SIP integration into a distribution channel at scale.

For enterprises in telco, banking, and utilities, this is the difference between a tool that handles calls and infrastructure that is part of the network itself.

***

### **Custom Voice Selection**

VoiceB now supports full voice customization per agent. When configuring an agent, you can select the language, filter by gender (All, Female, Male), and choose from a curated list of available voices — or leave it on the default system voice for the selected language.

Since we use ElevenLabs as our voice engine, any voice available in the ElevenLabs library can be added to your account. If you want to use a voice that isn't already on the list — including custom cloned voices or any ElevenLabs voice ID — just send us the voice ID and we'll add it to your selection.

A few things to keep in mind:

* Voices with an associated ElevenLabs cost will be invoiced directly to you at cost, with no markup from VoiceB
* Voice availability may vary by language
* Custom voices can be configured independently per agent

<div data-with-frame="true"><figure><img src="/files/Ne14xYOQAR8sV5l2httr" alt=""><figcaption></figcaption></figure></div>

This gives enterprise clients full control over brand voice and regional accent preferences without being limited to a predefined list.

***

### Prompt Control — Script Blocks

Building an effective AI sales agent has always required a balance between structure and flexibility. Too rigid, and the agent can't adapt to real conversations. Too loose, and it drifts off-script at the wrong moments.

<figure><img src="/files/tI2znkoKkRAl6UoAXbfU" alt=""><figcaption></figcaption></figure>

Script Blocks solve this by breaking the agent prompt into discrete, independently editable sections — each with a specific role in the conversation.

Instead of managing one long prompt where a small edit can unintentionally affect unrelated behavior, you work with named blocks: **Goal**, **Role**, **Objective**, **Objections**, **Guardrails**, and any custom blocks your setup requires. Each block has its own editing area with rich text formatting, header support, lists, and inline variable support using `{variable_name}` syntax.

This has meaningful practical consequences:

* You can update your objection-handling logic without touching the qualification flow
* You can iterate on the closing pitch independently of the opening framing
* You can add a new product block without restructuring the entire prompt
* Different team members can own different blocks without stepping on each other

Every change is automatically versioned. The **Script History** panel shows a full timeline of edits — including full snapshots, block reorders, and individual text changes. Each version supports **Preview**, **Diff**, and **Restore**, so you can compare any two states of the agent and roll back instantly if needed.

This moves prompt management from an art practiced by one technical person into a repeatable, auditable process that business teams can own directly — without opening a ticket, without waiting, and without risk of losing a version that was working.

> Your agent's behavior should be as easy to change as a Google Doc and as safe to iterate on as production code. Script Blocks get you both.

***

### Outcomes & Status

Every call in VoiceB is classified with a **Status** and an **Outcome**. This structure drives reporting, billing, CRM routing, and integrations.

<figure><img src="/files/V6DUBxD1iUt8UXGpBi78" alt=""><figcaption></figcaption></figure>

**Status** has two values:

* **Valid** — A real person was reached, the conversation had intent, and the lead fits your commercial scope.
* **Invalid** — The call did not produce a commercially meaningful interaction (voicemail, dropped call under 15 seconds, not interested, customer care).

**Outcomes for Valid calls:**

* **Sold** — The customer accepted the offer and provided the required data to complete the sale.
* **Talk to Agent** — The lead qualifies but requires a human for final closing or validation.

Both Sold and Talk to Agent can carry an additional **Transferred** label when real-time call transfer is enabled.

**Outcomes for Invalid calls:** Not Interested, Voicemail, Dropped, Customer Care.

Invalid calls are excluded from billing and from sales metrics — you only pay for commercially meaningful interactions.

[→ Full Status & Outcomes reference](https://docs.voiceb.ai/understanding-status-and-outcomes-in-voiceb)

***

### Tags — Custom AI Classification

Tags are a second, fully configurable outcome layer applied automatically by the AI after processing the full conversation.

<figure><img src="/files/Tpav8I0VNhvQ0CwvKQAm" alt=""><figcaption></figcaption></figure>

Where Outcomes tell you *what happened*, Tags tell you *what the call means for your business*. Each tag is defined in natural language:

* **True if** — a plain-language condition (e.g. "The user expressed interest in a solar panel add-on")
* **False if** — the inverse
* **Color** — for visual identification in the dashboard
* **Valid / SOLD override** — marks the tag as a completed conversion in reporting, independently of the voice outcome label

Tags enable precise CRM routing, lead enrichment, and segmentation without touching your existing reporting structure.

[→ Tags documentation](https://docs.voiceb.ai/tags-custom-ai-classification-for-calls-and-leads)

***

### Memory — Conversation Recall

Memory allows agents to resume previous conversations when the same phone number calls again.

When a returning caller is detected, the agent asks for name confirmation (GDPR compliance). If the name matches, the agent greets them by name, summarizes the previous conversation, and picks up exactly where it stopped — without re-asking already confirmed information.

The agent remembers: pricing discussed, tariff selected, objections raised, and the last completed step.

<figure><img src="/files/SwsKA734vkdFQGZBHlsh" alt=""><figcaption></figcaption></figure>

Memory is configured per agent with a configurable window in days (e.g. 14, 30, or 90 days). Each agent can have a different memory duration depending on its sales cycle.

Memory does not affect CRM logic — each call still generates its own transcript, summary, outcomes, and structured data. Memory affects the conversation experience, not the reporting layer.

[→ Memory documentation](https://docs.voiceb.ai/memory-conversation-recall)

***

### Call Transfer — AI to Human Handoff

When predefined conditions are met — lead qualified, customer requests a human, or business rules trigger escalation — VoiceB can transfer the call in real time to a human agent.

For WebRTC calls, VoiceB converts the VoIP session into a standard phone call (PSTN) and routes it to your call center with full conversational context. This means humans receive only pre-qualified, high-intent leads — never raw inbound traffic.

[→ Call Transfer documentation](https://docs.voiceb.ai/call-transfer-redirect-calls-to-a-human-agent)

***

### Real-Time Data Checks via Customer Endpoints

Agents can query your systems live during a conversation — with no perceived latency and no interruption to the call flow.

Common use cases:

* Retrieve a personalized price based on a customer identifier (e.g. CUPS, meter number, account ID)
* Check real-time agent availability before initiating a transfer
* Validate identity or existing customer status
* Fetch a counter-offer when a prospect rejects the initial price
* Check if the call is within business hours

Endpoints are configured per agent and called automatically when the relevant conversation trigger is reached.

[→ Real-Time Endpoints documentation](https://docs.voiceb.ai/real-time-data-checks-via-customer-endpoints)

***

### CRM Integrations

<figure><img src="/files/J6lRGMO20lBeKWLOsXBu" alt=""><figcaption></figcaption></figure>

#### HubSpot

Native two-way integration. Map VoiceB data fields to HubSpot contact and deal properties, set trigger conditions (which outcomes push data), and inject CRM Constants — fixed field values automatically added at sync time, such as campaign source, pipeline stage, or owner assignment.

[→ HubSpot integration guide](https://docs.voiceb.ai/how-to-connect-your-voiceb.ai-agent-with-hubspot)

#### Salesforce

Native integration with full field mapping. Connect via OAuth, map structured call data to your Salesforce schema, and control which outcomes trigger record creation or update.

[→ Salesforce integration guide](https://docs.voiceb.ai/how-to-connect-your-voiceb.ai-agent-with-salesforce)

#### Webhooks

Send call outcomes, structured data, transcripts, and tags to any external system via webhook on call completion.

***

### Campaign Dashboard

The Campaign Dashboard provides operational visibility across all your agents and inbound sources.

* **Search and filter** by phone number, outcome, agent, tag, date range, and more
* **Switchable bar chart** to visualize call distribution by Status, Outcome, or Tags — normalized to 100% for easy comparison
* **CSV export** of filtered call data for offline analysis or reporting
* Full call list with recordings, transcripts, summaries, and structured data per call

<figure><img src="/files/Prp9U3LrzbSCv6CZC2Yh" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/gLaFm5PhMR0zXUlRlaMI" alt=""><figcaption></figcaption></figure>

***

### Balanced Voice Routing

VoiceB automatically routes calls between **ElevenLabs** and **Bland.ai** — two of the most advanced AI voice providers available — based on real-time performance and availability.

This means your agents always use the best-performing voice infrastructure at any given moment, without manual configuration. If one provider degrades, routing shifts automatically. You get best-in-class voice quality and reliability, by default.\
\
[→ Balaned Voice Routing](https://docs.voiceb.ai/blog/voiceb-vs-elevenlabs-bland-retell-decagon-and-salesforce-agentforce)<br>

***

### Coming Soon — Direct Lead Source Connections

> *Google Ads · Meta Ads · and more*

Today, VoiceB handles what happens after a lead arrives. What comes next is closing the loop at the source.

We are building native connections to the major digital advertising platforms — starting with **Google Ads** and **Meta Ads** — so that leads generated by your campaigns flow directly into VoiceB agents the moment they are created, with zero manual steps in between.

When someone fills out a Lead Form on Meta or submits their details through a Google Ads lead extension, VoiceB will trigger an outbound call within seconds. The agent already knows the campaign context, the product the lead expressed interest in, and any data collected in the form. The conversation starts informed, not cold.

This closes a gap that has existed in performance marketing for years: the lag between lead creation and first contact. Every minute between a lead submitting a form and someone calling them back is intent decaying. With direct lead source connections, that window collapses to near zero.

**What this unlocks:**

The combination of unlimited calling capacity and instant response creates a fundamentally different economics for inbound demand. Instead of optimizing for lead quality at the top of the funnel because your team can only handle a finite volume, you can open the funnel wide — more campaigns, broader targeting, higher volume — and let VoiceB absorb all of it.

Every lead gets called. Every call is handled by a trained agent that knows the offer, handles objections, and closes or qualifies in real time. The constraint of human capacity disappears.

This means:

* **Bottom-of-funnel revenue delivered automatically** — from ad click to closed sale, end to end, without a human in the loop
* **Unlimited conversion capacity** — your ability to convert is no longer tied to headcount or call center hours
* **Massive top-of-funnel leverage** — invest more in acquisition knowing that every lead will be worked, not just the ones your team gets to

For companies generating thousands of inbound leads per month, this is not an efficiency improvement. It is a structural change in how revenue is generated from paid media.

> The future is not optimizing your lead handling. It is removing the ceiling on it entirely.

***

### 👉🏼 Full Feature Table 🧐

| Feature                      | Description                                                                     | Documentation                                                                                                |
| ---------------------------- | ------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ |
| Agent Studio                 | Script blocks, version history, A/B testing, conversational analytics           | [Feb 2026 release](https://docs.voiceb.ai/product-updates/february-2026-voiceb.ai-product-release)           |
| Inbound Phone Channel        | Dedicated DDI for AI-handled inbound calls                                      | —                                                                                                            |
| Click-to-Call API            | Trigger outbound AI calls via API on form submit or CTA                         | [→](https://docs.voiceb.ai/api-click-to-call)                                                                |
| Variables in Click-to-Call   | Personalize conversations by passing context at call time                       | [→](https://docs.voiceb.ai/using-variables-in-click-to-call-to-personalize-conversations)                    |
| Web Widget — WebRTC VoIP     | Browser-native voice channel, no phone required (SuperVoice)                    | [→](https://docs.voiceb.ai/new-voiceb-web-widget-and-webrtc-voip-channel)                                    |
| Status & Outcomes            | Structured call classification: Valid/Invalid + Sold/Talk to Agent/Transferred  | [→](https://docs.voiceb.ai/understanding-status-and-outcomes-in-voiceb)                                      |
| Tags                         | Custom AI classification layer per call, configurable in natural language       | [→](https://docs.voiceb.ai/tags-custom-ai-classification-for-calls-and-leads)                                |
| Memory — Conversation Recall | Cross-call continuity with GDPR-compliant name validation                       | [→](https://docs.voiceb.ai/memory-conversation-recall)                                                       |
| Call Transfer                | Real-time AI-to-human handoff, including VoIP-to-PSTN                           | [→](https://docs.voiceb.ai/call-transfer-redirect-calls-to-a-human-agent)                                    |
| Real-Time Endpoint Queries   | Live data checks during calls (pricing, availability, identity)                 | [→](https://docs.voiceb.ai/real-time-data-checks-via-customer-endpoints)                                     |
| HubSpot Integration          | Native CRM sync with trigger conditions and constants                           | [→](https://docs.voiceb.ai/how-to-connect-your-voiceb.ai-agent-with-hubspot)                                 |
| Salesforce Integration       | Native CRM sync with field mapping and outcome filtering                        | [→](https://docs.voiceb.ai/how-to-connect-your-voiceb.ai-agent-with-salesforce)                              |
| Webhook Integration          | Push call data to any external system on call completion                        | [→](https://docs.voiceb.ai/webhook-call-outcome-integration)                                                 |
| CSV Export                   | Download filtered call data from the Campaign Dashboard                         | —                                                                                                            |
| Campaign Dashboard           | Search, filter, visualize, and audit all calls across agents                    | —                                                                                                            |
| Data Collection Schema       | Define structured fields to capture during every call                           | [→](https://docs.voiceb.ai/create-a-data-collection-schema)                                                  |
| Balanced Voice Routing       | Automatic routing between ElevenLabs and Bland.ai for best-in-class reliability | [→](https://docs.voiceb.ai/blog/introducing-balanced-voice-routing-reliability-speed-and-scale-for-voice-ai) |
| Knowledge Base               | Upload product content, FAQs, and objection handling for agent reference        | —                                                                                                            |
| Script Blocks & History      | Structured prompt editor with full version control and restore                  | [Feb 2026 release](https://docs.voiceb.ai/product-updates/february-2026-voiceb.ai-product-release)           |
| CRM Trigger Conditions       | Control which outcomes fire CRM sync (Feb 2026)                                 | [Feb 2026 release](https://docs.voiceb.ai/product-updates/february-2026-voiceb.ai-product-release)           |
| CRM Constants                | Inject fixed field values at CRM sync time (Feb 2026)                           | [Feb 2026 release](https://docs.voiceb.ai/product-updates/february-2026-voiceb.ai-product-release)           |

***

### Security & Compliance

* Hosted on AWS with Twilio EU voice infrastructure — all data processed within Europe
* EU AI Act compliant: agents self-identify as AI at the start of every call
* SOC2 Type II in progress
* ISO 27001:2013 in progress
* Full call recording and transcription stored for audit
* GDPR-compliant Memory validation logic built in by default
* Role-based access control and end-to-end encryption in transit and at rest

***

#### Dashboard video

{% embed url="<https://www.loom.com/share/42f9f906163b44f3bc1d320762887c8a>" %}


# Pricing Guide

Updated on July 21st, 2026.

VoiceB pricing is designed around **outcomes, not software usage**.

We don't charge for seats, agents, minutes, or licenses. We price based on **what the Autonomous Sales Voice Agent (ASVA) actually delivers**: qualified demand, sales-ready conversations, and closed revenue.

This guide explains how the three pricing components work together — outcomes, telephony (DDI), and outbound call charges — and details exactly what is included in each plan (Starter, Enterprise, Corporate Custom) on [voiceb.ai/pricing](https://voiceb.ai/pricing).

***

**Pricing Philosophy**

Traditional contact-center and AI platforms price on agents, minutes, licenses, or conversations. This misaligns incentives: vendors get paid whether or not the customer wins.

VoiceB pricing is **performance-based**:

* If ASVA delivers value, you pay.
* If it doesn't, you don't.

This aligns VoiceB with Sales, Marketing, and Finance from day one.

***

**The Three Pricing Components**

VoiceB pricing has **three components**, all transparent and billed independently:

1. **ASVA outcome-based pricing** — what you pay for value delivered.
2. **DDI / DID telephony costs** — monthly fee per phone number.
3. **Outbound call charges** — apply only when the Outbound Reactivation module is active (Enterprise and Corporate Custom). Included by default for major markets; exceptions apply.

Each component is intentionally separated so Finance teams can model, audit, and forecast them without bundled surprises.

***

**1. ASVA Outcome-Based Pricing**

You pay only when ASVA produces a defined outcome. Three plans are available, all with access to both use cases — **AI Lead Qualification (L1–L2)** and **Autonomous Sales (L3–L4)**.

**Plan Comparison**

| Outcome                                      | Starter                     | Enterprise                  | Corporate Custom                    |
| -------------------------------------------- | --------------------------- | --------------------------- | ----------------------------------- |
| Unqualified / invalid calls                  | €0 / $0 / £0                | €0 / $0 / £0                | €0 / $0 / £0                        |
| Qualified call transferred to human agent    | €2.00 / $2.16 / £1.70       | €1.50 / $1.62 / £1.28       | Custom                              |
| Pre-sale (final step before payment)         | €9.00 / $9.72 / £7.65       | €7.00 / $7.56 / £5.95       | Custom                              |
| Completed sale with integration confirmation | €20.00 / $21.60 / £17       | €15.00 / $16.20 / £12.75    | Custom                              |
| Minimum monthly commitment                   | €700 / $756 / £595          | €2,000 / $2,160 / £1,700    | Volume-based                        |
| Minimum volume                               | —                           | —                           | 10,000 leads/month                  |
| Contract                                     | Annual, 3-month exit window | Annual, 3-month exit window | 12-month auto-renewal, 3-month exit |

Pricing shown in EUR, USD and GBP. Also available on the [pricing page](https://voiceb.ai/pricing).

**Plan Selection Logic**

* **Starter** — SMB and self-serve teams running a single inbound campaign with a 3-month PoC structure.
* **Enterprise** — High-volume inbound teams running multiple campaigns, with human onboarding, dedicated Customer Success, and access to the Outbound Reactivation module.
* **Corporate Custom** — Enterprise-grade volume (10,000+ leads/month) with dedicated SIP Trunk, dedicated CSM, 99.9% SLA, and fully custom per-outcome pricing.

**Outcome Definitions**

VoiceB uses **deterministic outcomes**, not subjective interpretations. Outcomes are configurable per customer and versioned for auditability. See [outcome definitions](https://docs.voiceb.ai/understanding-status-and-outcomes-in-voiceb) for the full reference.

***

**What's Included in Each Plan**

All three plans include access to all four autonomy levels (L1–L4) and both use cases (AI Lead Qualification and Autonomous Sales). Plans differ in commitment, support level, and operational scope.

**Starter — For SMB & self-serve**

* Minimum monthly commitment: €700 / $756 / £595
* 3-month PoC
* Annual contract with 3-month exit window
* 1 campaign — multiple AI agents per campaign
* AI agent for onboarding and AI Customer Success: included
* [**Agent Studio**](https://docs.voiceb.ai/agent-studio) (Beta) — AI assistant inside the platform:
  * Build and configure agents
  * Manage script blocks (Goal, Objections, Guardrails)
  * Query calls and outcomes in natural language
  * Run analytics, export CSV/Excel reports
  * Manage tags, configure widgets
  * Enrich scripts from any URL
* Best-in-class LLM orchestration — OpenAI, Anthropic, Google, Mistral, and many more
* Multi-language agents (see supported languages below)
* Call triggers — Inbound & callbacks:
  * Inbound DDI / DID (one number per agent)
  * Click-to-Call / Callback API — sub-4s reactive callbacks
  * VoIP / WebRTC web widget
* Tools (endpoint / MCP) — connect 3rd-party tools during the call (pricing checks, calendar availability, agent availability, and more)
* Post-call integrations — CRM-native or webhook:
  * HubSpot native integration
  * Salesforce native integration
  * Post-call webhook
* Customizable web widget
* Analytics — dashboard + CSV
* Phone costs included (within supported markets)
* Custom Integrations: €900 / DevDay ; $972 / DevDay or £765 / DevDay (optional)
* DID / DDI (one per agent) — price per number, see telephony section below
* Payment via Stripe

**Enterprise — For high-volume inbound teams**

*Everything in Starter, plus:*

* Minimum monthly commitment: €2,000 / $2,160 / £1,700
* Multiple campaigns — multiple AI agents per campaign
* Human onboarding: included
* Dedicated Customer Success: included
* Lower per-outcome pricing across all tiers (see plan comparison table above)
* **Outbound Reactivation module** — trigger autonomous calls from your own CRM:
  * Trigger API + automatic retries to maximize contact ratio
  * All post-call features included: outcomes, CRM write-back, closed-loop ad signals
  * Multi-language; runs on your own opted-in database — never cold lists
  * Requires a CRM integration to fire triggers
  * Governed by the Outbound Policy (see section 3)

**Corporate Custom — For enterprise-grade volume**

*Everything in Enterprise, plus:*

* Minimum volume: 10,000 monthly leads
* Contract: 12 months auto-renewal with 3-month exit window
* Dedicated Customer Success Manager included
* **Dedicated SIP Trunk** — Private Call Routing
* SLA: 99.9% uptime
* Custom outcome-based pricing
* Payment via wire transfer

***

**ASVA Autonomy Levels**

All plans include access to all four autonomy levels. Customers typically start at L1–L2 and scale over time without changing platforms.

| Autonomy Level                     | Description                                                  |
| ---------------------------------- | ------------------------------------------------------------ |
| **L1 — Qualification**             | Filter invalid calls, detect intent, capture structured data |
| **L2 — Intelligent Routing**       | Route sales-ready demand to the right human                  |
| **L3 — Assisted Autonomous Sales** | AI sells, humans step in for edge cases                      |
| **L4 — Full Autonomous Sales**     | AI closes end-to-end                                         |

Higher autonomy unlocks higher-value outcomes (transfers → pre-sales → closed sales). Pricing scales with the outcome produced, not the level configured.

***

**Supported Languages**

ASVA agents operate natively in the following languages, on both inbound and outbound flows:

Spanish · Catalan · Spanish (Latin America) · English (US) · English (UK) · German · French · Portuguese · Portuguese (Brazil) · Italian · Dutch · Dutch (Belgium) · Greek · Arabic · Romanian · Polish · Slovenian.

Additional languages are added for free on request — contact Customer Success.

***

**2. Telephony (DDI / DID) Costs**

VoiceB supports **PSTN callbacks** for enterprise deployments that require inbound phone numbers in addition to (or instead of) WebRTC.

For PSTN-based flows, each ASVA requires a **dedicated DID number**. DID numbers are region-specific, billed monthly, and treated as **infrastructure costs** — separate from outcome pricing.

**DDI / DID Monthly Pricing**

> One DID number per AI agent. Billed monthly.

| Region                      | Prefix | Price / Month |
| --------------------------- | ------ | ------------- |
| USA                         | +1     | $1.50         |
| UK                          | +44    | £2.50         |
| All other supported markets | —      | €5.50         |

The €5.50 standard rate applies to every supported market (EU-27, LATAM, and others), with the USA and UK as the only exceptions above.

**Local DDI — Regulatory Requirement**

> Due to local telecom regulations, some countries require a **registered local address and/or a company VAT number** to provision a local DDI. This applies to both inbound and outbound numbers. Where a customer cannot meet the local requirement, VoiceB can provision an alternative number format or advise on the best available option. Confirm requirements for your target countries with Customer Success before provisioning.

**Notes**

* DDI pricing applies regardless of plan or autonomy level.
* Customers using **WebRTC-only** (SuperVoice widget, click-to-call) may not require DIDs.
* Hybrid setups (WebRTC + PSTN) are fully supported.

***

**3. Outbound Reactivation — Policy & Call Charges**

**Availability**

The **Outbound Reactivation module** is available on **Enterprise and Corporate Custom** plans only. It is not included in Starter.

Reactive callbacks (Click-to-Call / Callback API — where a lead requests a call and ASVA calls back) are available on **all plans, including Starter**. These are not the Reactivation module.

**What the module does**

Reactivation runs on **your own opted-in database** — abandoned carts, expiring contracts, incomplete sign-ups, recent purchases — never on purchased or cold lists. It includes:

* Trigger API to fire calls from CRM events
* Automatic retry logic to maximize contact ratio
* All post-call features: deterministic outcomes, CRM write-back, and closed-loop conversion signals to your ad platforms
* Multi-language execution
* A required CRM integration to fire triggers

**Outbound Policy**

* Reactivation operates exclusively on the customer's own opted-in contacts. Purchased lists and cold outreach are not permitted.
* The customer is responsible for holding a **lawful basis** to contact each number.
* VoiceB enforces **per-market calling windows** and **opt-out suppression** at the platform level.
* In the event of misuse, VoiceB may **suspend the Outbound Reactivation module without suspending the underlying contract or inbound service**.
* Outbound is regulated differently in each market. Confirm target markets with Customer Success before launching campaigns.

**Call charges**

When the module is active, outbound **call charges are included by default** for the following markets:

* **North America** — United States, Canada
* **European Union** — all 27 member states
* **United Kingdom**
* **Oceania** — Australia, New Zealand
* **Major Latin America markets** — including Mexico, Brazil, Argentina, Chile, Colombia, Peru

For destinations outside this list, **contact Customer Success for confirmation** before launching campaigns. Some destinations (premium-rate ranges, certain African and Caribbean numbers, satellite phones) may carry per-minute surcharges billed separately at cost.

Inbound calls to your DID numbers are always included in the monthly DID fee.

***

**Pricing Modification Policy**

VoiceB reserves the right to modify **DDI / DID monthly fees** and **outbound call charges** at any time without prior notice, in line with changes from underlying telephony providers and currency markets.

ASVA outcome-based pricing is fixed for the duration of your contract. Modifications to telephony pricing do not affect outcome rates and are reflected on the next invoicing cycle.

***

**Why We Separate the Three Components**

We intentionally **do not bundle** telephony into outcome pricing because:

* Telephony costs vary significantly by geography and carrier.
* Customers and Finance teams require predictability and auditability.
* Bundled pricing creates hidden margins that distort unit economics at scale.

ASVA pricing reflects **value delivered**. DDI and call pricing reflect **infrastructure consumed**. Both are visible on every invoice.

***

**Example Pricing Scenarios**

**Scenario 1 — Qualification at Scale (L1, Starter)**

High inbound volume, significant noise, human sales team in place.

**You pay:**

* €0 for invalid/unqualified calls
* €2 per qualified lead transferred
* DDI monthly fee per agent

**Typical impact:** 20–50% call center hour savings, cleaner CRM, higher downstream conversion.

***

**Scenario 2 — Qualification + Routing (L2, Enterprise)**

Sales teams overwhelmed; misrouted or low-quality leads hurting conversion.

**You pay:**

* €1.50 per transferred sales-ready call
* DDI monthly fees
* Outbound charges included (when Reactivation module active)

**Typical impact:** Immediate conversion uplift, lower AHT, no headcount increase.

***

**Scenario 3 — Assisted Autonomous Sales (L3, Enterprise)**

Complex products. AI handles most of the conversation, humans intervene on edge cases.

**You pay:**

* €7 per pre-sale (final step before payment)
* €15 per completed sale with CRM integration confirmation
* DDI monthly fees
* Outbound charges included (when Reactivation module active)

**Typical impact:** Higher close rates, fewer agents required, faster ramp on new offers.

***

**Scenario 4 — Full Autonomous Sales (L4, Corporate Custom)**

Repeatable, compliant sales flows. End-to-end AI closing at 10,000+ leads/month.

**You pay:**

* Custom per-outcome pricing aligned to volume
* Dedicated SIP Trunk (private call routing)
* 99.9% SLA, dedicated CSM

**Typical impact:** 24/7 revenue generation, zero queue, no dependency on agent availability.

***

**What VoiceB Does&#x20;*****Not*****&#x20;Charge For**

* Seats or agents
* Minutes or call duration (within included markets)
* AI tokens
* Environments or sandboxes
* LLM orchestration overhead

You only pay for **outcomes**, **DDI numbers**, and **outbound calls outside included markets**.

***

**Enterprise Pilots & Commitments**

Most customers start with a **controlled KPI pilot**:

* Defined autonomy level and outcome targets
* Baseline metrics agreed upfront
* Clear success criteria
* Short pilot duration (typically the 3-month PoC included in Starter and Enterprise plans)

If outcomes are not achieved, customers do not scale. Long-term contracts are based on proven ROI, volume commitments, and an autonomy expansion roadmap.

***

**Summary**

* VoiceB pricing is **outcome-based** — three plans (Starter, Enterprise, Corporate Custom), all with access to all autonomy levels and both use cases.
* **Starter** includes the full platform, Agent Studio, integrations (HubSpot, Salesforce, webhook), MCP tools, multi-language agents, inbound and reactive callbacks, and AI-driven onboarding/CS — annual contract with 3-month exit.
* **Enterprise** adds multiple campaigns, human onboarding, dedicated Customer Success, lower per-outcome pricing, and the **Outbound Reactivation module**.
* **Corporate Custom** adds dedicated SIP Trunk, dedicated CSM, 99.9% SLA, custom outcome pricing, and wire-transfer billing — for 10,000+ monthly leads.
* **DDI pricing** is €5.50 / month for all supported markets, with the USA ($1.50) and UK (£2.50) as exceptions. Some countries require a local address and/or company VAT number to provision a local DDI.
* **Outbound Reactivation** (Enterprise and Corporate Custom only) runs on your own opted-in database; call charges are included for North America, EU, UK, Oceania, and major Latam markets — contact CS to confirm other destinations.
* DDI fees and outbound call charges may be modified at any time without prior notice.

Need help estimating pricing for your inbound volume or autonomy level? [Talk to our team](https://cal.com/team/voiceb-demo/voiceb-demo).

***

**🧠 Need Help? 📩 Contact our sales team** [**here**](https://cal.com/team/voiceb-demo/voiceb-demo)

***


# Getting Started with VoiceB.ai in 20 Minutes

This guide walks you through the fastest path to a live autonomous sales agent — from zero to handling real inbound calls, with a CRM receiving the results.

***

There are three steps: create your agent, set up call transfer for when you need a human, and connect your CRM.

***

### Step 1 — Create your agent with Agent Studio

Agent Studio is VoiceB's conversational AI interface for building and managing agents. You don't write prompts manually — you describe what you need, and Agent Studio configures the agent for you.

<figure><img src="/files/Gx3H6RLriDgKnmchQ1Fg" alt=""><figcaption></figcaption></figure>

**Start a new session in Agent Studio and tell it:**

* What your company sells
* Who your target caller is
* What a successful call looks like (a sale, a qualified lead, a booked appointment)
* Any key objections your sales team regularly handles

You can also give Agent Studio a URL — a product page, a pricing page, an FAQ — and it will read it to inform the agent's knowledge base and script.

Agent Studio will generate the agent's full configuration: Goal, Script Blocks, Objections, and Guardrails. You can review and edit each block independently. If something doesn't sound right, tell it — the agent improves through conversation, not by editing raw prompts.

<figure><img src="/files/qOB2np9SsA2JwOyQeo5m" alt=""><figcaption></figcaption></figure>

Once the agent is configured, Agent Studio will assign a **dedicated phone number (DDI)** to it. Any call to that number is handled autonomously by the AI agent. To go live, simply redirect your inbound traffic — from your website, your campaigns, or your existing phone line — to that number. That's it.

No code. No IT. No integration required to start receiving calls.

> For a full walkthrough of Agent Studio's capabilities — including analytics, script versioning, and agent comparison — see the [VOICEB Features →](https://docs.voiceb.ai/overview/voiceb-features)

***

### Step 2 — Set up Call Transfer (optional but recommended)

When a caller wants to speak to a human — or when your business rules require it — VoiceB can transfer the call in real time to your team.

To enable this, you need two things:

1. A condition that triggers the transfer (e.g. "caller asks to speak to an agent", "sale requires human verification", "product is out of scope")
2. A phone number to receive the transferred call

You can configure both directly in Agent Studio. Tell it the condition in plain language and provide the destination number. The agent will handle the conversation, qualify the lead, and hand off only when the condition is met — so your team receives warm, pre-qualified calls, not raw inbound traffic.

> [Call Transfer documentation →](https://docs.voiceb.ai/call-transfer-redirect-calls-to-a-human-agent)

***

### Step 3 — Connect your CRM

After every call, VoiceB generates structured data: outcome, transcript, summary, collected fields, and tags. This step sends that data automatically to your CRM.

#### HubSpot

Go to **Integrations → HubSpot** and connect your account. Once connected:

* Map VoiceB fields (name, email, outcome, tags, etc.) to your HubSpot contact and deal properties
* Set trigger conditions — choose which outcomes push data (e.g. only Sold and Talk to Agent, not Invalid or Dropped)
* Add CRM Constants if needed — fixed values injected at sync time, such as campaign source, pipeline stage, or owner

> [HubSpot integration guide →](https://docs.voiceb.ai/how-to-connect-your-voiceb.ai-agent-with-hubspot)

#### Salesforce

Go to **Integrations → Salesforce** and connect via OAuth. Once connected:

* Map structured call data to your Salesforce schema
* Control which outcomes trigger record creation or update
* Each call syncs automatically with full outcome data

> [Salesforce integration guide →](https://docs.voiceb.ai/how-to-connect-your-voiceb.ai-agent-with-salesforce)

***

### You're live

At this point your agent is receiving calls, handling the conversation autonomously, transferring when needed, and pushing results to your CRM without manual intervention.

<figure><img src="/files/zlEzZ1VmUS6p462eLwpH" alt=""><figcaption></figcaption></figure>

From here, Agent Studio lets you iterate: ask it why calls aren't converting, compare two script versions, or update the objection-handling block — all in natural language, without touching a prompt directly.

> For everything else — data collection schemas, tags, memory, real-time endpoint queries, and more — see the [full feature documentation →](https://docs.voiceb.ai/overview/voiceb-features)


# How to use the Agent Studio – Product Guide

By Alex Bisbe. June 29th 2026

{% hint style="warning" %}
**Beta · Admin only.** Agent Studio is in Beta and available to admin users. This guide assumes you already know what Agent Studio is — if not, start with the [Agent Studio reference page](https://docs.voiceb.ai/agent-studio).
{% endhint %}

This guide is about *technique*. The reference page tells you what Agent Studio can do; this one shows you how to ask for it so you get accurate, useful answers the first time — and how to chain those answers into real day-to-day work.

<figure><img src="/files/m7DbbLzUPwa5m4jWanVw" alt=""><figcaption></figcaption></figure>

## The one habit that changes everything: be specific

Agent Studio works against your live data, so vague questions get vague answers. The fastest way to improve a result is to add the three things the assistant can't guess: **scope, time range, and the shape of the output you want.**

Compare:

{% hint style="danger" %}
**Weak:** "How are my campaigns doing?"
{% endhint %}

{% hint style="success" %}
**Strong:** "For the last 30 days, show total calls and outcome breakdown per campaign in a table, sorted by call volume."
{% endhint %}

The second prompt names the period (last 30 days), the scope (per campaign), and the format (a sorted table). You'll get exactly that, instead of a generic summary you then have to refine.

A simple template that works for almost any analytics request:

```
[What] for [which campaigns/agents] over [time range], as a [table/chart/list], sorted by [metric].
```

## Work in a thread, not in one-shot questions

Agent Studio remembers the current conversation. The best workflows start broad and drill down, each question building on the last:

1. "Show me metrics for the last 30 days across all campaigns."
2. "Now focus on VOICEB.AI - WEB — why is the average duration only 31 seconds?"
3. "List the 10 most recent Dropped calls for that campaign with their transcripts."
4. "Export those to Excel."

Because the assistant keeps context, you don't repeat the campaign name or date range each time. When you start a genuinely new topic, open a fresh conversation (the **+** in the Conversations panel) so old context doesn't bleed into the new analysis.

<figure><img src="/files/BH3rguEEjLxp2srjDy8N" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
**Name your conversations.** Rename a thread the moment it becomes useful ("Q2 WEB campaign audit") so you can find it again. The Conversations panel is searchable.
{% endhint %}

## Choosing the right model for the job

All three models perform the same actions — the difference is how the answer reads. In practice:

| If you want…                           | Try                   | Why                          |
| -------------------------------------- | --------------------- | ---------------------------- |
| A fast, thorough dump of numbers       | **Gemini 3.5 Flash**  | Default; detailed and quick. |
| A tight, skimmable answer for a report | **Claude Sonnet 4.6** | Concise and well-structured. |
| A guided "walk me through it" flow     | **GPT 5.2 Chat**      | Goal-oriented, step-by-step. |

You can switch mid-thread without losing context. A common pattern: explore with the default, then switch to Claude and ask "summarize the key findings from this conversation in five bullets" when you need something to paste into a Slack update.

## Five workflows worth saving

{% stepper %}
{% step %}

## The Monday performance review

> "Compare this week's call volume and conversion to last week, per campaign, as a bar chart. Then list any campaign where conversion dropped more than 10%."

A two-minute health check that surfaces problems before they compound.
{% endstep %}

{% step %}

## Diagnose a weak agent before blaming it

VoiceB's rule of thumb: poor outcomes usually trace to traffic quality, not the agent. Use Agent Studio to check that first.

> "For \[campaign], show the outcome breakdown and average duration. Then list 15 recent Not Interested calls with transcripts so I can see what the incoming traffic looks like."

If the transcripts show off-target callers, the lead source is the issue — not the script.
{% endstep %}

{% step %}

## Build a new agent without the forms

> "Help me create a new inbound agent for residential energy. Walk me through it one step at a time."

The assistant builds the script blocks (Goal, Script, Objections, Guardrails) with you, one decision per turn. Review each block before confirming.
{% endstep %}

{% step %}

## Enrich an agent from a live web page

> "Read https\://\[yoursite]/faq and turn the key points into objection-handling notes I can add to my agent's script."

Great for keeping agent knowledge current with your public pages without copy-pasting.
{% endstep %}

{% step %}

## Tidy your call data with tags

> "Create a tag called 'Spanish Speaker' that applies when the caller speaks Spanish, and another called 'Pricing Objection' for calls where the customer pushes back on price."

Once tags exist, you can filter and report on them like any other outcome.
{% endstep %}
{% endstepper %}

## Reading charts and tables it gives you

When Agent Studio returns a chart, it also writes key takeaways beneath it. Read both — the takeaways often point you to the next useful question ("VOICEB.AI - WEB is 57% of calls but the shortest duration" is an invitation to drill in). If a chart isn't the format you wanted, just say "show that as a funnel instead" and it re-renders.

<figure><img src="/files/3HARPvKuY3sUXwyYeRkb" alt=""><figcaption></figcaption></figure>

## Guardrails: this is live data

Agent Studio can make **real changes** — creating agents, assigning DDIs, editing scripts, setting up tags. Two habits keep you safe:

* **Confirm before you commit.** The assistant proposes changes; read them before approving. Treat "create" and "assign" prompts with the same care you'd give the dashboard itself.
* **Verify numbers that drive decisions.** It's an AI assistant and can occasionally be wrong. Before you act on a figure in a board update or a client email, cross-check it in the relevant dashboard section.

{% hint style="info" %}
**Beta means evolving.** Capabilities and models will change. If something behaves unexpectedly, start a fresh conversation and rephrase — and check the [reference page](https://docs.voiceb.ai/agent-studio) for the current feature set.
{% endhint %}

## Quick reference: prompt patterns

```
Analytics    → "[metric] per [campaign/agent] over [range], as a [chart], sorted by [x]."
Drill-down   → "Now focus on [name] and show me [detail]."
Calls        → "List [N] recent [outcome] calls for [campaign] with transcripts."
Build        → "Help me create a [type] agent for [use case], step by step."
Web          → "Read [URL] and turn it into [script notes / knowledge]."
Export       → "Export [that result] to Excel."
Summarize    → "Summarize this conversation's findings in [N] bullets."
```

Start broad, drill down, confirm before committing, and verify what matters. That's the whole game.


# VoiceB Prompt & Script Guide

How to build an effective VoiceB Agent prompt + script

You no longer write prompts and scripts by hand. **Agent Studio** builds and refines your agent's configuration conversationally — you describe your business and your goals, and it generates the four core blocks that power every agent: **Goal, Script, Objections, and Guardrails**.

This guide shows you how to work with Agent Studio, and explains the best practices behind well-structured blocks so you can review, refine, and optimize what it produces — and get more natural, personalized, conversion-oriented conversations.

***

#### How it works in Agent Studio

<figure><img src="/files/0Nr25jsKudYC9NLSrTX5" alt=""><figcaption></figcaption></figure>

Start a session and tell Agent Studio:

* What your company sells
* Who your target caller is
* What a successful call looks like (a sale, a qualified lead, a booked appointment)
* Any key objections your team regularly handles

You can also give it a URL to build the knowledge base automatically (see Automate Your Knowledge Base with Web Scraping).

Agent Studio generates the agent's full configuration across the four blocks below. You review and edit each block independently — or simply tell it what to change in plain language. The agent improves through conversation, not by hand-editing raw prompts.

{% hint style="info" %}
**Knowledge vs. behavior:** Keep your reference knowledge (facts, pricing, FAQs) consolidated in a **single knowledge block** for easy maintenance. Your agent's *behavior* — persona, flow, objections, and boundaries — is structured across the four blocks below.
{% endhint %}

<figure><img src="/files/OHL3QsF9GpL5LphkDa8r" alt=""><figcaption></figcaption></figure>

***

#### The Four Core Blocks

Separating instructions by responsibility makes it easier for the AI to retrieve the right information — and simpler for you to update.

```
1. GOAL       → The mission, persona, and core objective
2. SCRIPT     → The conversational flow, stages, and phrasing
3. OBJECTIONS → Rebuttals, FAQs, and soft-landing formulas
4. GUARDRAILS → Hard boundaries, do-nots, and fallback behaviors
5. KNOWLEDGE BASE
6. ...
```

<figure><img src="/files/UGn0f2ASThobAiFWCzyG" alt=""><figcaption></figcaption></figure>

**🎯 Block 1 — Goal (the "who" and "why")**

The Goal block establishes the agent's identity, tone, and what a successful call looks like.

* **Define a clear persona:** role, company name, and emotional tone.\
  *e.g. "You are Sarah, a warm, professional customer success specialist at GreenEnergy Corp."*
* **Establish a single target outcome:** state exactly what the agent is trying to achieve.\
  *e.g. "Your primary goal is to qualify the lead and schedule a 15-minute consultation. Do not try to sell the product directly on this call."*
* **Acknowledge the starting state:** briefly mention how the lead was generated.\
  *e.g. "The contact requested a quote on our website yesterday."*

**💬 Block 2 — Script (the "how")**

The Script is a conversational map, not a rigid, word-for-word monologue.

* **Write for the ear, not the eye:** short, conversational sentences that are easy to pronounce. Avoid jargon.
* **Open warmly:** the first line the agent says should be short and welcoming.\
  *e.g. "Hi, I'm \[Agent Name], your AI assistant from \[Company]. Which service are you interested in today?"*
* **Use phase-based milestones:**
  * **Phase 1 — Hook & Permission:** acknowledge who they are, state why you're calling, ask for a moment.
  * **Phase 2 — Discovery & Qualification:** ask 1–2 open-ended questions instead of lecturing.
  * **Phase 3 — Pitch / Value Prop:** keep it under 3 sentences.
  * **Phase 4 — Close / Call-to-Action:** make a concrete, low-friction proposal.
* **Build in micro-agreements:** check in often (*"Does that make sense?"*) to keep the caller engaged.
* **Map the main flows / branches:**
  * *Interested – doesn't know what they want* → ask discovery questions
  * *Interested – knows what they want* → collect required data and move to close
  * *Already a customer* → redirect politely to customer service
  * *Wrong number* → end the call gracefully
  * *Wants a human agent* → collect data, then prepare transfer
* **Validation rules:** check email format, phone length, IDs where required. Normalize values (*"3 coma 45" → "3.45 kW"*).
* **Silence management:**
  * 1.5s → "Do you need me to repeat that?"
  * 3s → "Are you still there?"
  * 8s → end politely
* **Closing lines:**
  * Valid lead → "Thanks, an agent will contact you soon."
  * Non-lead → "Thanks for your time, call us again if you need."
* **Tone & voice:** \~150 words/minute, warm, professional, natural; use synonyms to avoid repetition.

{% hint style="info" %}
**Agent Instructions (DO NOT READ ALOUD):** Place non-spoken directives at the end of the Script block, in brackets, so the AI doesn't read them out.

```
Agent Instructions (DO NOT READ ALOUD)

- Speak slowly and clearly.
- Avoid long pauses or unnecessary delays.
- Do not ask for personal data outside of the online form.
- Use tone variation to sound natural and human.
- Always acknowledge the customer's responses before moving forward.
- Follow the sign-up steps strictly in sequence.
```

{% endhint %}

**🛡️ Block 3 — Objections (the "what if")**

Train your agent to handle pushback gracefully rather than ignoring it or sounding defensive.

Use the **Acknowledge → Pivot → Ask (APA)** formula:

* **Acknowledge:** validate the concern — *"I completely understand budget is top of mind right now\..."*
* **Pivot:** reframe the value — *"...which is exactly why we focus on cutting your operational costs by 30%..."*
* **Ask:** hand the microphone back — *"...how much are you currently spending on manual data entry?"*

List the top 3–5 objections with ready responses, for example:

* *"I don't have time right now."*
* *"Is this a robot?"* → "Yes, I'm an AI assistant helping the team today — and I'm fully equipped to get this sorted for you. Can I ask..."
* *"We already have a provider."*
* *"Just email me the information."*
* *Refuses privacy consent* → "I understand. Unfortunately I can't continue without your consent — if you change your mind, feel free to call us again."

**🚧 Block 4 — Guardrails (the "do nots")**

Guardrails keep your agent compliant, safe, and professional.

* **Define no-go zones:** topics the agent must never discuss.\
  *e.g. "Never promise specific pricing percentages; only state that pricing is customized."*
* **Handle interruptions:** instruct the AI to yield instantly if the human speaks over it.
* **Set fallback behaviors:** give the AI a line for when it doesn't know the answer.\
  *e.g. "That's a great technical question — let me note it so our engineer can address it in your scheduled demo."*
* **Identify transfer triggers:** define when to hand off to a human.\
  *e.g. "If the caller becomes angry, demands a manager, or uses profanity, transfer the call immediately."*

***

#### Conversation Design Golden Rules

* **Keep turns short:** no single agent response should exceed \~30–40 words (\~15 seconds). Long monologues make callers tune out or interrupt.
* **Format with Markdown inside blocks:** bolding, bullets, and numbered lists. The underlying model reads structured Markdown far better than walls of plain text.
* **One question at a time:** never ask double-barreled questions. Ask, wait, then ask the next.

***

#### Click to Call Variables

Pass customer data from your form directly into the dialogue to personalize the conversation in real time.

* Greeting → "Hi {{name}}, thanks for contacting us about {{product}}. How can I help you today?"
* In-flow → "I see you're interested in {{product}}. Let's review the details together."

***

{% hint style="danger" %}
**Alert – Character Limit**\
For best performance, the combined content of your **Goal, Script, Objections, and Guardrails** blocks should stay **below 21,000 characters**.\
\- Beyond this limit, the AI may behave unpredictably, pause, or fail to load.\
\- Keep blocks concise, avoid redundant text, and move long instructions or legal disclaimers into **Agent Instructions** at the end of the Script (in brackets).\
\- Always test the agent after adding content to confirm it runs smoothly.
{% endhint %}

***

#### Data Collection

Define what information the agent must request and validate. Typical fields:

* Name
* Email
* Phone
* Address (Street, Number, City, ZIP)
* Consent (Yes/No)
* Appointment date/time

**Create a Data Collection Schema**

{% hint style="info" %}
**Tips:** Repeat each piece of data back in question form to confirm it. Normalize values. Don't collect data you won't use.
{% endhint %}

***

#### Integrations

Connect your agent with **Salesforce**, **HubSpot**, or custom **Webhooks** to send collected data in real time.

{% hint style="info" %}
**Tips:** Always test the webhook with a sample call. Map your CRM fields to match the Data Collection schema.
{% endhint %}

***

#### Final Checklist

✅ Built and refined in Agent Studio\
✅ Clear Goal — persona + single target outcome\
✅ Script with phase-based milestones and modular flows\
✅ Short, warm opening line\
✅ Objections handled with the APA formula\
✅ Guardrails: no-go zones, interruptions, fallbacks, transfer triggers\
✅ Data Collection schema defined\
✅ Validation + silence-management rules applied\
✅ Standard closing messages\
✅ Click to Call Variables used for personalization\
✅ Turns short (30–40 words), one question at a time\
✅ Combined blocks under 21,000 characters\
✅ **Agent Instructions placed at the end of Script (DO NOT READ ALOUD)**

***

<figure><img src="/files/CQ3lUNQIUOTcS8fD6uyf" alt=""><figcaption></figcaption></figure>

{% hint style="success" %}
**Build or refine with Agent Studio:** Ask Agent Studio to review your live agent's blocks and suggest optimizations against these guidelines — or build a brand-new agent from scratch. It's all done in natural language.
{% endhint %}


# How to Create an Effective VoiceB Prompt: An Example from NexTel

When launching a VoiceB.ai agent, one of the most important steps is designing the **prompt** — the set of instructions that define how your AI agent introduces itself, interacts with users, and guides them through your process.

In this article, we’ll walk you through an ideal prompt structure using a real-world telco example from NexTel, including the **initial message**, **core script**, **step-by-step guidance**, and **objection handling**.

Let’s dive in.

***

### 🔹 1. Initial Message

The **initial message** is the very first thing your AI agent says. It sets the tone and starts the conversation.

> **Best practice:** Keep it short, natural, and welcoming. Prompt the user for their name to personalize the conversation.

**Example:**

```
Hello, my name is Jessica, I am your AI agent from NexTel. Thank you for contacting us. What is your name so I can address you properly?
```

***

### 🧠 2. Main Script: Conversation Flow + Product Pitch

The **script** is the backbone of your prompt. It explains your offering clearly and guides the user through the decision-making process.

In this case, NexTel offers various internet and mobile plans. The script leads the user through plan selection and signup, tailored to their preferences.

#### Example: Plan Selection

```
Hello, [user's name].

Let me help you choose the perfect plan for your needs, and then I will guide you through the sign-up process. If you're not on the website, go to NexTel's website, where you will see the available plans.

Are you looking for speed or price?
```

The agent then explains the plans based on user input:

* **Mobile Unlimited** – $15/mo: Unlimited calls, 50GB 5G. Best for mobile-only needs.
* **Fiber 600Mb + 25GB Mobile** – $35/mo: Great home+mobile combo.
* **Fiber 1Gb + Mobile Unlimited** – $40/mo: For high-speed, high-usage users.
* **Fiber 600Mb + 2 Mobiles + TV** – $47/mo: Ideal for families.

***

### 🛠️ 3. Step-by-Step Sign-up Assistance

Once the user chooses a plan, the agent guides them through each step of the signup form. This format keeps users engaged and reduces drop-off.

#### Example:

**Step 1:** Check coverage\
"Enter your address and phone number to check if our service is available in your area. Did it show that there is coverage?"

**Step 2:** Fill in personal information\
"Complete your personal details: name, ID, date of birth, email, phone. Have you completed it?"

**Step 3:** Customize your plan\
"Choose your fiber speed, number of mobile lines, and TV options. Have you completed it?"

**Step 4:** Schedule installation\
"Select a delivery or installation time that works for you. Have you completed it?"

**Step 5:** Payment\
"Enter your IBAN and billing address. Review and accept terms. Have you completed it?"

Once the process is complete, the agent confirms the order and outlines next steps.

***

### 💬 4. Objection Handling

Objections are part of any sales process. Prepare your agent with short, accurate responses to common questions.

#### Example Objections:

* **How can I check coverage?**\
  "Just provide your address and phone number, and I’ll check coverage instantly."
* **What’s the installation process like?**\
  "You’ll choose a date and time for installation or delivery. Our team will confirm all the details with you."
* **What payment methods do you accept?**\
  "We use secure direct debit. Just enter your IBAN and billing info during checkout."

> **Pro tip:** Only trigger objection responses when users ask directly. Keep your AI efficient and to the point.

***

### ✅ Conclusion: A Strong Prompt Drives Better Results

A well-structured VoiceB prompt makes the difference between a confusing user experience and a clear, guided journey that converts.

To summarize, your prompt should include:

* A friendly **initial greeting**
* A clear **goal-oriented script**
* A structured, interactive **step-by-step guide**
* Smart, preloaded **objections** with natural replies

Whether you're a telco like NexTel, a bank, a retailer, or an insurance company, VoiceB helps your users get things done faster — with no human needed (unless you want one later).

***

📣 **Want to launch your own VoiceB agent?**\
Book a demo or talk to our team and we’ll help you create a prompt that works — just like NexTel.


# Automate Your Knowledge Base with Web Scraping (BETA)

A great agent starts with great knowledge — and now **Agent Studio builds that knowledge for you** directly from your website. Instead of writing prompts or uploading documents, you give Agent Studio a URL, and it scans, structures, and adds your public content to the agent's knowledge base automatically. You can even scrape your website to **create a brand-new agent from scratch** — Agent Studio uses the content to generate the goal, script, and objections for you.

<figure><img src="/files/ozfmRSCcjD4htQRX9dCj" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
This capability used to be a separate beta that generated a `.txt` file on request. It's now built directly into **Agent Studio** — self-serve, no email or manual file handling required.
{% endhint %}

#### 🔍 What it does

When you give Agent Studio a URL while building or editing an agent, it:

* Crawls the public pages linked from that address, including:
  * Product pages
  * Pricing
  * Help center / FAQs
  * Blog posts
  * Documentation and resources
* Extracts the useful content: headlines, paragraphs, lists, answers, calls-to-action
* Cleans and structures it into the agent's knowledge base, formatted for optimal AI comprehension

No prompts to write from scratch, no support topics to enter by hand.

***

#### 🧠 Why it matters

Your agent is only as good as the knowledge it has. Building the knowledge base from your website means your agent:

✅ **Covers all your public info** — from product specs to FAQs, it draws on your published content.\
✅ **Speaks your language** — it absorbs your tone, brand, and messaging from the way you present content online.\
✅ **Saves time** — no copying and pasting; the knowledge base builds itself from a URL.

***

#### 📂 What VoiceB extracts

Here's a simplified example of how VoiceB structures scraped content:

```
Section: Plans

The Mobile Unlimited plan costs $15/month and includes unlimited calls and 50GB at 5G speeds.

The Fiber 600Mb plan includes internet and mobile for $35/month.

Section: FAQs

Q: How long does installation take?
A: Installation usually takes 1–3 days depending on availability.

Q: Can I change my plan later?
A: Yes, all plans are flexible and can be upgraded at any time.
```

> This structure is optimized for AI processing, ensuring accurate, human-like responses from your agent. You can review and refine everything in Agent Studio before going live.

***

#### 🗣️ Getting started

You can use web scraping two ways:

* **Create a new agent from scratch** — give Agent Studio your website URL and it will generate the full agent (goal, script, objections, and knowledge base) from your content.
* **Enrich an existing agent** — add a URL to an agent you've already built to expand its knowledge base.

Steps:

1. Open **Agent Studio** and start building a new agent (or edit an existing one).
2. Give it your website URL (or a specific page — pricing, FAQ, product).
3. Agent Studio scans the content and adds it to the knowledge base.
4. Review, edit, or expand the result, then go live.

{% hint style="success" %}
**Recommendation:** Ask Agent Studio to consolidate the entire knowledge base into a **single script block**. Keeping it in one block makes the agent far easier to maintain and optimize — you update one place instead of hunting across multiple blocks.
{% endhint %}

This is especially valuable for SaaS products, eCommerce catalogs, and telcos, banks, and insurers with dynamic pricing and FAQs.

> New to the platform? Start with [Getting Started in 20 Minutes →](https://docs.voiceb.ai/product-guides/getting-started-with-voiceb.ai-in-20-minutes)


# June 2026 — VoiceB.ai Product Release

By Alex Bisbe. July 6th 2026.

***

June moved VoiceB **from engineering-operated to operator-owned**.

Where May closed the contact gap — making sure every submitted lead actually gets reached — June changes who runs the system. **Agent Studio is now available to admin users**: the full agent lifecycle — create, configure, test, publish, and maintain — moves out of engineering and into the hands of the people who own the campaigns.

Until now, standing up a new agent or changing a live one meant an engineering ticket. That worked when deployments were counted on one hand. It does not scale to a platform signing multiple new enterprise logos every month.

The objective: **make agent creation and maintenance a configuration task, not an engineering project.**

***

**🚀 Key Highlights**

* **Agent Studio for admin users** — create, configure, and maintain autonomous agents end-to-end from the console, with no engineering involvement required
* **Structured script editing with version control** — typed Script Blocks, full change history, diff, and one-click restore on every edit
* **Routing and transfer as configuration** — human-handoff rules managed as tool settings, kept separate from the agent script
* **Callback retry hardening** — reliability and scheduling fixes across the Click-to-Call Retry Policy released in May

***

**✨ Agent Studio — The Full Agent Lifecycle, Now in Admin Hands**

Agent Studio is the operating console for autonomous agents. With this release, users with the admin role run the complete lifecycle without touching code or opening a ticket:

* **Create agents** — spin up a new agent from scratch, defining its goal, role, and conversation objectives through structured Script Blocks
* **Edit live scripts safely** — every change is auto-versioned with full history, diff, and restore, so a production agent can be edited with the same rigor as production code
* **Configure outcomes and tags** — define the outcome taxonomy and custom tags the Outcome Engine applies to every call
* **Manage routing and transfers** — transfer-to-human logic is configured as a tool, independent of the script, so routing changes never require rewriting the agent
* **Connect knowledge** — attach and update the agent's knowledge base, including web-scraped content
* **Publish and monitor** — push changes live and follow performance in the Campaign Dashboard, per outcome, tag, and agent

[Agent Studio →](https://docs.voiceb.ai/agent-studio)

<figure><img src="/files/m7DbbLzUPwa5m4jWanVw" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/BH3rguEEjLxp2srjDy8N" alt=""><figcaption></figcaption></figure>

*Images: Agent Studio — agent overview and creation view*

**What changes in practice**

| Before June                              | With Agent Studio for admins                                |
| ---------------------------------------- | ----------------------------------------------------------- |
| New agent = engineering ticket           | New agent = admin configures it in the console              |
| Script change = request, wait, deploy    | Script change = edit, diff, publish — and restore if needed |
| Routing change = code change             | Routing change = tool configuration                         |
| Onboarding gated by engineering capacity | Onboarding gated only by campaign readiness                 |

The result is speed exactly where it matters most: client onboarding and iteration on live agents. Changes that took days now take minutes — with a full audit trail underneath every edit.

*\[Screenshot: Script Blocks editor with version history panel]*

**Who gets it**

Agent Studio ships behind role-based access. Script editing, outcome configuration, routing, and publishing require the **admin** role; analyst and viewer roles keep read access to dashboards, transcripts, and recordings. Every published change is logged with author, timestamp, and version for full auditability.

***

**🔧 Improvements & Fixes — Callback Retries**

May introduced the Click-to-Call Retry Policy. June hardens it with fixes from the first weeks of production traffic:

* **Business-hours edge cases** — retries scheduled exactly on a window boundary now snap forward correctly, including across weekends and timezone offsets
* **Retry-chain reliability** — transient telephony failures no longer interrupt an active retry sequence; the chain resumes with the correct attempt index
* **Accurate retry indexing** — the `retry n/m` index in the Campaign Dashboard is now consistent across all failure outcomes
* **Duplicate-attempt prevention** — safeguards against double-dialing the same lead when multiple triggers fire in a short window

***

**Final Takeaway**

June was about **ownership**.

April made VoiceB easier to find, evaluate, and deploy. May made sure that once a lead raises its hand, it gets answered. June puts the operation itself in the operator's hands — autonomous agents are now created, changed, and maintained by the people running the campaigns, with version control and audit trails underneath every edit.

The engineering bottleneck on agent creation and maintenance is gone. What remains is the work that matters: making every conversation convert.


# May 2026 — VoiceB.ai Product Release

By Alex Bisbe. June 9th 2026

***

May moved VoiceB **from reaching the buyer to reaching the lead**.

Where April brought the product to the surface — deployable, priceable, understandable from the outside in — May closes the gap that sits one step further down the funnel: the lead who was acquired, paid for, and submitted, but didn't pick up the first time.

A single dial attempt reaches only a fraction of submitted leads. The new **Click-to-Call Retry Policy** turns that single attempt into a disciplined, compliant retry sequence — controlled by the operator, bounded by local calling rules, and fully visible in the dashboard.<br>

The objective: **maximize contact rate without violating local outbound calling rules.**

***

**🚀 Key Highlights**

* **Click-to-Call Retry Policy** — automatic re-dialing of submitted leads under operator-defined conditions: how many attempts, how long between them, which failure outcomes trigger a retry, and which hours and days are allowed
* **Exponential and custom backoff** — fast retries while intent is warm, spacing out as the lead cools, or a precise manual cadence
* **Business-hours enforcement** — retries snap forward to the next valid window so calls never land outside allowed times
* **Per-retry dashboard visibility** — every attempt logged as its own call with its retry index, filterable per lead and per outcome

***

**🔁 Click-to-Call Retry Policy — Every Lead Contacted, Inside the Rules**

Click-to-Call places an outbound callback the moment a lead is submitted — a form fill, a widget request, a CRM trigger. The first attempt connects most leads, but a meaningful share won't pick up the first time: they're driving, in a meeting, or away from their phone. Every one of those is paid-for demand that, until now, could go uncontacted.

[Click-to-Call Retry Policy →](https://docs.voiceb.ai/click-to-call-retry-policy)

The numbers make the case. A single dial attempt typically reaches 30–50% of leads, while three to four well-spaced attempts lift contact rates above 70%. That difference flows directly into conversion rate and CPA — it is one of the highest-leverage changes available on a live lead source, because the demand is already bought and paid for.

<div data-with-frame="true"><figure><img src="/files/uDMKkhAOHEMrnwAS8E4S" alt=""><figcaption></figcaption></figure></div>

**Operator-controlled, by agent**

The Retry Policy is configured per agent on the agent settings page. A single toggle activates it; when disabled, Click-to-Call makes exactly one attempt per lead and stops. Operators control four things:

* **Max attempts** — total dial attempts per lead, including the original call. Recommended 3–5 for most B2C use cases.
* **Retry on** — which failure outcomes trigger a retry: `no answered`, `busy`, `voicemail`, `failed`, `dropped`, `unreachable`. A typical setup retries on `no answered`, `busy`, and `voicemail`, and skips the number-level failures more attempts won't solve.
* **Backoff strategy** — how long to wait between attempts.
* **Business hours** — the window inside which retries are allowed to land.

**Two backoff strategies**

* **Exponential** — the wait grows after each attempt by a fixed factor, up to a maximum. Defaults of 15 min initial / factor 2 / 1440 min max give hot leads a fast retry while intent is warm, then space attempts out as the lead cools. The right default for most lead-gen flows.
* **Custom sequence** — define each delay manually as a comma-separated list of minutes (e.g. `15, 60, 240, 1440`). The system uses each value in order and reuses the last once the list is exhausted. Built for a precise cadence — retry quickly within the first hour, then once that evening, then once the next morning.

**Compliance by default**

Outbound calling outside business hours is illegal or restricted in most markets. The Business Hours block ensures every retry lands inside an allowed window:

* **Respect business hours** — any retry scheduled outside the window is moved forward to the next valid slot.
* **Start / end hour** — the daily window, in the local timezone of the campaign, not the browser.
* **Weekdays only** — retries scheduled for Saturday or Sunday are pushed to the following Monday.

A Spain campaign set to 9:00–20:00, weekdays only, will never place a retry on a Saturday morning or a Tuesday at 8:30 — even when the backoff math says it should. The controls let operators be aggressive on contact rate while staying inside the rules.

**Visible per attempt**

Every retry is recorded as its own call in the Campaign Dashboard with its retry index alongside the outcome — `retry 1/2`, `retry 2/2`, `retry 1/3`. Filter by phone number to view the full retry chain for any single lead, or by outcome to spot patterns: leads that consistently hit voicemail across all attempts point to a bad data source, not an agent problem. The same diagnose-before-you-blame discipline VoiceB applies everywhere, now extended to contact attempts.

<div data-with-frame="true"><figure><img src="/files/b3myQEBk2rO2UH526jAN" alt="" width="160"><figcaption></figcaption></figure></div>

**Recommended starting points**

| Use case                                                 | Max attempts | Retry on                     | Backoff                            | Business hours        |
| -------------------------------------------------------- | ------------ | ---------------------------- | ---------------------------------- | --------------------- |
| Standard B2C lead-gen (ES, UK, EU)                       | 4            | no answered, busy, voicemail | Exponential — 15 / ×2 / 1440       | 9:00–20:00, weekdays  |
| High-intent / hot leads (cart abandon, recent form fill) | 3            | no answered, busy            | Custom — `5, 30, 180`              | 9:00–21:00, weekdays  |
| Low-volume / high-value (insurance, financial)           | 5            | no answered, busy, voicemail | Custom — `15, 60, 240, 1440, 2880` | 10:00–19:00, weekdays |

***

**Final Takeaway**

May was about **recovering the demand that was already paid for**.

The Retry Policy doesn't acquire a single new lead. It makes sure the leads already submitted actually get reached — turning a 30–50% single-attempt contact rate into 70%+ across a disciplined, compliant sequence the operator controls and can audit per attempt.

March put VoiceB deeper inside the enterprise. April made it easier to find, evaluate, and deploy. May makes sure that once a lead raises its hand, it gets answered — every time the rules allow.


# April 2026 — VoiceB.ai Product Release

By Alex Bisbe. May 12th 2026

***

April moved VoiceB **from infrastructure to interface**.

Where March went deep into enterprise telephony and CRM integrations, April brings the product to the surface — the way buyers see it, deploy it, and price it.

The new Web Widget makes VOICEB deployable on any site with a single line of JavaScript and a full visual customization layer. The new Pricing Guide makes the commercial model explicit, with three plans and clean separation between outcomes, telephony, and outbound charges. And a new long-form article — *"What Happens When Lead Capacity Has No Ceiling"* — frames the strategic shift that autonomous inbound enables.

The objective: **make the platform deployable, priceable, and understandable from the outside in.**

***

**🚀 Key Highlights**

* **Web Widget Configuration** — one-line JavaScript deployment, multi-agent and multi-language at the campaign level, full visual customization, built-in compliance, real-time preview
* **Pricing Guide** — three plans (Starter, Enterprise, Corporate Custom), outcome-based ASVA pricing, transparent DDI and outbound call costs, all autonomy levels accessible on every plan
* **New Article:&#x20;*****What Happens When Lead Capacity Has No Ceiling*** — strategic narrative on how unlimited autonomous capacity restructures the entire demand engine

***

**🌐 Web Widget — One-Line Deployment for SuperVoice**

The Web Widget is the primary way to deploy AI voice agents on any website or landing page. It connects visitors directly to SuperVoice agents via WebRTC — no phone number required, no app to install, no friction.

[Widget Configuration Guide →](https://docs.voiceb.ai/widget-configuration-guide)

Installation is one line of JavaScript with a unique widget ID. From the Widget Configuration page, operators can fully preview and test the deployment before going live — the preview updates in real time as settings change, and a built-in test call validates agent behavior end-to-end.

<figure><img src="/files/DptqHxsV1he3rSdrRkMa" alt=""><figcaption></figcaption></figure>

**Multi-agent, multi-language at the campaign level**

A single widget deployment can serve multiple agents across multiple languages. Operators select which agents the widget exposes, set the display order per language, and the widget automatically renders a language selector for visitors. One embed, multiple agents, multiple markets — ideal for multi-language websites and landing pages serving diverse audiences from a single deployment.

**Two display variants**

*

```
<div data-with-frame="true"><figure><img src="/files/Lllz05utNUXhq9dpfQr9" alt="" width="159"><figcaption></figcaption></figure></div>
```

* **Bubble** — minimal floating bubble in the corner. Maximum discretion. Recommended for product pages, blogs, and content where the widget supplements the experience.
* **Compact** — a slightly larger floating element with a welcome message. More visible and explicit. Recommended for pricing pages, plan selectors, and checkout flows where voice engagement should be front-and-center.

**Full visual customization**

Every visual element is configurable to match brand identity:

* **Surface palette** — background, border, foreground, muted, hover states
* **Brand palette** — primary, hover, active, contrast
* **Metrics** — padding, corner radius, font size, icon size, exposed as individual numeric controls
* **Per-language text overrides** — title, greeting, agent card heading, description, and CTA button label

**Compliance built in**

Privacy Policy and Terms and Conditions URLs are first-class fields. When set, an explicit consent checkbox appears in the widget. With "Require terms acceptance" enabled, visitors must consent before initiating a call. Consent records — including the document content, up to 50 MB per document — are stored in the database, giving operators a complete audit trail of what each visitor consented to and which version of the documents was in effect. Built for GDPR, LOPDGDD, and equivalent regulatory frameworks.

**Closed-loop attribution**

Every widget load is an impression. Every interaction is a tracked click. UTM parameters from the page URL (`utm_source`, `utm_medium`, `utm_campaign`, `utm_content`, `utm_term`) are captured automatically and connect every voice conversation back to the marketing campaign that drove the visit — closing the attribution loop from ad spend to voice conversion at the conversation level.

***

**💶 Pricing Guide — Outcomes, Telephony, and Outbound, Cleanly Separated**

VoiceB pricing is outcome-based. The new Pricing Guide makes the commercial model fully explicit, with three plans available and a deliberate separation between value delivered and infrastructure consumed.

[Pricing Guide →](https://docs.voiceb.ai/pricing-guide)

**Three plans, all autonomy levels included**

| Outcome                                      | Starter | Enterprise | Corporate Custom |
| -------------------------------------------- | ------- | ---------- | ---------------- |
| Unqualified / invalid calls                  | €0      | €0         | €0               |
| Qualified call transferred to human          | €2.00   | €1.50      | Custom           |
| Pre-sale (final step before payment)         | €9.00   | €7.00      | Custom           |
| Completed sale with integration confirmation | €20.00  | €15.00     | Custom           |
| Minimum monthly commitment                   | €700    | €2,000     | Volume-based     |

All plans include access to all four autonomy levels — L1 (Qualification), L2 (Intelligent Routing), L3 (Assisted Autonomous Sales), L4 (Full Autonomous Sales). Customers typically start at L1–L2 and scale over time without switching plans. Pricing scales with the outcome produced, not the level configured.

**Three components, billed independently**

* **ASVA outcome-based pricing** — what gets paid for value delivered
* **DDI / DID telephony costs** — monthly fee per inbound phone number, billed per region (USA $1.50, UK £2.50, Spain €5.50)
* **Outbound call charges** — included by default for North America, EU, UK, Oceania, and major Latin America markets

This separation is deliberate. Telephony costs vary by geography and carrier — bundling them into outcome pricing would create hidden margins that distort unit economics at scale. Finance teams need predictability and auditability on each component independently.

**What VoiceB does not charge for**

Seats. Agents. Minutes within included markets. AI tokens. Environments or sandboxes. LLM orchestration overhead. The only charges are outcomes, DDI numbers, and outbound calls outside included markets.

***

**📖 New Article —&#x20;*****What Happens When Lead Capacity Has No Ceiling***

A new long-form piece on the strategic shift that autonomous inbound enables — written for CMOs, marketing leaders, and revenue operators.

[Read the article →](https://docs.voiceb.ai/blog/what-happens-when-lead-capacity-has-no-ceiling)

The thesis: most B2C companies are not capacity-constrained by their market — they are capacity-constrained by their contact center. That invisible ceiling shapes every upstream decision: which leads they buy, which channels they invest in, how they design the top of the funnel.

When the ceiling disappears, five structural shifts compound:

1. **Cheaper leads become economical**, because filtering is free
2. **New lead sources open up** — off-peak campaigns, broader targeting, partner traffic, organic, voice-on-landing-page
3. **Media-mix dependency breaks** — the operator controls the volume dial, not the agency
4. **No human threshold to overflow against** — acquisition capacity is decoupled from headcount
5. **TOFU stops being a filter and becomes a capture layer** — qualification moves from the form to the conversation

The article walks through each shift with production data from live enterprise deployments and closes on the compounding effect: more leads → better data → lower CPA → more budget → more leads. This is a market capture loop, not a cost-reduction story.

This is the strategic narrative that frames every other piece of the product — the widget, the pricing, the integrations, the SIP trunk. It is what the buyer is actually buying.

***

**Final Takeaway**

April was about **closing the gap between the platform and the buyer**.

The Widget makes deployment a one-line operation. The Pricing Guide makes commercial evaluation a self-serve exercise. The new article frames the strategic shift that explains why any of it matters.

Every release in March made VoiceB deeper inside the enterprise. April makes it easier to find, evaluate, deploy, and understand — from the outside in.


# March 2026 — VoiceB.ai Product Release

By Alex Bisbe. April 8th, 2026.

***

March focused on **enterprise connectivity and CRM depth**.

This release brings VoiceB deeper into enterprise telephony infrastructure and revenue systems. \
\
SIP Trunk moves VoiceB inside carrier-grade contact centers without disrupting existing infrastructure. The Salesforce integration replaces a simple webhook with a full CRM orchestration layer. \
\
And Trigger Conditions — now covering both HubSpot and Salesforce — give operators precise control over what flows into their CRM and when. \
\
The objective: **closer to the revenue system, further from the generic voice bot**.

***

**🚀 Key Highlights**

* **SIP Trunk** — private call routing with pre-call, in-call, and post-call data injection
* **Salesforce Integration** — CRM-native orchestration across any object, with deduplication and field mapping
* **Trigger Conditions for HubSpot and Salesforce** — outcome-based firing logic for both integrations

***

**📡 SIP Trunk — Private Call Routing for Enterprise Voice AI**

VoiceB can now connect directly to any enterprise PBX or Session Border Controller (SBC) via a dedicated, fully managed SIP trunk.

This is the architecture that Vodafone España runs in production. Instead of porting numbers or routing through shared telephony infrastructure, organizations keep full ownership of their carrier contracts and phone numbers — and connect them to VoiceB AI agents at the SIP layer.

[SIP Trunk documentation →](https://docs.voiceb.ai/sip-trunk-private-call-routing-for-enterprise-voice-ai)

**What this enables:**

* **Pre-call data injection** — Before the AI agent picks up, VoiceB fires a webhook to your CRM or backend. The response is injected into the agent's prompt, so every conversation starts with full caller context: account status, lead tier, prior interactions. The caller hears "hello" and the agent already knows who they are.
* **In-call SIP header routing** — Custom SIP headers attached by your PBX are captured and used to route calls to the right agent in real time. One trunk, multiple agents, zero PBX reconfiguration.
* **Post-call transfer with context** — When the agent hands off to a human, VoiceB fires a pre-transfer webhook with the full conversation data — transcript, intent, extracted fields — before the call arrives at the human agent's desk.

**Configuration is fully no-code from the Agent Studio dashboard:**

* IP Access Control — whitelist your SBC IP addresses
* SIP header capture table — define which headers to capture, their direction, and where to forward them
* Routing rules engine — evaluate headers and webhook fields to select the right agent (first-match-wins)
* Call Tester — dry-run validation mode and live test call mode, both from the dashboard

**Competitive position:** No other AI voice platform offers pre-call webhook templating with `{{placeholder}}` SIP header tokens, a routing rules engine native to the AI layer, or a pre-transfer webhook with HMAC-SHA256 signing. This is carrier-grade infrastructure, not a telephony workaround.

<div data-with-frame="true"><figure><img src="/files/kEwvJlM1Gv1Y2T9rhKDb" alt=""><figcaption></figcaption></figure></div>

***

**🔵 Salesforce Integration — CRM-Native Orchestration Layer**

The previous Salesforce connector created a Lead record and stopped there. The new integration makes VoiceB a full CRM automation layer for Salesforce.

[Salesforce integration documentation →](https://docs.voiceb.ai/how-to-connect-your-voiceb.ai-agent-with-salesforce)

**What changed:**

The integration is no longer limited to fixed Lead creation or simple webhook behavior. VoiceB now acts as a CRM-native orchestration layer.

You control:

* **Which Salesforce objects receive data** — Leads, Contacts, Opportunities, Cases, or any custom object. Not every call should create a Lead. A closed sale should create an Opportunity. A customer care interaction should update a Case.
* **How objects relate** — Define relationships between objects. An Opportunity can be linked to an existing Contact without creating a duplicate.
* **Which values are constant** — Fixed field values (pipeline owner, campaign source, agent ID) are injected at sync time regardless of call data. No extra middleware required.
* **How duplicates are handled** — Matching rules prevent duplicate Contact or Lead records when the same caller interacts more than once.

This allows enterprise-grade Salesforce automation driven entirely by call outcomes — without additional ETL tools, Zapier layers, or custom development.

**Why this matters at enterprise scale:** an energy client's architecture writes autonomous sales directly as *ventas brutas* to Salesforce, bypassing the human review step entirely. That flow requires an integration layer that can address any Salesforce object with structured data, not just create a Lead.

***

**🔗 Trigger Conditions for HubSpot and Salesforce**

First introduced for HubSpot in the February release, Trigger Conditions now apply to both HubSpot and Salesforce integrations.

[HubSpot integration documentation →](https://docs.voiceb.ai/how-to-connect-your-voiceb.ai-agent-with-hubspot)

Operators choose exactly which outcomes fire a CRM sync. The full outcome set available:

| Sales         | Scheduling | Routing       | Negative       | System      |
| ------------- | ---------- | ------------- | -------------- | ----------- |
| Sold          | Reschedule | Talk to Agent | Not Interested | Pending     |
| Customer Care |            |               | Dropped        | Busy        |
| Voicemail     |            |               | Failed         | Unreachable |
| Not Found     |            |               | No Answered    |             |

**What this prevents:**

* Polluting CRM pipelines with failed calls, voicemails, or customer care interactions
* Unwanted Contact creation on low-intent or invalid calls
* Attribution noise across sales and marketing reporting
* Duplicate pipeline entries from non-sales outcomes

<figure><img src="/files/88lBged01BWPkCXxgavE" alt=""><figcaption></figcaption></figure>

**Example configuration:** A client running Salesforce for closed-deal tracking and HubSpot for lead nurturing can configure VoiceB to sync *Sold* outcomes to Salesforce only, and *Reschedule* and *Talk to Agent* outcomes to HubSpot only — from a single integration panel, with no code.

The **Valid only** shortcut selects all sales-positive outcomes in one click. **All outcomes** resets to full sync for clients who want a complete call record in their CRM.

***

**Final Takeaway**

March was about **closing the gap between AI agent and revenue system**.

SIP Trunk puts VoiceB inside enterprise telephony infrastructure without disrupting it. The Salesforce integration makes VoiceB a CRM orchestration layer, not a webhook sender. Trigger Conditions ensure only the right outcomes reach the right systems.

Every feature in this release is about one thing: making the autonomous close — end-to-end, from call to CRM — something an enterprise can run in production and trust in their reporting.


# February 2026 — VoiceB.ai Product Release

By Alex Bisbe. March 5th, 2026.

***

February focused on **agent configurability and operational control**.

This release gives operators deeper tools to build, manage, and govern AI agents — from how scripts are structured and versioned, to how outcomes are classified, to how data flows into CRMs. The objective: **faster iteration, cleaner data, and enterprise-grade agent management**.

***

#### 🚀 Key Highlights

* New **Campaign Dashboard** with search, filters, and outcome visualization
* **Script Blocks** — structured, block-based agent prompt editor
* **Script History** — full version control with diff and restore
* **CRM Trigger Conditions** — outcome-based integration firing logic
* **CRM Constants** — fixed field values injected at sync time
* **Tags Editor** — natural language tag logic with Valid/SOLD override

***

#### 📊 New Campaign Dashboard

The Campaign Dashboard has been redesigned for operational clarity.

<figure><img src="/files/92RNaxFAdxR3xjG5QiL3" alt=""><figcaption></figcaption></figure>

What's new:

* **Search and filter bar** — filter calls by phone number, outcome, agent, tag, and more
* **Switchable bar chart** — visualize call distribution by Status, Outcome, or Tags, all normalized to 100%
* Cleaner call list with consistent column display across all inbound sources

<figure><img src="/files/Fd503xDi5dwZIkDQrdv2" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/jCtmgtLPHgoQG6tPITpR" alt=""><figcaption></figcaption></figure>

This makes it significantly faster to identify patterns, isolate segments, and report on campaign performance without leaving the dashboard.

***

#### 🧱 Script Blocks

Agent prompts are now built using a **structured block system** inside Agent Studio.

Each script is composed of typed blocks — Goal, Role, Objective, Objections, and more — each independently editable with rich text formatting, headers, lists, and inline variable support `{x}`.

<figure><img src="/files/hGkyYMW47LGwQKamllj5" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/BnAfoUhnV5hfvDZo2jLC" alt=""><figcaption></figcaption></figure>

What this unlocks:

* Clearer separation of prompt concerns
* Faster editing without touching unrelated sections
* Consistent structure across all agents and campaigns
* Foundation for future prompt automation and auto-generation

Documentation: [VoiceB Prompt & Script Guide →](https://docs.voiceb.ai/product-guides/voiceb-prompt-and-script-guide)

***

#### 🕓 Script History

Every change to an agent's script is now **automatically versioned**.

The Script History panel shows a full timeline of edits, including:

* Full snapshots
* Block reorders
* Individual script edits

Each entry supports **Preview**, **See Diff**, and **Restore** — allowing operators to review exactly what changed and roll back instantly if needed.

<figure><img src="/files/wYWZ0BYpVSWyszEouPvj" alt=""><figcaption></figcaption></figure>

This is the first step toward treating agent prompts with the same rigor as production code.

***

#### 🔗 CRM Integration: Trigger Conditions

CRM integrations can now be configured to fire **only on specific outcomes**.

Operators choose which outcomes trigger a sync — Sold, Reschedule, Talk to Agent, Pending, or any combination — and exclude everything else (Customer Care, Voicemail, Failed, Dropped, etc.).

What this prevents:

* Polluting CRM pipelines with invalid or low-intent calls
* Unwanted contact creation on failed or care-related interactions
* Attribution noise in marketing and sales reporting

<figure><img src="/files/hFktK12crYD6mu9k7Buj" alt=""><figcaption></figcaption></figure>

This pairs directly with the Outcome Governance framework introduced in January.

***

#### 📌 CRM Integration: Constants

Alongside field mapping, integrations now support **Constants** — fixed HubSpot field values that are injected at sync time regardless of call data.

Use cases include:

* Tagging all contacts from a specific campaign or DDI
* Assigning a fixed owner or pipeline stage on creation
* Injecting source or channel metadata automatically

***

#### 🏷️ Tags Editor: Natural Language Logic + Valid/SOLD Override

The Tags configuration interface has been redesigned around **natural language logic**.

Each tag is defined by:

* **True if** — a plain-language description of when the tag applies (e.g. *"The user is interested in mobile plans"*)
* **False if** — the inverse condition
* **Color** — for visual identification in the dashboard
* **Valid / SOLD status checkbox** — forces the tag to count as a completed sale in reporting

<figure><img src="/files/G2RTQUSrKZNX3Q7ndXzn" alt=""><figcaption></figcaption></figure>

The SOLD override is particularly significant: it decouples business-defined conversions from AI outcome labels. A checkout completion tag can be marked as SOLD independently of what the voice outcome was — ensuring revenue reporting reflects commercial reality.

***

#### Final Takeaway

February was about **giving operators control**.

Control over how agents are built and versioned. Control over what gets sent to the CRM and when. Control over how outcomes are defined and measured.

This release makes VoiceB meaningfully easier to configure, iterate on, and trust — across every layer of the agent stack.


# January 2026 — VoiceB.ai Product Release

By Alex Bisbe. February 5th, 2026.

January focused on **outcome governance and data integrity**.

This release strengthens how VoiceB classifies calls, enforces outcomes, and guarantees consistent labeling across inbound sources. The objective is simple: **cleaner data, stricter validity rules, and enterprise-grade outcome control**.

***

### 🚀 Key Highlights

* New **INVALID – CUSTOMER CARE** categorization
* Fixed **PENDING call logic** and display rules
* Guaranteed **FROM / TO consistency** for DDI and VoIP calls
* Introduction of **Outcome Governance**
* Custom, configurable **Outcome Tags** per AI agent or campaign

***

### 🛑 New INVALID Category: CUSTOMER CARE

VoiceB now automatically classifies **existing customer requests** as invalid demand.

If the AI detects:

* Invoice or billing questions
* Service or technical issues
* Incident reports
* Any customer-care related intent

The call is now closed as:

**`INVALID – CUSTOMER CARE`**

This prevents non-commercial demand from polluting sales metrics and ensures that VoiceB remains focused strictly on **revenue-driven conversations**.

***

### ❌ PENDING Call Logic Fix → INVALID – FAILED

We fixed the logic and display rules for calls stuck in **PENDING**.

What changed:

* PENDING is no longer a terminal state
* Failed or incomplete calls are now consistently classified as\
  **`INVALID – FAILED`**
* UI and reporting now reflect the real outcome

This eliminates ambiguity and removes artificial inflation of valid demand.

***

### ☎️ FROM / TO Consistency for DDI and VoIP

Inbound call labeling is now **fully consistent** across all sources.

Improvements include:

* FROM / TO fields always populated correctly
* Direct DDI inbound numbers displayed accurately
* VoIP calls no longer shown as “–”, now explicitly labeled as **VOIP**

This ensures:

* Reliable attribution
* Cleaner analytics
* Fewer reconciliation issues with telco teams

***

### 🧭 Outcome Governance

January introduces **Outcome Governance** as a first-class platform concept.

Key principles:

* Outcomes are no longer free-text or loosely enforced
* Each outcome follows explicit rules
* Required fields are enforced before an outcome can be applied

This is foundational for operating VoiceB in **regulated, high-volume enterprise environments**.

***

### 🏷️ Outcomes Are Now “Tags”&#x20;

#### 🧩 Custom Outcome Tags per AI Agent or Campaign

Outcomes are now implemented as **Tags**, enabling structured and extensible classification.

What this unlocks:

* Clear mapping between Tags and the **Valid / Invalid backbone**
* Multiple sub-tags under the same validity category
* A shared language between AI, reporting, and business teams

This makes outcome data both **machine-reliable** and **business-readable**

VoiceB now allows **custom outcome tags** configured per:

* AI agent
* Campaign
* Use case

Each custom tag:

* Is still mapped to Valid or Invalid
* Follows the same governance rules
* Can be tailored to vertical or client-specific logic

Documentation:\
<https://docs.voiceb.ai/tags-custom-ai-classification-for-calls-and-leads>

***

### Final Takeaway

January was not about new channels or shiny AI features.

It was about **making VoiceB’s data trustworthy**:

* Cleaner invalidation
* Stricter outcomes
* Configurable but governed classification

This is the groundwork required to scale autonomous sales **without breaking reporting, attribution, or enterprise trust**.


# December 2025 — VoiceB.ai Product Release

By Alex Bisbe. December 30th 2025.

December was not about shipping shiny features. It was about **consolidation, positioning, and infrastructure readiness**. We tightened the stack and clarified where VoiceB is heading.

***

#### 🚀 WebRTC VoIP Channel (Now the Default)

The **WebRTC VoIP channel** is now fully production-ready and positioned as the **default voice channel** across VoiceB.

**What this delivers**

* Click-to-Call from any digital journey
* Inbound VoIP numbers
* Pure VoIP channels, browser-native
* Lower latency, higher audio quality, zero telco friction

<figure><img src="/files/RsFOIf96lKyikSy4ySvL" alt=""><figcaption></figcaption></figure>

This is the foundation for scalable, conversion-driven voice.

***

#### 🧩 New Widget Configuration (VoIP-Only by Design)

We rolled out a **new VoiceB Widget**, redesigned with a clear stance: **VoIP first, phone second**.

<figure><img src="/files/xOpNCdcJXzhrvdVEDlB1" alt="" width="268"><figcaption></figcaption></figure>

**Key improvements**

* Unified widget for all WebRTC use cases
* Cleaner UX, faster load, improved mobile behavior
* Native SuperVoice support
* Built to scale across agents, languages, and industries

Legacy phone widgets are no longer the reference model.

Know more, [click here.](https://docs.voiceb.ai/new-voiceb-web-widget-and-webrtc-voip-channel)<br>

***

#### 🧠 SuperVoice vs Traditional Phone Channels (Strategic Clarity)

Let’s call it out clearly.

**Phone channels**

* Reactive and interruptive
* Expensive and capacity-bound
* Optimized for support, not revenue

**SuperVoice (WebRTC)**

* Embedded directly in the digital funnel
* Proactive guidance at decision points
* Unlimited concurrency
* Designed to **increase conversion and lead quality**

SuperVoice is not “AI answering calls”.\
It’s **AI voice acting inside the buying journey**.

<figure><img src="/files/ZrNZvfT3LDb3nUqcJ3Ba" alt=""><figcaption></figcaption></figure>

🚗 Example of SuperVoice for automotive [here](https://docs.voiceb.ai/blog/supervoice-video-demo-for-automotive).&#x20;

***

#### 👩🏻‍💻📲 VoIP-to-Phone Transfer

VoiceB now supports **VoIP-to-phone call transfer**, allowing conversations to **start in WebRTC and seamlessly escalate to human agents on regular phone lines**.

This widens the funnel with low-friction VoIP entry while ensuring that only **qualified, high-intent conversations** reach the call center—combining AI scalability with human closing power.\
\
Know more about Call Transfers: [here](https://docs.voiceb.ai/call-transfer-redirect-calls-to-a-human-agent).

***

#### ⚙️ Platform Hardening & Operational Control

Behind the scenes, December focused on control and reliability:

* Clearer separation between **agent logic** and **channel execution**
* Safer configuration and save flows
* Better version visibility and deployment confidence
* Reduced operational risk in high-volume environments

Less friction. More predictability.

***

#### 🔜 What’s Next

No ambiguity on direction:

* **VoIP over API** (programmatic VoIP channels)
* Deeper SuperVoice analytics tied directly to revenue KPIs
* Vertical-specific SuperVoice templates
* Expanded enterprise governance and controls

***

#### Bottom Line

December was about **locking the voice infrastructure**.

VoiceB already operates as an API-driven platform.\
With WebRTC VoIP now solid and VoIP on API coming next, we’re turning VoiceB into the **programmable voice layer for inbound sales**.

We’re not evolving call centers.\
We’re building what replaces them.


# November 2025 – VoiceB.ai Product Release

By Alex Bisbe. November 27th, 2025

November delivered a set of high-impact improvements across data handling, configuration usability, and agent testing workflows. The focus was to streamline operational readiness, eliminate friction in agent setup, and increase the value of Talk-To-Agent outcomes by surfacing actionable insights directly inside the platform and through integrations.

### **1. Talk To Agent – Collected Data Now Visible & Fully Forwarded**

Talk-to-Agent outcomes now include **all COLLECTED DATA captured by the AI**.\
This dataset is:

* Displayed in the VoiceB app inside the call details view.
* Delivered through Webhook payloads in real time to CRM, CDP, or internal systems.

<figure><img src="/files/8lI7JpZH6UJ1iE0HZDh7" alt=""><figcaption></figcaption></figure>

This closes the loop between qualification and handover, ensuring downstream teams receive structured, validated intel without manual extraction.

***

### **2. Script Save Experience – “Last Save” Indicator Added**

We added a persistent save-status indicator inside the script editor:

**Last save: October 18th, 05:33pm CET**\ <br>

<figure><img src="/files/OkyCf0vB95fOxynR1eIL" alt=""><figcaption></figcaption></figure>

This creates full transparency for operators and prevents accidental overwrites during multi-person workflows. Saves are now explicit, visible, and traceable.

***

### **3. Automatic Merge of Data Collection + Dynamic Variables**

VoiceB now merges **dynamic variables** with **Data Collection fields** automatically.

If the form and the AI agent collect fields with the same name (e.g., **name**, **last\_name**, **phone**, **email**, **product\_of\_interest**), the system merges them into a single record.\
No duplicates. No collisions. No manual cleaning.

<figure><img src="/files/vL7dECumbYzIOcnzWVf1" alt=""><figcaption></figcaption></figure>

This unlocks cleaner exports, cleaner CRM syncs, and more reliable downstream analytics.

***

### **4. Improved Data Collection Selector (UI Upgrade)**

In Agents Configuration, the Data Collection selector has been re-designed to eliminate guesswork:

* Displays the **schema field names** of the selected collection.
* Shows **tabbed navigation** of all fields included (see screenshot reference).
* Ensures operators instantly understand what the agent will request and what will be pushed downstream.

<figure><img src="/files/TPkuTtEcOaNJX8d3Ra42" alt=""><figcaption></figcaption></figure>

This turns data-collection setup into a visual, structured experience rather than a blind configuration.

***

### **5. Test Click-to-Call – New Instant Test Feature**

A **“Send Test Call”** function has been added at the bottom of the Click-to-Call configuration drawer (see screenshot).

This enables operators to:

* Trigger instant test calls to validate the workflow
* Verify variable injection
* Ensure the endpoint is correctly configured
* Validate call routing before a campaign goes live

<figure><img src="/files/fC4MReQURNCEA93tuak5" alt=""><figcaption></figcaption></figure>

This reduces go-live friction and increases operational confidence across teams.

***

## **Summary**

November’s release strengthens VoiceB’s operational backbone:

* Richer Talk-to-Agent handovers
* Cleaner and smarter data unification
* More transparent script operations
* Faster agent testing
* More intuitive configuration workflows

These enhancements sharpen execution quality for enterprise deployments and accelerate time-to-value for every new campaign launched.

***


# August–October 2025 Product Updates

By Alex Bisbe. October 31st 2025.

Over the last three months we’ve focused on tightening core reliability, closing feature gaps, and elevating VoiceB to enterprise-grade standards. The theme has been simple: eliminate friction, strengthen accuracy, and scale with confidence.

Below is a consolidated view of everything shipped from **August to October**.

***

## **1. Call Transfer — Completed & Fully Operational**

The transfer feature is now fully implemented across all inbound flows.

**What’s live:**

* Stable **SIP → Human Agent Transfer** with confirmation tone
* Automatic event logging (Transferred = TRUE)
* Console integration showing: timestamp, destination, AHT pre-transfer
* Fail-safe timeout + retry logic

This closes a key blocker for telecom, utilities, and banking use cases.

Documentation: <https://docs.voiceb.ai/call-transfer-redirect-calls-to-a-human-agent>

***

## **2. Outcome System Overhaul — Accuracy First**

We rebuilt the outcomes logic end-to-end to stop misclassification and give clients reliable reporting.

#### **New outcome rules deployed (August–September):**

* **SOLD**\
  Triggered only when **all required data fields** are captured (per campaign config).\
  This prevents over-reporting and aligns with Bysidecar/telecom requirements.
* **DROPPED <15s**\
  Any call **without customer interaction** and **below 15 seconds** is now classified as DROPPED.\
  Eliminates noise, tests, and accidental pickups from funnel metrics.
* **SUPPORT / INVALID / NO INTEREST** untouched but cleaned and normalized.

This has already reduced false positives and improved client trust in the AI outcomes.

***

## **3. Metrics Fixes & Additions (Urgent Blockers Closed)**

We took the opportunity to reset how we calculate and display core KPIs.

#### **New or fixed metrics:**

* **Accurate AHT** (including/excluding transfer time)
* **Call Length** available in both dashboard & CSV
* **Validated Calls** excluding tests when toggled OFF
* **Transferred Rate**
* **Real Conversion Rate** based on *valid sales interactions*, not total traffic

All metric fixes are now fully rolled out for Vodafone, Primagas, AXA and test accounts.

***

## **4. Dashboard Cleanup — Corrected Calculations**

The dashboard now operates on production-grade logic:

* All test calls are **removed** instantly when toggling “Exclude Tests”
* Aggregated KPIs (AHT, Valid %, Sold %, Transfer %) are now **correctly weighted**
* Duplicate calls filtered automatically
* Fixed “To:” calculations for hourly and daily breakdowns
* Unified timeline so that transcripts + metrics always match

This significantly improves reporting integrity and prevents back-and-forth with clients.

***

## **5. CSV Export — Expanded & More Useful**

The CSV export now includes everything customers expect from a modern call-center analytics suite.

#### **New CSV columns:**

* **Transferred (YES/NO)**
* **Call Length (seconds)**
* **Outcome (normalized)**
* **ASR Confidence**
* **Test Call (flag)**
* **AHT**
* Timestamps (start, end, transfer event)

This closes the analytics gap and supports external BI pipelines.

***

## **6. Vanta Integration — ISO 27001 + SOC2 (Initiated)**

We officially kicked off the full compliance track with Vanta.

#### **What’s in flight:**

* Automated monitoring of infrastructure
* Policies uploaded: Access control, Secure SDLC, Incident Response, Vendor management
* Device management enforced
* Internal audit preparation for SOC2 Type II
* Evidence collection for ISO 27001 Annex A controls

This pushes VoiceB into enterprise-ready territory and unlocks procurement for Tier-1 accounts.

***

## **7. Platform Stability Improvements (August–October)**

We deployed multiple reliability upgrades across the system:

#### **ASR/TTS Stability**

* Rebalanced load across ElevenLabs + fallback
* Reduced voice drift issues
* Eliminated accidental “code reading” incidents with new guardrails

#### **Agent Behavior**

* Better recovery from long silences
* Improved objection routing
* Tiered confidence-based re-prompts

#### **Data Collection**

* Stricter validation of phones, emails, ID numbers
* Removal of partial/ambiguous data capture

These fixes materially improved conversion and customer perception.

***

### **8. New Public Status Page — Live SLA Tracking**

We launched [**status.voiceb.ai**](https://status.voiceb.ai) a real-time public status page that gives customers full transparency into platform performance and uptime.\
The page tracks:

* **99.99% SLA** across all core services
* SuperVoice uptime
* SuperQualification uptime
* API health
* SIP/Twilio connectivity
* Incident history and live updates

This rolls out enterprise-grade visibility and strengthens trust with high-volume telecom, utilities, and financial clients.

***


# July 2025 – Product Updates

Here’s what’s new in VoiceB.ai this month — including powerful features, important fixes, and our brand new Help Center.

### 🌐 New Website & Help Center Launch

We’ve launched our completely **redesigned website** [**voiceb.ai**](/) and a fully searchable **Help Center** on [docs.voiceb.ai](https://docs.voiceb.ai)!

* The new docs include:
  * Full product documentation
  * Setup guides
  * Blog articles
  * API references
  * Release notes (like this!)

→ 🔍 You can now find everything you need in one place — faster and easier.

***

### 🚀 New Features

#### ✅ Required Fields in Data Collection Schema

Each field can now be marked as **required**. These required fields determine whether the call qualifies as **SOLD**.\
→ Available in the Data Collection Schema editor.

#### 📩 Post-Call Webhook with SOLD Validation

The webhook now evaluates if a call qualifies as SOLD based on required fields.\
→ If no fields are required, **1 collected field** is enough to mark a call as SOLD.\
\
![](/files/b1CamUiIotPHXFK1ZvCB)

#### 🔒 Locked Schema Improvements

You can now **open locked Data Collection Schemas** to view their configuration.\
→ Only the **required fields** are editable in locked mode.\
\
![](/files/rQsZTqY5P0CDnqzOwswm)

***

### 🛠 Fixes & Improvements

#### 🧱 Campaign Dashboard Stability

Clicking on calls with "Pending" status no longer causes the app to crash.

#### 🗣 ElevenLabs Provider Enhancements

Our ElevenLabs voice integration now accurately detects:

* `TALK_TO_AGENT`
* `RE_SCHEDULE`

#### 🖥 Click-to-Call Drawer UI Fix

The **"Show credentials"** toggle is now correctly visible inside the instruction drawer.

***

### 📢 Stay in the Loop

Check this section regularly or subscribe to our updates to stay informed on everything new in VoiceB.ai.

Need help or have feedback? → Contact <hello@voiceb.ai>


# Help Center

Start here. Browse every Help Center article from one page.

Find setup steps, scripting guides, APIs, and integrations.

{% hint style="info" %}
Need help fast? Email **<hello@voiceb.ai>** or use the in-app chat.
{% endhint %}

### Browse by category

<table data-view="cards"><thead><tr><th>Category</th><th data-card-target data-type="content-ref">Start here</th><th data-hidden data-card-cover data-type="image">Cover image</th></tr></thead><tbody><tr><td>Getting started</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/JWASdHkWGngL9HakHACR">/spaces/ggIr2RvhGm49D4vbjurT/pages/JWASdHkWGngL9HakHACR</a></td><td><a href="/files/Bt86PDF1e0iPQh6lSyVq">/files/Bt86PDF1e0iPQh6lSyVq</a></td></tr><tr><td>Setup &#x26; access</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/89FB48EssGPRVyrLUUoQ">/spaces/ggIr2RvhGm49D4vbjurT/pages/89FB48EssGPRVyrLUUoQ</a></td><td><a href="/files/Bt86PDF1e0iPQh6lSyVq">/files/Bt86PDF1e0iPQh6lSyVq</a></td></tr><tr><td>Prompting &#x26; scripts</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/X9Hua40etHyeqihqdRJt">/spaces/ggIr2RvhGm49D4vbjurT/pages/X9Hua40etHyeqihqdRJt</a></td><td><a href="/files/Bt86PDF1e0iPQh6lSyVq">/files/Bt86PDF1e0iPQh6lSyVq</a></td></tr><tr><td>Channels (Widget / Click-to-Call / DDI)</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/lLi7X6cOoUknRdJif4p7">/spaces/ggIr2RvhGm49D4vbjurT/pages/lLi7X6cOoUknRdJif4p7</a></td><td><a href="/files/Bt86PDF1e0iPQh6lSyVq">/files/Bt86PDF1e0iPQh6lSyVq</a></td></tr><tr><td>Integrations (HubSpot / Salesforce / Webhooks)</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/lpAhnGn0oUkg6JIefyYN">/spaces/ggIr2RvhGm49D4vbjurT/pages/lpAhnGn0oUkg6JIefyYN</a></td><td><a href="/files/Bt86PDF1e0iPQh6lSyVq">/files/Bt86PDF1e0iPQh6lSyVq</a></td></tr><tr><td>Status, outcomes, and reporting</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/ZnL8VKgTZdHeV6eyOC9w">/spaces/ggIr2RvhGm49D4vbjurT/pages/ZnL8VKgTZdHeV6eyOC9w</a></td><td><a href="/files/Bt86PDF1e0iPQh6lSyVq">/files/Bt86PDF1e0iPQh6lSyVq</a></td></tr><tr><td>Pricing</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/ONtUo2WEmbtnp74Nz2Dx">/spaces/ggIr2RvhGm49D4vbjurT/pages/ONtUo2WEmbtnp74Nz2Dx</a></td><td><a href="/files/Bt86PDF1e0iPQh6lSyVq">/files/Bt86PDF1e0iPQh6lSyVq</a></td></tr><tr><td>Legal &#x26; compliance</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/ViVgMtXMGVFEL0d23szE">/spaces/ggIr2RvhGm49D4vbjurT/pages/ViVgMtXMGVFEL0d23szE</a></td><td><a href="/files/Bt86PDF1e0iPQh6lSyVq">/files/Bt86PDF1e0iPQh6lSyVq</a></td></tr><tr><td>Call and lead outcomes</td><td><a href="https://docs.voiceb.ai/understanding-status-and-outcomes-in-voiceb">https://docs.voiceb.ai/understanding-status-and-outcomes-in-voiceb</a></td><td><a href="/files/Bt86PDF1e0iPQh6lSyVq">/files/Bt86PDF1e0iPQh6lSyVq</a></td></tr></tbody></table>

### Popular articles

* [Getting set up](/help-center/getting-set-up)
* [Setting permissions](/help-center/getting-set-up/setting-permissions)
* [Inviting Members](/help-center/getting-set-up/inviting-members)
* [How to Update a VoiceB Agent Prompt & Script](/how-to-update-a-voiceb-agent-prompt-and-script)
* [Webhook – Call Outcome Integration](/webhook-call-outcome-integration)

### Browse all articles

<details>

<summary>Getting started</summary>

* [Welcome to VOICEB.ai DOCS](/)
* [Getting Started with VoiceB.ai: A Complete Tutorial](/getting-started-with-voiceb.ai-a-complete-tutorial)
* [What we do](/overview/what-we-do)
* [VOICEB Features](/overview/voiceb-features)

</details>

<details>

<summary>Setup &#x26; access</summary>

* [Getting set up](/help-center/getting-set-up)
* [Setting permissions](/help-center/getting-set-up/setting-permissions)
* [Inviting Members](/help-center/getting-set-up/inviting-members)

</details>

<details>

<summary>Prompting &#x26; scripts</summary>

* [VoiceB Prompt & Script Guide](/product-guides/voiceb-prompt-and-script-guide)
* [How to Create an Effective VoiceB Prompt: An Example from NexTel](/product-guides/how-to-create-an-effective-voiceb-prompt-an-example-from-nextel)
* [How to Update a VoiceB Agent Prompt & Script](/how-to-update-a-voiceb-agent-prompt-and-script)
* [Automate Your Knowledge Base with Web Scraping (BETA)](/product-guides/automate-your-knowledge-base-with-web-scraping-beta)

</details>

<details>

<summary>Channels (Widget / Click-to-Call / DDI)</summary>

* [New VoiceB Web Widget & WebRTC VoIP Channel](/widget-configuration-guide/new-voiceb-web-widget-and-webrtc-voip-channel)
* [API – Click-to-Call](/api-click-to-call)
* [Using Variables in Click to Call to Personalize Conversations](/using-variables-in-click-to-call-to-personalize-conversations)
* [Call Transfer: Redirect Calls to a Human Agent](/call-transfer-redirect-calls-to-a-human-agent)
* [Test Numbers](/test-numbers)

</details>

<details>

<summary>Data, outcomes, and integrations</summary>

* [Understanding Status & Outcomes in VoiceB](/understanding-status-and-outcomes-in-voiceb)
* [Create a Data Collection Schema](/create-a-data-collection-schema)
* [Webhook – Call Outcome Integration](/webhook-call-outcome-integration)
* [Real-Time Data Checks via Customer Endpoints](/real-time-data-checks-via-customer-endpoints)
* [How to Connect Your VoiceB.ai Agent with HubSpot](/how-to-connect-your-voiceb.ai-agent-with-hubspot)
* [How to Connect Your VoiceB.ai Agent with Salesforce](/how-to-connect-your-voiceb.ai-agent-with-salesforce)
  * [How to register a local External Client App in Salesforce](/how-to-connect-your-voiceb.ai-agent-with-salesforce/how-to-register-a-local-external-client-app-in-salesforce)

</details>

<details>

<summary>Product updates</summary>

* [December 2025 — VoiceB.ai Product Release](/product-updates/december-2025-voiceb.ai-product-release)
* [November 2025 – VoiceB.ai Product Release](/product-updates/november-2025-voiceb.ai-product-release)
* [August–October 2025 Product Updates](/product-updates/august-october-2025-product-updates)
* [July 2025 – Product Updates](/product-updates/july-2025-product-updates)

</details>

<details>

<summary>Pricing</summary>

* [Pricing Guide](/pricing-guide)

</details>

<details>

<summary>Use cases</summary>

* [How SuperVoice Converts Website Visitors to Customers](/how-supervoice-converts-website-visitors-to-customers)
* [Implementing SuperQualification for B2C & B2B Lead Scoring](/implementing-superqualification-for-b2c-and-b2b-lead-scoring)

</details>

<details>

<summary>Legal &#x26; compliance</summary>

* [Terms & Conditions](/legal-and-compliance/terms-and-conditions)
* [Privacy Policy](/legal-and-compliance/privacy-policy)
* [Data Sub-processors](/legal-and-compliance/data-sub-processors)
* [Cookies](/legal-and-compliance/cookies)

</details>


# Getting set up

<details>

<summary>Step1: Creating your organization</summary>

#### ✅ Organization setup (done by us)

To ensure a secure and frictionless onboarding, the **creation of your VoiceB.ai organization is handled directly by our support team**. You don’t need to create it yourself — we’ll do it for you.

Once your organization is created, we’ll assign you as the **Admin**, and you’ll be ready to start inviting collaborators.

***

#### 👥 Inviting your team

From your **Organization settings**, you can easily invite team members and manage their roles.

**To invite your team:**

1. Go to **Settings** → **Organization** → **Members**
2. Click on **Invite**
3. Enter the teammate’s email and select their **role** (Admin or Member)
4. Click **Send invitation**

🧩 *Admins* have full access to script configuration, DDI/API setup, and analytics.\
👤 *Members* can monitor conversations, review transcripts, and assist with content.

> You can always edit or revoke invitations later from the Members panel.

📸 *Example view:* `Members screenshot`![](/files/09eEa5tJdAP4VQyv6xQK)

***

#### 🧾 Profile Management

Every user can manage their profile and emails under **Settings** → **Account**.

* Change your profile name or image
* Add or remove secondary email addresses

📸 *Example view:* `Profile screenshot`

<figure><img src="/files/NXTeDHoMzSQ8dxlZNyZf" alt=""><figcaption></figcaption></figure>

***

🔐 **Need help or want to modify your organization setup?**\
Contact our team through the in-app chat or email us at **<hello@voiceb.ai>** and we’ll assist you right away.

</details>

<details>

<summary>Step 2: Create a campaign</summary>

#### 📦 Create your first campaign

To launch your VoiceB.ai agent, you’ll first need to create a campaign. This will define the context, goal, and time window for the agent to operate in.

**🪄 How to create a campaign:**

1. Go to the **Dashboard**
2. Click **“New Campaign”**
3. Fill out the campaign form as follows:

| Field             | Description                                                                                                                                                                                                 |
| ----------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Campaign Name** | Give your campaign a name — e.g. “Black Friday 2025” or “Website Assistant - Spain”                                                                                                                         |
| **Type**          | <p>Choose one of the three campaign types:<br>• <strong>SuperVoice (Digital Journey Assistant)</strong><br>• <strong>SuperQualification (Lead Qualification)</strong><br>• <strong>Performance</strong></p> |
| **Description**   | Optional — add internal notes or context for your team                                                                                                                                                      |
| **Start Date**    | Set the start date. Leave empty to activate immediately                                                                                                                                                     |
| **End Date**      | Optional — set an end date if this is a temporary or seasonal campaign                                                                                                                                      |

Once you’re done, click **Create**.

📸 *Example view:*<br>

<figure><img src="/files/k7Ghq1DrJaS7uyEM11YU" alt=""><figcaption></figcaption></figure>

***

> ✅ **Tip:** You can edit or pause the campaign at any time, and assign different agents or scripts later from the Config panel.

📍**Next up:** Let’s configure your AI Agent — you’ll define the **Goal**, write the **Script**, and upload **Objections** in Step 3.

</details>

<details>

<summary>Step 3: Create an AI Agent and Prompt (Goal + Script + Objections)</summary>

In this step, you'll define the behavior and purpose of your VoiceB.ai agent. This is where the magic happens — you’ll set your agent's goal, write the conversation flow, handle objections, and choose voice settings.\
\
Quick tour of the VoiceB dashboard and features: <https://docs.voiceb.ai/overview/voiceb-features>

***

#### 1️⃣ Configure your AI Agent

Go to your campaign’s **Config** section and complete the following fields:

| Field            | Description                                                                                |
| ---------------- | ------------------------------------------------------------------------------------------ |
| **Agent’s Name** | Choose a name that reflects the purpose (e.g. *Anna – SuperVoice*, *Luis – Qualification*) |
| **Provider**     | Select your preferred voice engine: **ElevenLabs** (recommended) or **Bland.ai**           |
| **Language**     | Pick the language and voice accent (e.g. English US, Spanish, etc.)                        |
| **Type**         | Select **Inbound** for most cases. Use **Callback** only if triggered by API.              |

***

#### 2️⃣ Define the Agent’s Goal

The **Goal** tells the AI what it's trying to achieve. Keep it focused and outcome-driven.

Examples:

* “Identify the user’s intent (purchase, support, or qualification), assist them through the process using real-time voice interaction.”
* “Qualify leads by asking a set of structured questions and route only qualified ones to sales.”

***

#### 3️⃣ Write the Initial Message

This is the **first sentence** your AI agent will say. Keep it friendly, clear, and brief.

Example:\
`Hi! I’m Anna, your AI assistant from [your company] — the most...`

If this field is left empty, the agent will wait for the user to start the conversation.

***

#### 4️⃣ Add Your Script

This is the core of the interaction — your product pitch, explanation, and flow logic.

Example template:

```
pgsqlCopyEditHello, (name of the user)

Let me explain — VoiceB is not just another chatbot or automation system. We don’t replace human agents; we support them with real-time, scripted conversations powered by AI voice.

We guide users through product selection, form completion, and answer questions instantly. All of it is trained using your own public website and documents.
```

> ✅ **Tip:** Use short sentences, conversational tone, and include soft confirmations (e.g. “Would you like to know more?”)

***

#### 5️⃣ List Common Objections

VoiceB lets you pre-train the AI with answers to common objections. Add at least 3–5 common ones.

Examples:

* **"How long does it take to implement?"**\
  → “Just one line of JS — live in under a day.”
* **"Can it integrate with our CRM or APIs?"**\
  → “Yes, we support secure APIs and webhook-based integrations.”
* **"What about updates to the content?"**\
  → “We scan your site daily. If you update a form, price, or product — the AI adapts automatically.”

***

#### 6️⃣ Select Data Collection (Optional)

If your flow requires collecting user input (e.g. name, phone, preferences), you can configure **Data Collection** here.\
If not required, leave it as **Without Data Collection**.

***

#### 7️⃣ Connect a DDI or Inbound Number

From the **DDI** dropdown, select the number the AI should listen to.\
If you're embedding on a website only, this can be optional.

***

#### 8️⃣ Add Integrations

You can connect tools like **HubSpot**, **Salesforce**, or use our API/Webhook integration to send lead data or call outcomes in real time. (See STEP 4)

***

When everything is complete, click **Save**.

📍Next up: In **Step 4**, you’ll connect the agent to a real-world channel — via phone (DDI) or API-based callback.

</details>

<details>

<summary>Step 4: Connect VoiceB to Your Channel </summary>

Inbound via DDI, Callback API, or SIP Trunk + Post-Call Integration to HubSpot, Salesforce or Webhook.

Now that your AI agent is fully configured, it's time to make it operational — by connecting it to your users.

VoiceB.ai supports multiple routing options for **inbound voice conversations**, whether they originate from your website, a phone number, or a telephony system. After the conversation, outcomes can be pushed to your CRM or custom backend — instantly.

***

#### 🔊 SuperVoice (Website Voice Assistant via JavaScript)

Use case: Voice-qualified journeys embedded directly in your website.

**How it works:**

* Add **one line of JavaScript** to your site — that’s it.
* The AI widget appears, users click to speak.
* Calls are routed through VoIP to a VoiceB DDI or to a callback number.

> ✅ This is the fastest way to go live — no training, no dev effort.

***

#### 📞 SuperQualification (Inbound Calls via DDI or SIP)

Use case: Qualify leads from incoming phone calls using AI.

**VoiceB supports three routing modes:**

**1. 100% call routing to VoiceB**

Forward all your traffic to a dedicated VoiceB **DDI** (Direct Dial-In). The AI answers all calls, qualifies, and stores or routes outcomes.

**2. Partial routing (Overflow, After-hours, A/B tests)**

Only send some calls — based on time, volume, or logic — to VoiceB.\
This is ideal for:

* After-hours or weekend routing
* Overflow handling during peak loads
* Piloting VoiceB with a small % of users

📌 **Important:** Ensure your telephony system passes the **original Caller ID** when forwarding calls to VoiceB. This is critical for call tracking, CRM matching, and analytics.

**3. SIP Trunk Integration (Advanced)**

Connect your SIP trunk directly to our platform for full control and scale.\
Perfect for enterprise-scale routing and PBX-level configuration.

📧 To set up SIP trunking, contact **<hello@voiceb.ai>** — our support team will assist you.

***

#### 🔁 API-based Callback

VoiceB can also place outbound calls **on demand**, triggered by your product or form.

* Send a call request to our API (e.g. when a user submits a lead)
* VoiceB instantly calls the user
* The AI takes over the interaction with your preconfigured script

Ask support to activate this feature for your account.

***

#### 📤 Post-Call Integrations

Once a call is completed, VoiceB automatically sends all call data — including transcripts, agent info, call time, and qualification status — to your systems.

**✅ Native integrations:**

* **HubSpot**: Map fields like status, agent, call time, and transcript into your CRM automatically.
* **Salesforce**: Native sync for call outcomes and contact activity.

📸 *Example UI:*

<figure><img src="/files/dEmqcn35ypgLWTHNOOdU" alt=""><figcaption></figcaption></figure>

**🌐 Webhook:**

* Push data to your own system in real time
* Supports:
  * Custom auth token
  * Optional HMAC-SHA256 signature verification
  * Full payload including transcript and collected data

📸 *Webhook example:*

<figure><img src="/files/IhrO3W3oPpCISFFFKmHr" alt=""><figcaption></figcaption></figure>

```json
jsonCopyEdit{
  "agent": "Claire (VoiceB.ai)",
  "from": "+123456xx90",
  "to": "+349303xx814",
  "start": "...",
  "end": "...",
  "status": "SOLD",
  "transcript": [...],
  "collectedData": {
    "name": "John Doe",
    "email": "john@example.com",
    "address": "Calle Ejemplo 123, Ciudad"
  }
}
```

***

🎉 That’s it! You’re now ready to receive real calls — qualified, assisted, and fully integrated into your workflow.

</details>

{% hint style="info" %}
**Book a demo** [**here**](https://cal.com/team/voiceb-demo/voiceb-demo)**.**
{% endhint %}


# Setting permissions

VoiceB.ai keeps permissions simple and secure by offering two clearly defined roles: Admin and Member.  Assign roles based on how each team member will interact with your AI agents and data.

* Assign **Admin** access only to users who need to build or modify agent flows, SIP Trunk settings, or Tools
* Use the **Member** role for marketing, operations, or sales teams who need to track performance and handle leads
* You can change a user's role at any time from the **Organization → Members** section

<div data-with-frame="true"><figure><img src="/files/FA5cuV5CP5xw68TA7dgx" alt=""><figcaption></figcaption></figure></div>

## Setting permissions

{% hint style="info" %}
**VoiceB tip:** Keep your AI agent secure — grant Admin rights only to users directly responsible for configuration and script maintenance.
{% endhint %}

#### 👥 Permission Levels

| **Role**   | **Capabilities**                                                                                                                                                                                                                                    |
| ---------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Admin**  | Full access: configure agents in **Agent Studio**, edit scripts and objections, manage **SIP Trunk** configuration and **Tools**, manage integrations, and manage users — including inviting new users and changing roles between Member and Admin. |
| **Member** | Limited access: view analytics, monitor call outcomes, listen to recordings, and access collected lead data — but cannot edit or configure agents.                                                                                                  |

***

#### 🛠️ What Admins can access

Admins have exclusive access to the platform's configuration areas:

* [**Agent Studio**](https://docs.voiceb.ai/how-to-update-a-voiceb-agent-prompt-and-script) — create and configure agents, edit scripts, objections, and script blocks
* [**SIP Trunk configuration**](https://docs.voiceb.ai/sip-trunk-private-call-routing-for-enterprise-voice-ai) — connect and manage your dedicated enterprise call routing
* **Tools** — configure the tools your agents can call during conversations
* **User management** — invite additional users and change any user's role from Member to Admin (or back)

{% hint style="info" %}
Only Admins can promote a Member to Admin. Grant this role deliberately, since Admins can modify live agent configuration.
{% endhint %}

***

#### 🔐 Best Practices

* Assign **Admin** access only to users who need to build or modify agent flows, SIP Trunk settings, or Tools
* Use the **Member** role for marketing, operations, or sales teams who need to track performance and handle leads
* You can change a user's role at any time from the **Organization → Members** section


# Inviting Members

Once your VoiceB organization is created, you can easily invite your team members and assign them roles based on their level of access. Roles define what each user can see and do inside the platform.

{% hint style="info" %}
**VoiceB tip:** Assign at least two Admins to your organization — this ensures you always have full access even if someone is unavailable or leaves the team.
{% endhint %}

### 🔐 Roles Overview

| Role       | Permissions                                                                                                |
| ---------- | ---------------------------------------------------------------------------------------------------------- |
| **Admin**  | Full access: can create/edit campaigns and agents, manage team members, update billing, and view analytics |
| **Member** | Can view dashboards, transcripts, and analytics. Cannot edit scripts or manage campaigns                   |

***

### 🚀 How to Invite Members

1. Navigate to **Settings → Organization → Members**
2. Click **Invite**
3. Enter the email address of the team member
4. Choose their **Role** (Admin or Member)
5. Click **Send invitation**

The invited user will receive an email to join the organization and complete their setup.

📸 *Example view:*

<figure><img src="/files/BiNctRkpodzYhjjQozpP" alt=""><figcaption></figcaption></figure>

Choose their **Role** (Admin or Member)\
![](/files/GZ6al9YIhuhM9xHeNjbo)

***

### ✏️ Changing a Member’s Role

Admins can update roles at any time:

* Go to **Settings → Members**
* Use the dropdown under the **Role** column next to each member
* Select **Admin** or **Member**

> ✅ **Tip:** Only Admins can change roles or remove users.

***

Need help managing team permissions?\
Reach out to **<hello@voiceb.ai>** — we’ll help you set everything up securely.


# Getting Started with VoiceB.ai: A Complete Tutorial

VoiceB.ai is redefining how businesses connect with customers using AI-powered voice agents that sell, qualify, and assist — just like a real person, but available 24/7 and scalable beyond human limit

<sup>Alex Bisbe   3/24/2026  6 min read</sup>

{% embed url="<https://www.loom.com/share/42f9f906163b44f3bc1d320762887c8a>" %}

### Getting Started with VoiceB.ai: A Complete Tutorial

Whether you're in finance, insurance, telco, or retail, launching your first AI voice campaign with VoiceB is easier than ever. Here's your step-by-step guide.

#### **Step 1: Create Your Account**

The journey starts with a simple sign-up. Creating your VoiceB.ai account gives you immediate access to the platform so you can begin building your first AI sales agent. No technical background? No problem — the setup is designed for speed and simplicity.

#### **Step 2: Set Up a Campaign**

What kind of calls do you want your AI agent to handle? Whether it’s **telesales**, **lead qualification**, or **appointment booking**, you can define your campaign's goal and structure. VoiceB lets you tailor the interaction to your sales process, ensuring each call delivers value.

#### **Step 3: Train the AI**

Your AI agent is only as good as the knowledge it’s trained on. With VoiceB, training is fast, intuitive, and powerful. You can:

* Upload your sales scripts
* Add product documentation and brochures
* Include internal FAQs and business-specific objection handling
* **Automatically scrape your public website** to generate a rich knowledge base

VoiceB’s AI uses this data to deliver dynamic, accurate, and natural responses — handling everything from product questions to complex objections with confidence.

#### **Step 4: Connect & Go Live**

Once your campaign and training are in place, it’s time to launch. You can:

* Buy a dedicated phone number, or
* Integrate VoiceB with your existing call system using the API

Within minutes, your AI agent can begin handling inbound or outbound calls, fully operational and ready to scale.

#### **Step 5: Pay for Performance**

With VoiceB, there are no hidden costs or monthly surprises. You only pay **per connected lead**, making pricing fully transparent and aligned with your business outcomes. It’s performance-based, just like a great sales team.

#### **Final Thoughts**

In just five steps, you can go from concept to conversion — deploying a sales-ready AI agent that never sleeps. With VoiceB.ai, you're not just automating calls; you're transforming how your business grows.

Ready to see it in action? [Talk to Sales](https://cal.com/team/voiceb-demo/voiceb-demo) and start building your AI-powered future today.


# Agent Studio

By Alex Bisbe. June 29th 2026.

{% hint style="warning" %}
**Beta · Admin only.** Agent Studio is currently in Beta and available exclusively to admin users in your organization. Features and behavior may change. As the assistant works with live data, always verify critical figures before acting on them.
{% endhint %}

## What is Agent Studio?

Agent Studio is a conversational copilot built into your VoiceB.ai dashboard. Instead of clicking through screens, exports, and configuration forms, you describe what you want in plain language and the assistant does it for you — querying your data, generating charts, creating and editing agents, and managing your web widget, all from a single chat.

You can reach it from **Agent Studio** in the left-hand navigation.

<figure><img src="/files/My8R41vp8fJS0TdjbbMD" alt=""><figcaption></figcaption></figure>

*The Agent Studio home screen, with suggested prompts to get you started.*

## Getting started

Open Agent Studio and either type your question in the input box at the bottom or tap one of the suggested prompts, such as:

* List all my campaigns with their agent counts
* Show me metrics for the last 30 days
* Compare agent performance across campaigns
* Export calls from this week to Excel

Your previous sessions are saved. Open the **Conversations** panel (top right) to search, revisit, rename, or delete past chats.

<figure><img src="/files/YuOcnpWJh9SHFC8dLV4p" alt=""><figcaption></figcaption></figure>

*The Conversations panel keeps a searchable history of your sessions.*

## Choosing a model

Agent Studio lets you pick the underlying model from the selector at the top right. All models can perform the same actions — the difference is in tone and how the response is structured, so you can choose whichever style you prefer.

Currently available:

| Provider  | Model            | Notes                                 |
| --------- | ---------------- | ------------------------------------- |
| Google    | Gemini 3.5 Flash | Default. Fast, detailed responses.    |
| Anthropic | Claude Sonnet 5  | Concise, well-organized replies.      |
| OpenAI    | GPT 5.3 Chat     | Goal-oriented, step-by-step guidance. |

<figure><img src="/files/qvbhLlSkWIAo1BOsrm3t" alt=""><figcaption></figcaption></figure>

*Switch models at any time from the selector. The action set is identical across models.*

## What you can do

### Querying, analytics & reporting

* **Monitor performance** — fetch aggregate metrics for your campaigns, including total calls, average call duration, conversion rates, and outcome breakdowns.
* **Track trends** — view call volume over time, agent-by-agent comparisons, and visual funnels.
* **Inspect calls** — list calls with filters (date range, agent, outcome such as *sold* or *busy*), and open full transcripts, summaries, and collected lead data for individual calls.
* **Visualize data** — generate interactive charts (pie, bar, line, funnel, radar, heatmap, and more) directly in the chat.
* **Export reports** — prepare and export call logs and campaign data to CSV or Excel for offline analysis.

<figure><img src="/files/PlixPuALir5XjITYu4MR" alt=""><figcaption></figcaption></figure>

*Ask for metrics and Agent Studio returns a structured performance overview across campaigns.*

<figure><img src="/files/TcILJStflgBZbbQ7iiSi" alt=""><figcaption></figcaption></figure>

*Outcomes are rendered as interactive charts, with written key takeaways alongside.*

### Creating & modifying AI agents

* **Create campaigns & agents** — spin up new voice campaigns or design new conversational agents from scratch.
* **Write & edit scripts** — configure and refine agent behavior step by step using structured script blocks (Goal, Script, Objections, Guardrails, or custom blocks). You can add, edit, reorder, and delete blocks.
* **Adjust agent settings** — change language, model, voice provider, and memory.
* **Assign phone numbers (DDIs)** — check available numbers in your organization and assign or unassign them to agents.
* **Organize with tags** — set up and edit custom AI-evaluated tags that are applied to calls automatically based on criteria you define (for example, "Interested Customer" or "Spanish Speaker").

<figure><img src="/files/xoS3BqGJuVTLNteo2CU2" alt=""><figcaption></figcaption></figure>

Agent Studio walks you through creation and edits one decision at a time, so you don't have to fill out long forms.

### Widget management

* **Configure your web widget** — adjust your click-to-call or VoIP web widget, including layout (Bubble vs. Compact), brand colors, button styles, rounded corners, spacing, and per-language texts.
* **Preview live** — review changes before saving.
* **Assign agents** — add or remove agents from the widget.
* **Track engagement** — monitor impressions, click-through rates, call initiation rates, and UTM campaign sources.
* **Manage compliance** — configure GDPR consent requirements and privacy policy / terms URLs, and review logs of who accepted your terms before calling.

### Web knowledge fetching

Paste any URL — an FAQ page, product catalog, or pricing sheet — and Agent Studio reads and summarizes the content to help you write or enrich an agent's knowledge base.

## Example prompts

```
Show me a list of my active campaigns.
Compare agent performance in my latest campaign with a bar chart.
Help me create a new inbound agent.
Look up the transcript for my last call.
Export this month's calls to Excel.
Read https://example.com/pricing and add it to my agent's knowledge.
```

## Things to know

{% hint style="info" %}
**Verify critical data.** Agent Studio is an AI assistant and may occasionally produce inaccurate information. For anything that drives a decision, confirm the underlying figures in the relevant dashboard section.
{% endhint %}

* Agent Studio acts on your organization's live data and can make real changes (creating agents, assigning DDIs, editing scripts). Review proposed changes before confirming.
* The feature is Beta and limited to admin users. If a teammate needs access, an organization admin can advise on availability.
* Switching models mid-conversation is supported; the available actions remain the same.


# How to Update a VoiceB Agent Prompt & Script

How to Configure a VoiceB Agent: Script Blocks Guide

*By Alex Bisbe. Last updated: March 2026*

***

### Before You Start

* You need access to the VoiceB dashboard at app.voiceb.ai
* Agent configuration is done per agent, not globally
* Changes do not go live until you explicitly click **Publish**
* The agent continues running on its last published version until you publish a new one

***

### Step 1: Navigate to Your Agent

1. Log in to app.voiceb.ai
2. From the left sidebar, select your campaign
3. Under **Agents**, click the agent you want to configure
4. The agent configuration panel will open, showing settings and Script Blocks below

***

### Step 2: Configure Basic Settings

Before editing the script, verify the top-level agent settings:

**Type** — the channel this agent handles (e.g., Click to Call, SIP Trunk, WebRTC Widget).

**Memory (days)** — how many days of previous call summaries the agent can access for returning callers. Set to 0 to disable. 14 days is a reasonable default for most sales use cases.

**Initial Message** — the first sentence the agent will say when a call connects. Keep it short: agent name, company, and a single opening question. Example: *"Hello, my name is Angela, I am your AI agent from NexTel. Thank you for contacting us. What is your name so I can address you properly?"*

<figure><img src="/files/Z926A4433GzfUR2HlzfM" alt=""><figcaption></figcaption></figure>

***

### Step 3: Understand Script Blocks

Script Blocks are the core of agent configuration. Instead of a single long text prompt, the agent's intelligence is broken into typed, independently editable blocks. Each block has a label and a specific purpose. The agent receives all blocks concatenated in order — the structure helps both you as an editor and the LLM as a reader.

From the screenshots, a typical agent has these blocks:

* Goal
* Role
* Objective
* Guardrails
* Script
* Available Plans (or equivalent product/pricing block)
* Objections

You can add custom blocks beyond these. Blocks can be reordered by dragging the handle (≡) on the left of each block.

<div data-with-frame="true"><figure><img src="/files/RJy9a6yfTRNotnmrJAlZ" alt=""><figcaption></figcaption></figure></div>

***

### Step 4: Block-by-Block Writing Guide

#### Goal

**Purpose:** One sentence that defines the single commercial outcome this agent is optimizing for. The LLM uses this as its north star throughout the conversation.

**How to write it:**

* One sentence only. No context, no explanation.
* Business-measurable. Focus on the outcome, not the behavior.
* Avoid multiple objectives — pick the primary one.

**Good examples:**

* *"Assist the customer in choosing the best NexTel fiber plan, check coverage, and complete the order including payment and installation scheduling."*
* *"Qualify the caller's energy consumption and close an autonomous sale of the most suitable Holaluz electricity plan."*
* *"Identify whether the caller is a valid sales prospect and transfer sales-ready leads to a human agent."*

***

#### Role

**Purpose:** Defines the agent's persona, communication style, and voice. This block shapes *how* the agent speaks, not *what* it does.

**How to write it:**

* Start with a single sentence defining the persona (name, company, role).
* Then add 4–6 voice style guidelines as a short list.
* Specify any structural decision logic (e.g., phone-based flow vs. website-assisted flow).

**What to include:**

* Tone (energetic, professional, empathetic, confident — pick what fits your brand)
* Pace guidance (never slow or robotic)
* What to avoid (monotone delivery, scripted-sounding answers)
* Any dual-path logic the agent must handle (e.g., caller is on the website vs. not)

**Example:**

> You are an engaging, dynamic, fast, and persuasive NexTel sales agent handling real phone-based subscriptions.
>
> Voice style guidelines:
>
> * Energetic, confident, natural pace (never slow or robotic).
> * Professional but engaging.
> * Clear articulation.
> * Slight enthusiasm when presenting benefits.
> * Never sound monotone or scripted.

***

#### Objective

**Purpose:** The step-by-step conversation structure the agent must follow. This is the tactical execution layer below the Goal.

**How to write it:**

* Use a numbered list of steps in conversational order.
* Each step should be a clear, actionable instruction.
* Include branching logic (if X, then Y).
* Specify what data to collect at each stage.
* Define when to escalate, transfer, or close.

**Example structure:**

> 1. Greet the caller and ask for their name.
> 2. Ask if they are calling about a new subscription or an existing one.
> 3. If new subscription: ask for their address to check fiber coverage.
> 4. Present the plan options available at their address.
> 5. Recommend the most suitable plan based on stated needs.
> 6. Confirm the caller's decision and collect billing details (IBAN).
> 7. Schedule the installation date.
> 8. Confirm the order summary before closing.

***

#### Guardrails

**Purpose:** Hard limits on what the agent must never do, regardless of what the caller asks. This is your compliance and brand protection layer.

**How to write it:**

* Use a short, direct list of prohibitions.
* Focus on the most critical failure modes for your specific use case.
* Include legal, pricing, and transfer rules.

**Recommended guardrails for most agents:**

* Never invent pricing, plans, or availability not listed in the script.
* Never confirm an order before collecting all required data fields.
* Never contradict what the caller has already been told in this call.
* If the caller is angry or requests a human agent, transfer immediately without attempting to retain.
* Do not engage with topics outside the defined scope of this agent.
* Do not make commitments about timelines or service levels not specified in the script.

***

#### Script

**Purpose:** The full conversation intelligence — product knowledge, pitch logic, qualification rules, data collection, and edge case handling. This is the most detailed block and the one you will iterate on most.

**How to write it:**

* Write in the second person, addressing the agent directly (*"When the caller asks about price..."*).
* Use clearly labeled sections (e.g., "Coverage check," "Plan presentation," "Closing").
* Include conditional branches explicitly: *"If the caller says they already have a contract, acknowledge it and proceed to..."*
* Define exactly when and how to ask for personal data.
* Specify what the agent should do if a required data field is refused.
* Include instructions for the most common conversation detours.

**Length:** This block will typically be the longest. Do not artificially constrain it. A complex product (telco, energy, insurance) may require 500–1,500 words here.

**What not to do:**

* Do not repeat content from Goal or Objective verbatim — the Script should add detail, not restate.
* Do not write a rigid word-for-word script — write logic and intent, not a teleprompter.

***

#### Pricing / Available Plans Block

**Purpose:** A dedicated block for all pricing, plan details, and offer conditions. Keeping this separate from the Script makes it easy to update without touching the conversation logic.

**How to write it:**

* List every available plan or product option with their key parameters.
* Include conditions, eligibility rules, and promotional terms if any.
* If prices change frequently, make this the only block you need to update.
* The agent will reference this block when presenting options to the caller.

**Example format:**

> FIBER 300 — €29.99/month. Download speed: 300Mb. Includes unlimited calls. Promotional price for first 12 months.
>
> FIBER 600 — €39.99/month. Download speed: 600Mb. Includes unlimited calls + one TV package.
>
> FIBER 1GB — €49.99/month. Download speed: 1Gb. Includes unlimited calls + two TV packages. Best value for households with 4+ devices.

***

#### Objections

**Purpose:** Pre-programmed responses to the most common resistance points. Well-written objection handling is one of the highest-ROI improvements you can make to a sales agent.

**How to write it:**

* Format each objection as: the caller's statement → the agent's response.
* Acknowledge before rebutting — never dismiss the objection.
* Keep responses short (2–4 sentences maximum per objection).
* End every response with a forward-moving question or statement.
* Aim for 5–10 objections covering the most common resistance patterns.

**Essential objections to cover:**

* *"I want to think about it."* → Acknowledge, create mild urgency without pressure, offer a specific next step.
* *"It's too expensive."* → Reframe on value or long-term cost, compare to what they currently pay.
* *"I already have a contract with another provider."* → Ask when it expires, position VoiceB's client offer as the obvious switch.
* *"I prefer to speak to a human."* → Transfer immediately. Do not argue.
* *"I'm just checking options."* → Lower commitment language, move to a softer close (e.g., check coverage, send info).
* *"I need to check with my partner/family."* → Acknowledge, offer to schedule a callback or summarize what they'd be sharing.

***

### Step 5: Adding New Blocks

To add a block beyond the defaults:

1. Scroll to the bottom of the Script Blocks section
2. Click **+ Add Block**
3. Name the block clearly (e.g., "Agent Instructions," "Competitor Handling")
4. Write the block content and save

Custom blocks are useful for separating concerns that don't fit the standard taxonomy — for example, a block specifically for handling competitor comparisons, or an internal instructions block that governs agent behavior in edge cases.

<div data-with-frame="true"><figure><img src="/files/YhBzfm2fEYbidPMidZ6I" alt=""><figcaption></figcaption></figure></div>

***

### Step 6: Reordering Blocks

The order of blocks affects how the LLM reads and weighs the instructions. The recommended order for most agents is:

Goal → Role → Objective → Guardrails → Script → Pricing → Objections

To reorder: click and drag the **≡ handle** on the left side of any block to its new position. The change is reflected immediately in the editor but does not go live until published.

<div data-with-frame="true"><figure><img src="/files/B6HMVcxGxsFz4EsPoZK2" alt=""><figcaption></figcaption></figure></div>

***

### Step 7: Publishing Changes

Changes are not live until explicitly published. The bottom bar of the editor will show **"You have unpublished changes"** with an orange indicator whenever you have edits pending.

* **Publish** — deploys your current draft to production. The agent will start using the new configuration on the next incoming call.
* **Reset** — discards all unpublished changes and reverts to the last published version.

Do not close the browser tab while changes are unpublished without first publishing or resetting.

<div data-with-frame="true"><figure><img src="/files/qVbJqLUeej7dHaTpaqeu" alt=""><figcaption></figcaption></figure></div>

***

### Step 8: Script History

Every time you publish, a full snapshot is saved automatically. The Script History panel (top right of the Script Blocks section, **Script History** button) shows the complete version log.

For each entry you can:

* **Preview** — see the full content of that version before taking any action
* **See Diff** — compare that version side-by-side with the current one to see exactly what changed
* **Restore** — roll back the agent to that version (this becomes a new unpublished draft, which you then publish)

The current live version is labeled **Current** in green.

Use Script History as your safety net. Before making significant changes to a high-volume live agent, take note of the current version timestamp so you can restore it quickly if call quality degrades.

<figure><img src="/files/6l0pifqGjYE3lrv2leD4" alt=""><figcaption></figcaption></figure>

***

### Quick Reference Checklist

* Navigate to the agent in Agent Studio
* Verify Model, Memory, and Initial Message settings
* Edit the relevant Script Block(s)
* Check the block order (Goal → Role → Objective → Guardrails → Script → Pricing → Objections)
* Review unpublished changes using Preview or Diff
* Click **Publish**
* Monitor the next set of calls in the Dashboard to confirm performance

***

*For a complete example of how all blocks work together in a real deployment, see:* [*How to Create an Effective VoiceB Prompt: An Example from NexTel*](https://docs.voiceb.ai/product-guides/how-to-create-an-effective-voiceb-prompt-an-example-from-nextel)


# Widget Configuration Guide

By Alex Bisbe. April 2026.

### Overview

The VoiceB Web Widget is the primary way to deploy AI voice agents on your website or landing page. It connects visitors directly to your SuperVoice agents via WebRTC VoIP — no phone number required, no app to install, no friction.

<div data-with-frame="true"><figure><img src="/files/ATU1cxdaXgOqZGjX8pV5" alt="" width="332"><figcaption></figcaption></figure></div>

This guide covers every configuration option available in the **Widget** section of the VoiceB platform (`app.voiceb.ai → Campaign → Widget`).

> **New to the VoiceB Widget?** Start with [New VoiceB Web Widget & WebRTC VoIP Channel](https://docs.voiceb.ai/new-voiceb-web-widget-and-webrtc-voip-channel) for the product overview and channel strategy. This article focuses on the configuration details.

***

### Installation

Deploying the widget requires a single line of JavaScript. Copy the embed snippet from your Widget Configuration page and paste it into your website's HTML:

```html
<script
    async
    src="https://widget.voiceb.ai/loader.js"
    data-widget-id="YOUR-WIDGET-ID"
/>
```

Each widget has a unique `data-widget-id`. You'll find your specific snippet in the **Setup** section of the Widget Configuration page.

<figure><img src="/files/NQw5AP0JdFmilgKkGx6G" alt=""><figcaption></figcaption></figure>

The widget works on any website, landing page, or web application — anywhere you can add a script tag.

***

### Widget Configuration Options

#### 1. Select Agents (Multi-Agent Support)

The widget supports **multiple agents simultaneously**. You can connect as many agents as you need to a single widget deployment.

In the **Select Agents** section, check the agents you want to make available through the widget. Each agent is identified by name and a language flag indicating their configured language (e.g., 🇺🇸 for English, 🇪🇸 for Spanish, 🇧🇷 for Portuguese).

**Key capability:** The widget operates at the **campaign level**, meaning one widget deployment can handle multiple agents across different languages and use cases. This makes it ideal for multi-market websites and landing pages serving diverse audiences.

***

#### 2. Agent Display Order

Once agents are selected, you can control the order in which they appear to the visitor by dragging and dropping agent cards in the **Agent Display Order** section.

Agents are organized **per language**. For example, if your widget serves English (EN-US) and Spanish (ES) visitors, you'll see two separate columns — one for each language — where you can independently set the display priority for agents within that language group.

***

#### 3. Variant (Bubble vs. Compact)

Choose how the widget appears on your page:

* **Bubble** — Displays a small floating bubble in the corner of the screen showing just your logo (or the default VoiceB icon). Minimal footprint, maximum discretion. When clicked, the widget expands to show the full agent interface.

<div data-with-frame="true"><figure><img src="/files/fBo8volIjnwLQiTksMTb" alt="" width="159"><figcaption></figcaption></figure></div>

* **Compact** — Shows a slightly larger floating element with your logo and a welcome message (e.g., "Talk to SuperVoice"). More visible and explicit, which can drive higher engagement on high-intent pages.

<div data-with-frame="true"><figure><img src="/files/FT38OflROmjSQz8lhQN8" alt="" width="263"><figcaption></figcaption></figure></div>

Both variants include the VoiceB logo icon alongside a phone icon, and display "Powered by voiceb.ai" at the bottom.

**Recommendation:** Use **Bubble** on pages where the widget supplements the experience (e.g., product pages, blog). Use **Compact** on conversion-critical pages (e.g., pricing, plan selectors, checkout flows) where you want to maximize voice engagement.

***

#### 4. Language Selector

When your widget includes agents in more than one language, the widget automatically displays a **language selector** (e.g., "EN-US | ES") so visitors can choose their preferred language.

<div data-with-frame="true"><figure><img src="/files/TijSLwzrYWASoYdyx3IT" alt="" width="375"><figcaption></figcaption></figure></div>

You can control language display behavior in **Language Display**:

* **Show language badges** — Displays language flags next to agent names.
* **Always show language badges (even with a single language)** — Forces the language selector to appear even if all agents share the same language. Useful for consistency across deployments.

The **Default Language** setting determines which language tab is pre-selected when the widget loads.

***

#### 5. Logo

Customize the widget icon by providing a URL to your own logo image. The recommended size is **64×64 pixels**. If the URL is empty or fails to load, the default VoiceB icon will be used.

You can also enable **"Hide icon when URL is empty or fails to load"** to suppress the icon entirely in case of loading errors, rather than falling back to the VoiceB default.

***

#### 6. Privacy Policy & Terms and Conditions (Compliance)

The widget supports **full regulatory compliance** for data privacy and consent collection:

* **Privacy Policy URL** — Link to your Privacy Policy page. When set, a "I accept the Privacy Policy" checkbox appears in the widget.
* **Terms and Conditions URL** — Link to your Terms and Conditions page. When set, this is shown alongside the Privacy Policy link.
* **Terms Required** — When enabled ("Require terms acceptance"), the visitor **must** check the consent box before initiating a call.

**Compliance storage:** When a visitor accepts the Privacy Policy and/or Terms and Conditions, VoiceB stores the consent record — including the document content — **up to 50 MB per document** in the database. This ensures you have a complete audit trail of what each visitor consented to, and which version of the documents was in effect at the time of consent. This is critical for GDPR, LOPDGDD, and other regulatory frameworks.

> **Important:** If neither a Privacy Policy URL nor a Terms URL is set, no checkbox or legal link will be shown in the widget.

***

#### 7. Widget Styles (Full Visual Customization)

The widget is fully customizable to match your brand identity. Every visual element can be adjusted.

<div><figure><img src="/files/Cs2aYjd4GwToF7wtT63e" alt=""><figcaption></figcaption></figure> <figure><img src="/files/CE84XLnEqhIyr6RsWYCy" alt=""><figcaption></figcaption></figure> <figure><img src="/files/A7w4GOME4eKdj9dWw9sC" alt=""><figcaption></figcaption></figure> <figure><img src="/files/lw4PUkJiY2CHuQFAV0Pu" alt=""><figcaption></figcaption></figure></div>

**Surface Palette**

Control the core appearance of the widget:

* **Background** — Main background color of the widget panel.
* **Background Hover** — Background color on hover states.
* **Border** — Border color for the widget container.
* **Muted** — Color for secondary/muted text and elements.
* **Foreground** — Primary text color.

**Brand Palette**

Set your brand colors for interactive elements:

* **Primary** — Main brand color used for buttons and CTAs (e.g., the "Talk now" button).
* **Primary Hover** — Hover state for primary elements.
* **Primary Active** — Active/pressed state for primary elements.
* **Contrast** — Text color used on top of primary-colored elements (ensures readability).

**Metrics (Sizing & Spacing)**

Fine-tune dimensions, spacing, and corner radius:

| Setting            | Description                                        | Default |
| ------------------ | -------------------------------------------------- | ------- |
| Horizontal Padding | Left/right padding inside the widget               | 20      |
| Vertical Padding   | Top/bottom padding inside the widget               | 20      |
| Main Radius        | Corner radius of the main widget panel             | 16      |
| Bubble Radius      | Corner radius of the floating bubble               | 50      |
| Bubble Padding     | Internal padding of the bubble                     | 6       |
| Button Radius      | Corner radius of CTA buttons                       | 8       |
| Inputs Radius      | Corner radius of input fields (e.g., phone number) | 8       |
| Compact Font Size  | Font size for the compact variant text             | 14      |
| Compact Padding    | Padding inside the compact variant                 | 12      |
| Compact Icon Size  | Size of the icon in the compact variant            | 36      |

**Toggles**

* **Show agent card borders** — Adds visible borders around individual agent cards.
* **Show language badge borders** — Adds borders around the language selector badges.
* **Show section dividers** — Adds horizontal dividers between widget sections.

***

#### 8. Widget Texts (Per-Language Customization)

Customize all user-facing text in the widget on a per-language basis.

Select the language you want to edit from the **Edit Language** dropdown, then configure:

**General**

* **Title** — The header displayed at the top of the widget (e.g., your company name or a tagline like "Talk to SuperVoice" / "Habla con SuperVoice").
* **Greeting Text** — The welcome message shown in the compact variant or as a subtitle (e.g., "How can we help you?").

**Agent Cards**

For each agent, you can override:

* **Heading** — The title shown on the agent's card.
* **Description** — The description text below the heading.
* **Button Label** — The CTA button text (e.g., "Talk now", "Habla ahora").
* **Alignment** — Left or center alignment for the card content.

***

#### 9. Analytics: Impressions, Clicks & UTM Tracking

The widget automatically tracks:

* **Impressions** — Every time the widget loads on a page, an impression is recorded.
* **Clicks** — Every interaction (opening the widget, clicking "Talk now", clicking "We call you for free") is tracked.
* **UTM parameters** — The widget captures UTM tags (`utm_source`, `utm_medium`, `utm_campaign`, `utm_content`, `utm_term`) from the page URL. This connects every voice conversation back to the marketing campaign that drove the visit, enabling closed-loop attribution from ad spend to voice conversion.

This data feeds into your VoiceB Dashboard alongside call outcomes, durations, and conversion rates.

***

### Preview & Testing

You can **preview and fully test the widget** directly from the Widget Configuration page — no need to deploy it to a live site first. The widget preview panel appears on the right side of the configuration screen and updates in real time as you change settings.

<div data-with-frame="true"><figure><img src="/files/NPV57QRIGFXXUJitN57a" alt=""><figcaption></figcaption></figure></div>

This allows you to:

* Test agent selection and language switching.
* Verify styling and brand consistency.
* Confirm legal consent flows work correctly.
* Make a real VoIP test call to verify agent behavior.

***

### Quick Setup Checklist

1. **Create and activate your agents** in Agent Studio (set language, voice, goal, script, objections).
2. Navigate to **Campaign → Widget**.
3. **Select agents** and set the display order per language.
4. Choose your **variant** (Bubble or Compact).
5. Upload your **logo** and set your **brand colors**.
6. Add your **Privacy Policy** and **Terms and Conditions** URLs. Enable "Require terms acceptance" if needed.
7. Customize **widget texts** for each language.
8. **Preview and test** the widget in the configuration panel.
9. Copy the **embed script** and paste it into your website HTML.
10. Go live.

***

### Related Articles

* [New VoiceB Web Widget & WebRTC VoIP Channel](https://docs.voiceb.ai/new-voiceb-web-widget-and-webrtc-voip-channel) — Product overview, channel strategy, and SuperVoice vs. phone-based agents.
* [Call Transfer: Redirect Calls to a Human Agent](https://docs.voiceb.ai/call-transfer-redirect-calls-to-a-human-agent) — How to set up VoIP-to-PSTN handoff for hybrid AI + human workflows.

***

*© 2026 Sekondary S.L. DBA VoiceB.ai. All rights reserved.*


# New VoiceB Web Widget & WebRTC VoIP Channel

We’ve launched a new VoiceB Web Widget built exclusively on WebRTC VoIP.

By Alex Bisbe. December 30th 2025.

### New VoiceB Web Widget - Executive Summary

This is not a UI change.\
It’s a **channel shift**.

From now on, the widget runs **only on VoIP (WebRTC)** and is the foundation of **SuperVoice**—our product for driving more conversions inside digital journeys.

Phone numbers remain supported, but **SuperVoice ≠ phone calls**.

<div data-with-frame="true"><figure><img src="/files/o49iST63o1fAsEUVY8x6" alt="" width="277"><figcaption></figcaption></figure></div>

<div data-with-frame="true"><figure><img src="/files/Hy0Mtle9BunrBwbEJ7Tm" alt="" width="288"><figcaption></figcaption></figure></div>

<div data-with-frame="true"><figure><img src="/files/3sQhFz1Ps71AuHeQ31qH" alt="" width="242"><figcaption></figcaption></figure></div>

\
\
Web widget.

***

### What’s New

#### 1. New VoiceB Web Widget (VoIP-Only)

The new widget:

* Works **only with WebRTC VoIP**
* Connects users instantly with an AI agent in the browser
* Replaces chat, forms, and callbacks
* Is designed for **conversion, not support**

No PSTN. No dial tones. No friction.

<figure><img src="/files/wnqZwr8IAWhw4tQFtvmm" alt=""><figcaption></figcaption></figure>

Displayed on the image above is the latest widget located in the bottom left corner.

***

#### 2. WebRTC VoIP as a First-Class Channel

With WebRTC, VoiceB agents can:

* Start conversations instantly
* Speak while the user is still on the page
* Guide decisions in real time
* Close or qualify before intent decays

This is the backbone of **SuperVoice**.

***

### Available Voice Channels in VoiceB

| Channel                  | Use Case                               | Product            |
| ------------------------ | -------------------------------------- | ------------------ |
| **WebRTC VoIP (Widget)** | Digital journeys, configurators, forms | **SuperVoice**     |
| Inbound Phone Number     | Legacy call traffic                    | SuperQualification |
| Click-to-Call (PSTN)     | Transitional setups                    | SuperQualification |
| API (coming soon)        | Programmatic voice                     | Platform           |

👉 The **Widget only works with VoIP** by design.

***

### Why WebRTC VoIP Wins vs Phone Channels

#### Phone Calls (PSTN)

* Contextless
* High drop-off
* User leaves the website
* Slow, expensive, legacy
* Built for call centers

#### WebRTC VoIP (SuperVoice)

* **User stays in the digital journey**
* Zero latency
* Zero app switching
* Higher intent
* Lower cost
* Built for **conversion at the moment of truth**

This is not an optimization.\
It’s a **replacement**.

***

### SuperVoice vs Phone-Based AI Agents

#### SuperVoice (Widget + WebRTC)

Best for:

* Car configurators
* Telecom plan selectors
* Insurance quotes
* Complex forms
* Assisted checkouts

What it does:

* Guides users step by step
* Answers objections live
* Increases completion rates
* Converts anonymous traffic into qualified demand

#### Phone Channels

Best for:

* Existing inbound traffic
* Overflow handling
* Legacy operations

Different tools. Different goals.

***

### How to Configure SuperVoice (VoIP Widget)

#### Step 1: Configure the Agent

Path:\
**Campaign → Agent → Type**

Select:

* **VoIP**

Set:

* Language
* Voice provider
* Goal
* Script
* Objections

Activate the agent.

<figure><img src="/files/cTiuJUVEL2eSs3lhVLQG" alt=""><figcaption></figcaption></figure>

***

#### Step 2: Configure the Widget

Path:\
**Campaign → Widget**

Here you can:

* Select VoIP agents
* Customize UI and brand
* Set default language and texts
* Add legal URLs
* Generate the embed script

Deploy the script on your site.

That’s it. SuperVoice is live.\ <br>

<div><figure><img src="/files/HJkeCjsui940wwOlffXa" alt=""><figcaption></figcaption></figure> <figure><img src="/files/qwrVWDyK1MRdloAavVwG" alt=""><figcaption></figcaption></figure> <figure><img src="/files/qPHzWTjanbwmRUQ8QN35" alt=""><figcaption></figcaption></figure> <figure><img src="/files/zQl4hbBSKr8vvGNw1j45" alt=""><figcaption></figcaption></figure></div>

***

### Analytics & Control

Every WebRTC conversation is tracked:

* Calls
* Duration
* Outcomes
* Conversion rate
* Agent performance

Same VoiceB dashboards.\
More signal. Less noise.\ <br>

<figure><img src="/files/ZKmadqjgX9ufAKmZZUGI" alt=""><figcaption></figcaption></figure>

***

#### 👩🏻‍💻 📲 VoIP to Phone Call Transfer (AI → Human Handoff)

VoiceB supports **hybrid call flows** where conversations **start in VoIP (WebRTC)** and are seamlessly **transferred to a human agent on a traditional phone line**.

**How it works**

* A user starts a call via the **VoiceB Web Widget (VoIP/WebRTC)**.
* The AI agent handles the conversation: greeting, intent detection, qualification, and data collection.
* When predefined conditions are met (lead qualified, request to speak to an agent, business rules), VoiceB **converts the VoIP call into a regular phone call (PSTN)**.
* The call is transferred to the call center with full conversational context.

**Why this matters**

* **Lower entry friction**: WebRTC removes the need to dial or switch devices.
* **Wider funnel**: More users engage when the first step is VoIP.
* **Smarter escalation**: Only qualified conversations reach human agents.
* **Better agent efficiency**: Humans focus on closing, not filtering.

**Recommended use cases**

* Inbound sales qualification before agent transfer
* High-traffic digital journeys (telecom, energy, automotive, finance)
* Reducing call center load while maintaining conversion quality

This approach combines the scalability of VoIP with the reliability of traditional call centers.

How to setup a transfer call: [click here](https://docs.voiceb.ai/call-transfer-redirect-calls-to-a-human-agent).

***

### Bottom Line

* The **Widget is new**
* The **Widget is VoIP-only**
* **WebRTC is the future of inbound sales**
* **SuperVoice is built for digital conversion, not phone calls**

If the user is already on your website, **making them dial a number is a failure**.

SuperVoice fixes that.


# API – Click-to-Call

Use the VoiceB API to trigger outbound calls automatically. Ideal for form submissions, lead callbacks, or click-to-call features from your website or CRM.

To enable API-triggered outbound calls in VoiceB.ai, the configuration is done **at the agent level**.

Before you can access your API credentials and endpoint:

1. Go to your agent’s configuration screen
2. Set the **Type** to `Click to Call`
3. Click **Save**
4. The **"Show Credentials & Instructions"** button will then appear

Clicking that button reveals your **API Token**, **Secret**, and full usage instructions — all specific to that agent.

> ⚠️ Each agent has its own credentials and endpoint. Make sure you configure and save the correct agent before using the API.

<figure><img src="/files/AdbAHNuP6YoKbcFsAo1J" alt=""><figcaption></figcaption></figure>

Once configured, you’ll be able to securely trigger real-time outbound calls from your backend or product flows.

***

### 🔐 API Authentication

You’ll need:

* `API Token`
* `API Secret`

> 📌 Credentials are available in  **Agent -> Show Credentials & Instructions** in your VoiceB.ai dashboard.

Always trigger calls **from your backend** to avoid exposing secrets.

{% hint style="info" %}
If you need to trigger a call from a system that can't sign the request, like Pardot, check the option **Disable Signature Verification** from agent configuration.&#x20;
{% endhint %}

***

### 📍 Endpoint

```
POST https://app.voiceb.ai/api/click-to-call
```

### 🧪 Sample Request Payload

```json
{
  "phone": "+34612345678",
  "variables": {
      "name": "John",
      "product": "VoiceB AI"
  }
}
```

***

### 🔐 Headers

* `api-key`: your API Token
* `signature`: HMAC SHA256 of the request body using your API Secret

***

#### 📦 Variables

The `variables` object in the request allows you to send data that can be utilized during interactions with the AI agent. By including specific key-value pairs, you can tailor the conversation and provide context or parameters that the AI can use to generate more accurate and relevant responses.

To utilize variables defined in the request, you need to explicitly reference them in your script using `{{variable name}}`. For example, if your request includes `"name": "John"` and `"product": "VoiceB AI"`, your script should instruct the agent with something like: "The client's name is `{{name}}` and is interested in `{{product}}`." Additionally, avoid using variables prefixed with "voiceb" or "system" to prevent conflicts with reserved system variables.

It is essential to explicitly reference all necessary variables in your script, as failing to do so may result in unsuccessful calls. The request may include any number of variables, but ensure that all required ones are clearly specified. For optional variables, it is acceptable to have them as empty strings, but they should never be undefined to avoid potential errors.

***

### 🧰 Error Codes

| **Code** | **Message**              | **Status** |
| -------- | ------------------------ | ---------- |
| ctc:001  | Missing headers          | 400        |
| ctc:002  | Error fetching agent     | 500        |
| ctc:003  | Invalid API key          | 403        |
| ctc:004  | Agent is disabled        | 400        |
| ctc:005  | Agent not configured     | 400        |
| ctc:006  | Invalid signature        | 403        |
| ctc:007  | Call already in progress | 400        |
| ctc:008  | Error initiating call    | 500        |

***

### 🔧 Code Samples

#### Node.js

```javascript
const axios = require('axios');
const crypto = require('crypto');

const API_KEY = process.env.API_KEY;
const SECRET_KEY = process.env.SECRET_KEY;
const ENDPOINT = process.env.ENDPOINT;

function generateSignature(secret, data) {
  return crypto.createHmac('sha256', secret).update(data).digest('hex');
}

async function makeRequest() {
  const body = {
    phone: '+34612345678',
    variables: {
      name: "John",
      product: "VoiceB AI"
    }
  };
  const textBody = JSON.stringify(body);
  const signature = generateSignature(SECRET_KEY, textBody);

  try {
    const response = await axios.post(ENDPOINT, body, {
      headers: {
        'api-key': API_KEY,
        'signature': signature,
        'Content-Type': 'application/json',
      },
    });

    console.log('Response:', response.data);
  } catch (error) {
    console.error('Error making request:', error.response ? error.response.data : error.message);
  }
}

makeRequest();
```

#### Python

```python
import os
import hmac
import hashlib
import requests
import json

API_KEY = os.getenv('API_KEY')
SECRET_KEY = os.getenv('SECRET_KEY')
ENDPOINT = os.getenv('ENDPOINT')

def generate_signature(secret, data):
    return hmac.new(secret.encode(), data.encode(), hashlib.sha256).hexdigest()

def make_request():
    body = {
        'phone': '+34612345678',
        'variables': {
            'name': 'John',
            'product': 'VoiceB AI'
        }
    }
    text_body = json.dumps(body)
    signature = generate_signature(SECRET_KEY, text_body)

    headers = {
        'api-key': API_KEY,
        'signature': signature,
        'Content-Type': 'application/json',
    }

    response = requests.post(ENDPOINT, headers=headers, data=text_body)
    print('Response:', response.json())

if __name__ == '__main__':
    make_request()
```

***


# Click-to-Call Retry Policy

By Alex Bisbe. May 5th 2026.

Click-to-Call agents in VoiceB place an outbound callback the moment a lead is submitted (form fill, widget request, CRM trigger, etc.). The first attempt connects most leads, but a meaningful share will not pick up the first time — they're driving, in a meeting, or simply away from their phone.

The Retry Policy lets your Click-to-Call agent automatically dial the same lead again under conditions you control: how many times, how long to wait between attempts, which failure types should trigger a retry, and which hours and days are allowed.

The goal is simple: **maximize contact rate without violating local outbound calling rules.**

***

### Why this matters

Every lead that goes uncontacted is paid-for demand that never converts. Industry data and our own production deployments show that a single dial attempt typically reaches only 30–50% of leads, while three to four well-spaced attempts can lift contact rates above 70%. That difference flows directly into your conversion rate and your CPA.

At the same time, outbound calling is regulated. Spain, the UK, the US, and most other markets restrict when calls can be placed (typically business hours, weekdays), how aggressively a number can be re-dialed, and how voicemails or no-answers are handled. The Retry Policy is designed so you can be aggressive on contact rate while staying inside those rules by default.

***

### How retries work

When the first call attempt ends in a failure outcome (for example, the lead does not answer), VoiceB checks the retry configuration:

1. Is the failure outcome one of the configured retry triggers?
2. Have we reached the maximum number of attempts?
3. According to the backoff strategy, when is the next attempt due?
4. Does that next attempt fall inside business hours?

If all conditions are met, VoiceB schedules the next call automatically. If the next attempt would fall outside business hours, it is snapped forward to the next valid window. The retry counter advances on every attempt, and every call is logged independently in the dashboard with its retry index visible.

***

### Configuration

The Retry Policy is configured per agent, on the agent settings page.

<figure><img src="/files/vgHfH22jYyO9XFzNAt2s" alt=""><figcaption></figcaption></figure>

#### Enable retries

A single toggle activates the retry logic. When disabled, Click-to-Call agents make exactly one attempt per lead and stop.

#### Max attempts

The total number of dial attempts allowed per lead, **including the original call**. A value of 4 means: 1 first call + up to 3 retries. We recommend 3 to 5 for most B2C use cases. More than 5 attempts on the same number is rarely productive and can raise compliance concerns in stricter markets.

#### Retry on

Defines which failure outcomes should trigger a retry. Available outcomes:

* **no answered** — the lead did not pick up
* **busy** — the line was busy
* **voicemail** — the call went to voicemail
* **failed** — the call did not connect at the carrier level
* **dropped** — the call dropped mid-connection
* **unreachable** — the number could not be reached at all

A typical configuration retries on `no answered`, `busy`, and `voicemail`, and does not retry on `failed`, `dropped`, or `unreachable` — those usually indicate a number-level problem that more attempts will not solve. Adjust based on what your campaign data shows.

#### Backoff strategy

<div data-with-frame="true"><figure><img src="/files/bEzCYTg4mclifgMDC59H" alt=""><figcaption></figcaption></figure></div>

Controls how long the system waits between attempts. Two options:

**Exponential**

The wait time grows after each attempt by a fixed factor, up to a maximum delay.

* *Initial delay (min):* time between the first call and the first retry. Default 15.
* *Factor:* multiplier applied after each attempt. Default 2 (15 → 30 → 60 → 120…).
* *Max delay (min):* upper bound on the wait. Default 1440 (24 hours).

Exponential is the right default for most lead-gen flows: hot leads get a fast retry while the intent is still warm, then attempts space out as the lead cools.

**Custom sequence**

You define each delay manually as a comma-separated list of minutes, e.g. `15, 60, 240, 1440`. The system uses each value in order; once the list is exhausted, the last value is reused for any further attempts. Useful when you want a precise cadence — for example, retry quickly within the first hour, then once that evening, then once the next morning.

<div data-with-frame="true"><figure><img src="/files/uDMKkhAOHEMrnwAS8E4S" alt=""><figcaption></figcaption></figure></div>

#### Business hours

Outbound calling outside business hours is illegal or restricted in most markets. The Business Hours block ensures every retry lands inside an allowed window.

* **Respect business hours:** when enabled, any retry scheduled outside the window is moved forward to the next valid time slot.
* **Start hour / End hour:** the daily window, expressed in the **local timezone of the campaign** (not your browser timezone).
* **Weekdays only:** when enabled, retries scheduled for Saturday or Sunday are pushed to the following Monday.

A Spain campaign configured 9:00–20:00, weekdays only, will never place a retry on a Saturday morning or a Tuesday at 8:30 — even if the backoff math says it should.

***

### Monitoring retries on the dashboard

<div data-with-frame="true"><figure><img src="/files/fKcnA85YYis0i2FfstNa" alt="" width="320"><figcaption></figcaption></figure></div>

Every retry is recorded as its own call in the Campaign Dashboard, with the retry index visible alongside the outcome:

* `retry 1/2` — first retry, two retries allowed
* `retry 2/2` — second and final retry
* `retry 1/3` — first retry, three retries allowed

This makes it straightforward to see, per lead, how many attempts were made and what outcome each one produced. You can filter calls by phone number to view the full retry chain for any single lead, or filter by outcome to spot patterns (for example, leads that consistently hit voicemail across all attempts may indicate a bad data source rather than an agent problem).

***

### Recommended starting points

Three configurations that work well in production. Tune from there based on your campaign data.

**Standard B2C lead-gen (Spain, UK, most of EU)**

* Max attempts: 4
* Retry on: no answered, busy, voicemail
* Backoff: Exponential — initial 15, factor 2, max 1440
* Business hours: 9:00–20:00, weekdays only

**High-intent / hot leads (cart abandon, recent form fill)**

* Max attempts: 3
* Retry on: no answered, busy
* Backoff: Custom sequence — `5, 30, 180`
* Business hours: 9:00–21:00, weekdays only

**Low-volume / high-value (insurance, financial products)**

* Max attempts: 5
* Retry on: no answered, busy, voicemail
* Backoff: Custom sequence — `15, 60, 240, 1440, 2880`
* Business hours: 10:00–19:00, weekdays only

***

### Compliance notes

The Retry Policy is built so that the default behavior stays inside outbound calling rules in our primary markets, but **you are responsible for confirming the configuration against your own legal requirements**. In particular:

* Some jurisdictions cap the total number of attempts on a single number within a given period.
* Some markets require an opt-out path that survives across retry attempts.
* Voicemail handling rules vary by country.

If you are deploying in a new market, review the Retry Policy with your legal or compliance team before activating the agent. The Business Hours and Weekdays Only controls are the primary technical safeguards, but they are not a substitute for a compliance review.

***

### Related

* [Outcome Engine and how failure outcomes are classified](https://docs.voiceb.ai/understanding-status-and-outcomes-in-voiceb)
* [Click-to-Call API — triggering the first call from your CRM or website](https://docs.voiceb.ai/api-click-to-call)


# Understanding Status & Outcomes in VoiceB

Understanding Status & Outcomes in VoiceB

By Alex Bisbe. Last update: March 2026

VoiceB classifies every call into a **Status** and an **Outcome**. Together, they define *lead quality*, *intent*, and the *next operational step*. This structure powers reporting, billing, routing, and your downstream integrations (Salesforce, HubSpot, Webhooks).

***

### 1. Status Categories

Each call receives one of two high-level statuses:

<div align="left" data-with-frame="true"><figure><img src="/files/m7M2CWolwW6Q1g1bG93v" alt="" width="97"><figcaption></figcaption></figure></div>

**✅ Valid** The AI Agent reached a real person, the conversation had intent, and the lead fits your commercial scope.

**❌ Invalid** The call didn't result in a commercially meaningful conversation. No sales opportunity is created.

***

### 2. Outcomes for INVALID Status

Invalid leads are automatically filtered out. These calls **do not count** toward sales metrics and are excluded from revenue-based pricing.

<div align="left" data-with-frame="true"><figure><img src="/files/L77AHokTaNN9zXqz67K8" alt="" width="129"><figcaption></figcaption></figure></div>

#### 1. Not Interested

The lead is not a target, not interested in the offer, or explicitly rejects the product. Examples:

* "I'm not interested."
* "This is not for me."
* Wrong product segment.
* Wrong target audience.

#### 2. Voicemail

The AI reached a voicemail system. No conversation took place.

#### 3. Dropped

The call lasted **under 15 seconds** and had no meaningful customer interaction. Typical cases:

* Hang-ups
* Accidental dials
* Early disconnections

#### 4. Customer Care

The call is valid but **out of sales scope**. The caller is an existing customer seeking support, billing, technical help, or account management — not a new purchase or sales interaction. No sales qualification or conversion is possible, so the call is marked Invalid for commercial KPIs.

<div data-with-frame="true"><figure><img src="/files/T5xphT0DK8mzihJxAqoc" alt=""><figcaption></figcaption></figure></div>

#### 5. Busy

The caller's line was busy or engaged at the time of the call. No conversation took place.

#### 6. Unreachable

The call connected but was never answered. The number exists but no one picked up.

#### 7. Not Found

The phone number dialed was invalid or could not be found. The call could not be connected.

#### 8. Failed

A technical or system error prevented the call from connecting. This is an infrastructure-level outcome, not a commercial one. Failed calls should be monitored for patterns — a spike may indicate a telephony configuration issue.

#### 9. Pending

The call was still in progress at the moment the CSV export was generated. Pending is a **snapshot state**, not a final outcome. If you see Pending calls in an export, re-export after a few minutes to retrieve the final outcome once the call has completed.

***

### 3. Outcomes for VALID Status

Valid leads indicate a real opportunity. VoiceB uses two primary outcomes inside Valid.

<div align="left" data-with-frame="true"><figure><img src="/files/yvurIuSC8kKAlcI1UoHk" alt="" width="238"><figcaption></figcaption></figure></div>

#### 1. Sold

The customer agrees to move forward *and* provides the required personal data to complete the sale. This is the strongest commercial outcome.

Sold is triggered when:

* The AI successfully pitches the offer.
* The customer accepts.
* The required personal data is collected (at minimum: the fields configured as mandatory for your process).

#### 2. Talk to Agent

The lead meets your qualifying criteria but cannot be fully closed by the AI. Typical situations:

* Customer requires a final human confirmation.
* Complex cases needing agent validation.
* The customer is interested but does not provide all required data.
* AI cannot finalize the process due to business rules.

This outcome means **valid lead, warm intent**.

***

### 4. Transferred (Real-Time Handoff)

Both **Sold** and **Talk to Agent** can carry an additional label: **Transferred**.

<div align="left" data-with-frame="true"><figure><img src="/files/UoSIpDWLAvQReCtvY5n4" alt="" width="220"><figcaption></figcaption></figure></div>

This appears only when real-time call transfer is enabled for your account. It indicates that the call was successfully routed to a live agent during the conversation.

Examples:

* Valid → Talk to Agent → Transferred
* Valid → Sold → Transferred (only if your flow includes a verification step)

***

### 5. Why Status & Outcomes Matter

This classification enables:

* **Accurate KPIs** (conversion, agent handoff rate, sales rate)
* **Precise billing models** (only Valid outcomes count)
* **Lead routing** (real-time transfers, CRM pushes, HubSpot/Salesforce sync)
* **Operational clarity** (support, sales teams, QA)

Data from both Status and Outcomes is synced via Salesforce, HubSpot, Webhooks, and Custom CRM Connectors.

***

### Summary Table

| Status    | Outcome        | Meaning                                    | Billing  |
| --------- | -------------- | ------------------------------------------ | -------- |
| ❌ Invalid | Not Interested | No intent, wrong target, rejected offer    | Excluded |
| ❌ Invalid | Dropped        | <15s, no interaction                       | Excluded |
| ❌ Invalid | Voicemail      | Reached voicemail                          | Excluded |
| ❌ Invalid | Customer Care  | Out of sales scope                         | Excluded |
| ❌ Invalid | Busy           | Caller's line was busy/engaged             | Excluded |
| ❌ Invalid | Unreachable    | Number valid but never answered            | Excluded |
| ❌ Invalid | Not Found      | Phone number invalid or not found          | Excluded |
| ❌ Invalid | Failed         | Technical error, call could not connect    | Excluded |
| ❌ Invalid | Pending        | Call still in progress during CSV export   | Excluded |
| ✅ Valid   | Sold           | Customer accepts + required data collected | Billed   |
| ✅ Valid   | Talk to Agent  | Valid lead needing human follow-up         | Billed   |

***

### Tags (Custom Outcomes)

VoiceB also supports **Tags** — a fully customizable second classification layer applied automatically by the AI after understanding the full conversation.

<figure><img src="/files/VRj73IGYfqy6MoWBZokv" alt=""><figcaption></figcaption></figure>

While Status explains if a call was successful and Outcome explains what happened, Tags explain what the call means for your business. Use them to classify calls by product interest, customer type, or intent, enrich your CRM, and route leads with greater precision — without affecting existing reporting.

<figure><img src="/files/amtVmoix95pGO6eIFxPT" alt=""><figcaption></figcaption></figure>

👉 Learn more about Tags and how to configure them [here](https://docs.voiceb.ai/tags-custom-ai-classification-for-calls-and-leads).

***


# Tags — Custom AI Classification for Calls & Leads

By Alex Bisbe. Last update: January 26th 2026

Tags turn VoiceB into a **semantic classifier**. The AI understands each call end-to-end and **auto-labels it using your custom tags**. This gives us a **second, fully customizable outcome layer** on top of standard call results. Net effect: cleaner lead qualification, smarter routing, and higher signal pushed to CRM and agents.

***

### What problem Tags solve

Traditional outcomes (Valid, Invalid, Transferred, Dropped, etc.) are necessary but generic. Tags let us **encode business logic**—products, intents, customer type, ambiguity—directly into the call. The AI applies these tags in real time based on what the caller actually says.

**Bottom line:** more precision, less manual work, better decisions downstream.

***

### How Tags work (at a glance)

1. **Define tags** (unlimited): name, condition for *true*, condition for *false*, and color.
2. **AI listens to the call** and evaluates intent and context.
3. **Tags are automatically applied** to the call.
4. **Tags appear everywhere**: dashboard, call detail, exports, CRM, routing logic.

Think of Tags as **custom outcomes tailored to your business**.

<figure><img src="/files/KywTwrQA9cWqKJKyEfJY" alt=""><figcaption></figcaption></figure>

***

### Common use cases

* **Lead classification**: Mobile, Fiber, Convergent, SME, Enterprise
* **Customer filtering**: Existing customer vs new prospect
* **Intent clarity**: No clear interest, comparison shopping, price-driven
* **Routing**: Send only “Mobile + High Intent” calls to sales
* **CRM enrichment**: Push structured intent data, not raw transcripts

***

### How to create Tags

Go to **Choose Campaign → Tags** and add as many as needed.

For each tag, define:

* **Tag name** (e.g. *Mobile*, *Home Internet*)
* **True if**: when the AI should apply the tag
* **False if**: when it should explicitly not apply
* **Color**: visual clarity in the dashboard

You’re not training a model. You’re **describing intent in plain language**. VoiceB handles the rest.

***

### How Tags appear in the Dashboard

Once enabled, Tags are visible directly in the **Calls table** and **Call detail view**:

* Multiple tags can be applied to a single call
* Tags coexist with standard outcomes (Valid, Invalid, Transferred, etc.)
* Visual color coding makes patterns obvious at scale

This allows instant analysis like:

* *Valid + Mobile + Fiber*
* *Invalid + Customer Care*
* *No Clear Interest + Dropped*

<figure><img src="/files/GnfpXs6NvkqhUcyh1iev" alt=""><figcaption></figcaption></figure>

***

### Tags as a “second outcome”

Outcomes answer **what happened**.\
Tags answer **what this call means**.

That’s the key shift.

With Tags:

* Outcomes stay standardized
* Classification becomes **fully custom**
* Routing and CRM logic becomes deterministic

In practice, Tags act as a **business-specific outcome layer** without breaking reporting or analytics.

<div data-with-frame="true"><img src="https://docs.voiceb.ai/~gitbook/ogimage/ZnL8VKgTZdHeV6eyOC9w" alt="Understanding Status &#x26; Outcomes article" width="375"></div>

***

### CRM & routing integration

Tags can be used to:

* Populate CRM fields automatically
* Trigger routing rules (only transfer if Tag = X)
* Filter reports and exports
* Inform agents before they even pick up the call

This is especially powerful in high-volume inbound environments where **every unnecessary transfer costs money**.

***

### Best practices

* Start with **high-level intent tags** (Product, Customer Type)
* Avoid over-engineering on day one
* Use Tags to **reduce ambiguity**, not describe everything
* Review tag distribution weekly to refine definitions

***

### Why this matters

Tags move VoiceB from call handling to **decision intelligence**.

We’re not just answering calls.\
We’re structuring intent, at scale, automatically.

That’s what unlocks:

* Higher conversion
* Lower operational cost
* Cleaner data
* Fully autonomous inbound sales

If you want to push Tags into routing or CRM logic, that’s already supported.


# Webhook – Call Outcome Integration

## VoiceB.ai Webhook – Call Outcome Integration

Webhooks let you receive structured call outcomes, transcripts, and lead data after every call. Use this to:

* Save conversations to your database
* Trigger follow-ups or automations

***

### 🔗 How it Works

1. Configure your webhook URL in **Settings → Integrations**
2. Optionally enable **signature signing** for secure verification
3. VoiceB will POST call result data in real-time

***

### How to Add a Webhook Integration

1. **In your VoiceB account**, go to **Campaigns**.
2. Select the **Agent** you want to configure.
3. Once inside the **Agent Config** (Name, AI Provider, Goal, Script…), scroll down to **Integrations**.
4. Click on the **“+”** button.
5. Choose **Webhook** from the list.
6. You will now see the configuration screen with instructions and examples.

<figure><img src="/files/gWB5EWkyAHorRy5N89a6" alt="" width="563"><figcaption></figcaption></figure>

<figure><img src="/files/CwqmK6gF6Zh8qOC4wTQN" alt="" width="563"><figcaption></figcaption></figure>

<figure><img src="/files/ddABEFJNtF5k7rrRq4GR" alt="" width="563"><figcaption></figcaption></figure>

***

### 📬 Sample Webhook Payload

```json
{
  "agent": "AgentName",
  "from": "+1234567890",
  "to": "+0987654321",
  "status": "SOLD",
  "transcript": [...],
  "collectedData": {
    "name": "John Doe",
    "email": "john.doe@example.com",
    "address": "Calle Ejemplo 123, 45678, Ciudad"
  }
}
```

***

### 🛡️ Signature Verification (Optional)

Header:

```ini
Authorization=Bearer XXXXXXXXXXXXXXXXXX
x-webhook-signature=YYYYYYYYYYYYYYYYYYY
```

Use your **signature secret** to verify the authenticity of the payload.

***

### Trigger Conditions

By setting trigger conditions, you can define specific outcomes that will determine if the integration runs for that call or not. Allowing you to determine which ones ends in your system/CRM

<figure><img src="/files/7m39kocr6LBInN7i7XpG" alt=""><figcaption></figcaption></figure>

### 🧰 Error Codes (Same as API)

***

If you need help setting this up, reach out to <hello@voiceb.ai>

Would you like a diagram to visualize the full API/webhook data flow?&#x20;


# Create a Data Collection Schema

Updated on July 10th 2026

#### 🧾 How to Create a Data Collection Schema in VoiceB.ai

VoiceB.ai lets you collect structured data from users during calls, making it easy to capture key information like names, addresses, product preferences, and more. Here's how to set up your own **Data Collection Schema** — and how it integrates with your CRM or webhook.

***

**✅ Step-by-Step: Creating a New Schema**

**Go to "Data Collection >> New Schema"** Click the **+ New Schema** button to start a new data set.

<div data-with-frame="true"><figure><img src="/files/302XRt6kn9Ua0aoI6nij" alt="" width="251"><figcaption></figcaption></figure></div>

<div data-with-frame="true"><figure><img src="/files/vsGG0Ewfk75JMjHG1ELZ" alt=""><figcaption></figcaption></figure></div>

1. **Give it a Name and Description**
   * Example:
     * Name: `Basic personal data`
     * Description: `Default data collection for Telco customers, includes name, phone, email, etc.`
2. **Add as Many Fields as You Need**\
   For each field, enter the following:

   * **Field name** – e.g., `email`
   * **Format** – e.g., `text`, `number`, `date`
   * **Description** – what the field is for (e.g., `Customer email address`)
   * **Example** – e.g., `jane.doe@email.com`
   * **Required** – toggle ON if this field is mandatory for the sale to be considered complete

   <div data-gb-custom-block data-tag="hint" data-style="warning" class="hint hint-warning"><p><strong>Every schema must have at least one Required field.</strong> You won't be able to save a schema where all fields are optional.</p></div>
3. **Click the Arrow (⬆️) to Add the Field**
   * You can repeat this process and add unlimited fields before saving.
4. **Click Save**
   * Once saved, the schema is locked and cannot be edited. If you need changes, create a new schema (see Updating a Schema below).

***

**🎯 Required Fields & the SOLD Outcome**

Required fields are directly tied to how VoiceB classifies call outcomes:

* A call is only classified as **SOLD** when the agent has collected **all Required fields** in the assigned schema.
* If any Required field is missing at the end of the call, the call will **not** be classified as SOLD — even if the customer verbally agreed to the purchase.
* Optional fields never block the SOLD outcome. Use them for nice-to-have data (preferences, referral source, etc.).

{% hint style="info" %}
**Design tip:** Mark as Required only the fields that are genuinely essential to process the sale (e.g., full name, phone, contract address). Every extra Required field adds friction to the conversation and can reduce your close rate.
{% endhint %}

***

**🚫 Agents Without Data Collection**

Data collection is optional at the agent level. If an agent doesn't need to capture structured data (e.g., a pure qualification or routing agent), simply **don't assign any schema** in the agent's configuration. The agent will run normally without data collection, and outcomes will be classified according to the agent's own logic.

<div data-with-frame="true"><figure><img src="/files/EXhNUcfZ4MNIqHYI317g" alt=""><figcaption></figcaption></figure></div>

***

**🔄 Updating a Schema: Create a New One**

Schemas are **immutable** once saved. This is by design: it guarantees performance and accuracy, because every historical call remains linked to the exact schema version it was collected under.

To change a schema:

1. **Create a new schema** with the updated fields.
2. **Reassign it in each agent** that should use it (agent configuration → Data Collection).
3. The old schema stays intact for historical data — nothing breaks retroactively.

Never work around this by trying to "reuse" a schema for a different purpose. A clean new schema per change keeps your data consistent and your outcome classification reliable.

***

**📌 Using the Schema**

Once your schema is created, you can **assign it to any AI Agent**:

* Go to the agent's configuration screen
* Select the desired schema under **Data Collection**
* This tells the agent what to collect during each conversation

***

**🔗 Integration & Post-Call Automation**

* If you're using **CRM integrations (like HubSpot or Salesforce)**:
  * Your data fields will appear in the **mapping section**, ready to sync automatically.
* If you're using **webhooks**:
  * You'll receive **all collected data**, plus:
    * Post-call **transcript**
    * **Summary**
    * And every **data field collected**

This ensures you get the full picture from every conversation — structured, synced, and ready to use.

***

{% hint style="info" %}
**Pro Tip:** Use descriptive names and examples to avoid confusion, especially when multiple team members are managing different schemas.
{% endhint %}

Need help structuring complex data? **Reach out to our team** — we'll help you build the right schema for your flow.

***

***


# Memory - Conversation Recall

By Alex Bisbe. February, 17th 2026.

Persistent commercial continuity — agent by agent.

Memory allows your agent to securely resume conversations exactly where they stopped.

Instead of restarting from zero when a prospect calls again, the agent:

* Recognizes the phone number
* Requests name validation (GDPR compliance)
* Recovers the previous summary
* Continues from the last completed step

This is a structural UX upgrade — not just a feature.

***

## Why Memory Matters (UX & Revenue Impact)

Without Memory:

* Users repeat information
* Agents re-ask already confirmed details
* Momentum is lost
* Sales cycles extend
* Conversion drops

With Memory:

* Context persists
* Conversations feel human
* The commercial journey continues naturally
* Sales cycles shorten

This is especially powerful in:

#### B2B

Where decision cycles span days or weeks.

#### Energy, Telco, Insurance, Financing

Where customers compare before committing.

Memory removes friction and preserves intent.

***

## 🔐 Data Protection & Validation Logic

Memory is secure by design.

If a caller has contacted before, the agent must say:

> “I see that you called before. For data protection reasons, can you confirm your name so we can resume the conversation?”

#### If the name matches:

* The agent greets the caller by name.
* Summarizes the previous conversation.
* Continues exactly where it stopped.
* Avoids repeating confirmed information.

The agent remembers:

* Price discussed
* Tariff selected
* Surname
* Objections
* Last completed step

#### If the name does NOT match:

* The system assumes it is a different person.
* The conversation resets.
* The agent starts almost from zero.

This prevents cross-user data exposure when multiple people share a phone number.

***

## Already a Customer Logic

If after validation previous calls show the user is already a client:

* Redirect to Customer Support for service issues.
* If they want a new contract, resume the commercial process from the last completed step.

***

## How to Activate Memory (Agent by Agent)

Memory is configured individually per agent.

To activate:

1. Go to **Agent → Config**
2. Find **Memory (days)**
3. Enter the number of days
4. Click **Save**

If set to:

* `0` → Disabled
* `14` → Remembers last 14 days
* `30` → Remembers last 30 days

Each agent can have a different memory window depending on its objective.<br>

<figure><img src="/files/6BtAm935Ti9guhoVquA1" alt=""><figcaption></figcaption></figure>

***

## Recommended Memory Windows

Choose the number of days based on your sales cycle.

#### B2B

Recommended: **15 to 90 days**

#### High-consideration B2C

Recommended: **15 to 30 days**

#### Fast-decision B2C

Recommended: **7 to 15 days**

***

### ⚠ Important: Pricing & Promotion Expiry

If:

* Prices change frequently
* Promotions expire quickly
* Conditions vary weekly

Use a shorter Memory window.

The agent will remember previously discussed pricing.\
If pricing has changed, you must either:

* Add prompt logic to revalidate pricing upon resume
* Or reduce the memory window

This is critical for:

* Energy tariffs
* Telecom promotions
* Limited-time offers

***

## Dashboard: Viewing Conversation Chains

When the same customer calls multiple times within the memory window, the Dashboard shows:

* **Call Details 1/2**
* **Call Details 2/2**
* Navigation buttons (Next / Previous)

Each call has its own Call ID, but the interface lets you navigate across the conversation chain.

<figure><img src="/files/l4BE6NjdujtzTu1Hs5LB" alt=""><figcaption></figcaption></figure>

This allows you to:

* Follow the evolution of the sale
* Understand how objections progressed
* See how the deal was closed
* Review summaries incrementally

Memory enhances the live interaction — the dashboard lets you audit the journey.

***

## CRM & Post-Call Data

Memory does NOT merge calls internally.

Each call still generates:

* Transcript
* Summary
* Structured outcomes
* Tags
* Status (Valid, Sold, etc.)

After each call, post-call data is sent automatically:

* To **HubSpot**
* To **Salesforce**
* Or via **Webhook** to any external system

Depending on your integration setup:

* You can create a new Lead
* Update an existing Lead
* Add a Task/Activity per call

Each call remains traceable and independent at CRM level.

Memory affects the conversation experience — not the reporting logic.

***

## Script Requirement for Proper Behavior

You must include Memory-aware logic in your prompt:

```
# PREVIOUS CALL CONTEXT

If the caller has previously contacted us:

Say:
“I see that you called before. For data protection reasons, can you confirm your name so we can resume the conversation?”

If the name matches:
- Greet them by name.
- Summarize previous conversation.
- Continue exactly from the previous step.
- Avoid repeating confirmed information.

If the name does not match:
- Restart the conversation from the beginning.
```

***

## What Makes This Unique

This is not simple context recall.

It required:

* Phone-number-based session linking
* Controlled summary injection
* Name validation gating
* Secure reset logic
* Dashboard navigation support
* CRM-safe post-call isolation

It moves VoiceB from stateless automation to persistent commercial memory.

That shift increases:

* Trust
* Intelligence perception
* Conversion probability
* Sales efficiency

This is foundational infrastructure for autonomous revenue execution.

***


# Using Variables in Click to Call to Personalize Conversations

With our new **Click to Call Variables** feature, you can now pass extra customer information directly from your **Click to Call form** into the AI conversation. This allows your AI agent to greet customers by name, mention their product of interest, or adapt responses based on other attributes like **country, age, or any custom field** you choose.

This is a powerful way to create more **personalized, relevant, and effective conversations**.

***

### How It Works

When a customer submits your **Click to Call form**, you can send additional data as part of the request. This information is passed in the `"variables"` object, and the AI agent can use it dynamically during the call.

#### Example Request Body

```json
{
  "phone": "+34xxxxxxxxx",
  "variables": {
    "name": "John",
    "last_name": "Doe",
    "product": "VoiceB AI",
    "country": "Spain",
    "age": "32"
  }
}
```

In your agent’s script, you can then call any of these variables using **`{{variable_name}}`** syntax.

**Example in Script:**

```
Hello {{name}}, I see you are interested in {{product}}.  
Let’s go over the details together.
```

***

### Setup Instructions

1. **Login** to your VoiceB account.
2. **Select your campaign** from the left-hand menu.
3. **Select the agent** you want to configure.
4. Go to **Configuration > Click to Call**.
5. Click the **Show Credentials & Instructions** button.
6. Copy your **API Token, Secret, and Endpoint**.
7. When sending your POST request to the Click to Call API, include the `variables` object with any customer data you want available during the call.

<figure><img src="/files/qxavUW7RF0EJ0Dr0Lvzg" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/ePUi3wZDWL2MDihzPwcV" alt=""><figcaption></figcaption></figure>

***

### Example Use Cases

* **Personalized Greetings**: Greet the customer by name.
* **Product Relevance**: Mention the specific product they selected on your site.
* **Regional Adaptation**: Adapt the conversation depending on the customer’s country or language.
* **Age-Targeted Messaging**: Offer different plan recommendations based on the age group.
* **Custom Data Fields**: Any other field you include in `variables` can be used in real time during the call.

***

### Technical Notes

* Always send requests from your **backend** to keep your credentials secure.
* You can pass as many variables as you want inside the `"variables"` object.
* Use `{{variable_name}}` in your script to insert these values dynamically.

***

✅ **With this new feature, every Click to Call conversation can now feel like a tailored, one-to-one experience.**


# Call Transfer: Redirect Calls to a Human Agent

Last update: by Alex Bisbe on December 30th 2025.

Our AI agent can now seamlessly **transfer calls to a phone number** (such as a call center agent) depending on the conversation.

This feature is ideal when:

* The user prefers to speak with a human.
* The sales request is out of the AI’s defined scope.
* The call is part of a validation or compliance process.

When a call is transferred, you’ll see new **outcome tabs** in the call logs (e.g., *Talk to Agent*, *Transferred*), making it easy to track results (see screenshots ).\ <br>

<figure><img src="/files/kSS0BL6rBVXwLmGNSe1S" alt="" width="356"><figcaption></figcaption></figure>

<figure><img src="/files/r1cfalnwPkyVbM2bNk76" alt="" width="271"><figcaption></figcaption></figure>

***

### How to Set Up Call Transfers

Follow these steps to configure call transfers for your agent:

1. **Go to the Agent Configuration page.**\
   In the **Call Transfers** section, click **New Transfer**.
2. **Add a Transfer Name.**\
   Example: *Billing Department*.
3. **Set the DDI (phone number).**\
   Enter the phone number where the call should be transferred.
4. **Define the Condition.**\
   Write the condition that must be met for the AI to transfer the call.\
   Example: *“The customer asks to speak with the billing department.”*
5. **Add multiple transfers if needed.**\
   You can configure as many transfers as required for different scenarios.
6. **Click Save.**\
   Your transfer rules will now be active.

<figure><img src="/files/d9csNS5cAx5nii8Ijpzs" alt=""><figcaption></figcaption></figure>

***

### Provider Compatibility

✅ **Works with both ElevenLabs and Bland** – so you don’t need to worry about balancing between AI providers. The transfer logic applies equally to both.

***

#### 📲 New: Transfer VoIP Calls to a Human Agent via Telephony

VoiceB now supports **transferring VoIP-generated calls to human agents or call centers using traditional telephony (PSTN)**.

This enables a **hybrid call model** where conversations start in **VoIP/WebRTC + AI** —low friction, browser-native—and are escalated to **human agents on regular phone lines** once qualification criteria are met.

**How it works**

* The call starts via **VoIP (WebRTC)** from the widget or inbound VoIP channel.
* The AI agent handles greeting, intent detection, qualification, and data collection.
* When a transfer condition is met (qualified lead, user request, business rule), VoiceB **converts the VoIP call into a standard phone call**.
* The conversation is transferred to a **human agent or call center** with full context.

**Why this matters**

* **Lower friction at entry**: VoIP increases engagement and funnel volume.
* **Smarter escalation**: Only qualified conversations reach human agents.
* **Operational efficiency**: AI absorbs low-value traffic; humans focus on closing.
* **Enterprise-ready**: Works with existing call center infrastructure.

**Typical use cases**

* Inbound sales qualification before agent handoff
* High-volume digital journeys (telecom, energy, automotive, finance)
* Reducing call center load without sacrificing conversion

***

### Example

* **User:** “I want to talk to an agent.”
* **AI:** “Please hold, I’ll transfer you to an agent.”
* **System:** Transfers to the configured DDI.

Outcome in dashboard: **Valid → Talk to Agent → Transferred**.


# Test Numbers

By Alex Bisbe. October 22nd, 2025.

#### ✅ Overview

You can now **add test phone numbers** to your organization to safely test agents, call flows, and prompts — without affecting your analytics or campaign statistics.

This feature is perfect for **QA teams, demos, or training sessions**, ensuring that internal tests don’t distort your real conversion or performance data.

<figure><img src="/files/0b3afmOA4W00ir84wIVJ" alt="" width="175"><figcaption></figcaption></figure>

***

#### ⚙️ How It Works

When a number is added as a **Test Number**, all calls from or to that number are automatically **excluded** from:

* Conversion Rate
* Orders and Valid Calls
* Agent Performance Reports
* Campaign and Monthly Analytics

You can still **see these calls in the logs** by toggling **Show Test Calls** at the bottom of your dashboard.

***

#### 🪜 How to Add a Test Number

1. Navigate to **Settings → Organization Settings → Test Numbers**
2. Enter the phone number in international format (e.g. `+34XXXXXXXXX`)
3. (Optional) Add internal notes such as *“Demo QA number”* or *“Marketing test line”*
4. Click the **➕ button** to save the number

<figure><img src="/files/PuiKXzlpEwbiHDZtGa1o" alt=""><figcaption></figcaption></figure>

Your new test number will appear in the list, with the creation date and edit/delete actions available.

***

#### 🧹 Managing Test Numbers

| Action                       | Description                        |
| ---------------------------- | ---------------------------------- |
| ✏️ **Edit**                  | Modify notes or update the number  |
| 🗑️ **Delete**               | Remove a test number               |
| 🕒 **Created / Last Edited** | Track when it was added or updated |

***

#### 📊 Filtering Analytics

At any time, you can choose to **show or hide test calls** in analytics:

> **Dashboard → Show Test Calls**

* **ON** → Displays all data, including test calls
* **OFF** → Excludes all test numbers (recommended for reports and performance tracking)

<figure><img src="/files/razhYZrfFWhWxGU8ZE63" alt="" width="375"><figcaption></figcaption></figure>

***

#### 🏢 Where It’s Configured

This configuration is **company-wide**, meaning it applies to all your campaigns.

To access:\
**Organization → Configuration → Test Numbers**

All changes take effect immediately across the platform.

***

#### 💡 Best Practice

Before running a demo or internal validation, add your QA or team numbers as Test Numbers.\
This ensures your **conversion, call validity, and order metrics remain clean** and reflect only real customer interactions.


# Real-Time Data Checks via Customer Endpoints

By Alex Bisbe. November 20th 2025.

Our AI Agent can connect directly to your backend systems to fetch live data—prices, availability, contract details, eligibility checks, or any other information exposed through your API. This ensures real-time accuracy during the conversation and eliminates the traditional “let me check that” delays of human agents.

{% hint style="info" %}
**Important:** This feature requires a **manual integration** by our engineering team. \
Please contact us to enable it.
{% endhint %}

***

### **How It Works**

The AI Agent makes on-demand requests to your API endpoints during the call. When the customer asks for a price, a coverage check, or any contract detail, the agent calls your system, retrieves the response, and uses it instantly in the conversation.

Your endpoint can return any payload (typically JSON). The AI Agent will:

* Interpret and speak the result
* Drive the decision tree based on the data
* Push the returned fields into the data-collection process if relevant

Everything works in real time—no hard-coded values, no stale information.

***

### **Use Cases**

#### **1. Live Pricing**

Fetch dynamic prices, discounts, or bundles based on:

* Postal code
* Consumption or usage
* Cart configuration
* Credit profile
* Product variant

The agent then verbalizes the exact offer instantly.

#### **2. Eligibility Verification**

Validation examples:

* Fiber coverage or network availability
* Product stock
* Contract status or renewal options
* Financing or scoring thresholds

If the endpoint returns “eligible,” the agent proceeds; otherwise, it reroutes.

#### **3. Dynamic Offer Generation**

Your API can return personalized recommendations, bundles, or cross/upsell paths.\
The agent uses this data to craft a tailored pitch in real time.

#### **4. Automatic Data Capture**

Any information returned by your endpoint can be automatically included in the data-collection process.

For example, your endpoint returns:

* *Price*
* *Bundle ID*
* *Coverage result*

These values can be stored directly—no additional customer questions required.

***

### **Data Sync With Your Systems**

All collected data—including fields obtained through your endpoints—can be pushed into:

* **Salesforce**
* **HubSpot**
* **Webhooks (POST callbacks)**
* **Your internal CRM or BI tools via custom integration**

This guarantees fully aligned data across your pipeline, from the call to your system of record.

***

### **Technical Overview**

* The AI Agent triggers your endpoints only when needed.
* Works with REST or GraphQL.
* Response times typically 100–400ms.
* The agent processes structured and unstructured fields.
* Full audit trail available through VoiceB’s call console.

{% hint style="info" %}
**Activation requires a manual integration. Contact us to enable this.**
{% endhint %}


# SIP Trunk — Private Call Routing for Enterprise Voice AI

Connect your existing telephony infrastructure directly to VoiceB.ai's AI agents through a dedicated, fully managed SIP trunk — no number porting, no middleware, no compromise.

***

### Overview

The **SIP Trunk** feature in VoiceB.ai allows enterprise customers to route calls from their existing PBX or Session Border Controller (SBC) directly into VoiceB.ai's AI agents over a private, dedicated SIP trunk. Instead of relying on shared telephony infrastructure or porting phone numbers to a third-party carrier, organizations keep full control of their telecom stack while seamlessly connecting it to conversational AI.

When a call arrives at your PBX, your SBC forwards it via SIP INVITE to VoiceB.ai's termination URI. VoiceB.ai processes the incoming SIP headers, optionally calls a pre-call webhook to pull real-time CRM context, applies routing rules to select the right AI agent, and then bridges the caller into the conversation — all within milliseconds.

<div data-with-frame="true"><figure><img src="/files/302XRt6kn9Ua0aoI6nij" alt="" width="251"><figcaption></figcaption></figure></div>

The feature is accessible from the main navigation sidebar under **SIP Trunk**.

*The SIP Trunk dashboard showing trunk metadata (name, SID, termination URI, region, security posture) and the IP Access Control list for whitelisting client SBC addresses.*

***

### How It Works

The call flow is straightforward:

1. **Caller dials your business number** — The number stays on your carrier; nothing changes for the end user.
2. **Your PBX/SBC forwards the call** — Via SIP INVITE to the VoiceB.ai termination URI (e.g., `vidaone-espana-…-voiceb-v-spain.pstn.twilio.com`), including any custom SIP headers your system attaches.
3. **VoiceB.ai processes the call** — It captures configured SIP headers, fires the pre-call webhook to fetch dynamic context from your CRM or backend, evaluates routing rules, and selects the appropriate AI agent.
4. **AI agent handles the conversation** — The agent uses the injected context (caller identity, account data, lead info) to personalize the dialogue in real time.
5. **Transfer back to your PBX** — When the agent needs to hand off to a human, VoiceB.ai fires a pre-transfer webhook with conversation data and routes the call to your SBC address.

***

### Configuration Guide

The SIP Trunk settings page is organized into clearly separated configuration sections, accessible via the left-hand tab navigation.

#### 1. Trunk Information

This read-only panel displays the core identifiers for your trunk:

| Field               | Description                                              |
| ------------------- | -------------------------------------------------------- |
| **Trunk Name**      | Human-readable identifier (e.g., `voiceb-v-spain`)       |
| **Trunk SID**       | Unique system ID for API references                      |
| **Termination URI** | The SIP endpoint your SBC should send INVITE requests to |
| **Security**        | Active security mechanisms (TLS, IP ACL)                 |
| **Region**          | Data-processing region (e.g., Dublin IE1)                |

<div data-with-frame="true"><figure><img src="/files/jvGfzbirsHPZ4TU83FfV" alt=""><figcaption></figcaption></figure></div>

#### 2. Security

This section manages two layers of protection:

**IP Access Control** — Whitelist the IP addresses of your SBC so only authorized sources can send SIP traffic to the trunk. Each entry has a label and an IP address or CIDR range.

**SIP Credentials** — Optionally configure SIP digest authentication credentials. When enabled, clients must authenticate with username/password in addition to IP whitelisting, adding a second factor of security.

&#x20;*IP Access Control showing two whitelisted Vodafone SBC addresses, plus the SIP Credentials section for optional digest authentication.*

#### 3. Twilio Trunk Configuration

These settings are read from and saved directly to the underlying Twilio infrastructure:

| Setting                          | Options                                                                          |
| -------------------------------- | -------------------------------------------------------------------------------- |
| **Secure Trunking (SRTP + TLS)** | Toggle on/off — encrypts both media and signaling                                |
| **Call Transfer**                | Disable all · Enable all · SIP only                                              |
| **Call Recording**               | Do not record · Record from ringing · Record from answer · Dual-channel variants |
| **Recording Trim**               | Do not trim · Trim silence                                                       |
| **Disaster Recovery URL**        | Fallback TwiML URL if the primary endpoint becomes unreachable                   |

#### 4. Client Integration

Configure how VoiceB.ai communicates back to your infrastructure during call transfers:

**Client SBC Address** — The FQDN or IP (plus port) of your PBX/SBC where transferred calls should be routed.

**Pre-Transfer Webhook URL** — A URL called with full conversation data (transcript, intent, extracted entities) before each transfer, so your CRM can prepare the receiving human agent.

**Webhook Signing Secret** — An HMAC-SHA256 secret for verifying webhook payload authenticity via the `x-webhook-signature` header.

&#x20;*Client Integration configuration with SBC address, pre-transfer webhook URL, and HMAC signing secret.*

#### 5. Pre-Call Webhook

This is one of VoiceB.ai's most powerful differentiators. On every incoming call — before the AI agent even picks up — VoiceB.ai fires a webhook to your backend system.

**Pre-Call Webhook URL** — The endpoint VoiceB.ai calls (POST or GET) on each inbound call.

**Pre-Call Payload Template** — A JSON template using `{{placeholder}}` tokens that map to SIP headers and call metadata. For example:

```json
{"caller": "{{from}}", "account_id": "{{X-Account-Id}}"}
```

This lets you inject dynamic CRM data, customer tier, open support tickets, or any other context into the AI agent's prompt — making each conversation hyper-personalized from the first second.

<div data-with-frame="true"><figure><img src="/files/MICb9sZL65MSPKXIdUKM" alt=""><figcaption></figcaption></figure></div>

#### 6. Phone Numbers (DDIs)

Assign phone numbers (DDIs) to the trunk. Assigned numbers are routed through the custom SIP trunk instead of native provider routing.

The **Linked Agents** table shows which AI agents are associated with this trunk and whether they use the webhook-based transfer tool. A simple toggle lets you link or unlink agents instantly.

#### 7. SIP Headers

Configure which SIP headers to capture on inbound calls and inject on transfers. Each header is configured with:

| Property        | Description                                                            |
| --------------- | ---------------------------------------------------------------------- |
| **Header Name** | The SIP header name (e.g., `X-Original-Caller`)                        |
| **Direction**   | INBOUND, OUTBOUND, or BOTH                                             |
| **Behavior**    | ROUTING (used in routing rules) or PASSTHROUGH (forwarded to webhooks) |
| **Required**    | Whether the header must be present                                     |
| **Forward to**  | Where the header value is sent (Webhooks, Agent, etc.)                 |

The **Caller ID Header** selector lets you pick which SIP header carries the real caller number, replacing the SIP `From` field throughout the system (in the agent, in leads, in logs).

<div data-with-frame="true"><figure><img src="/files/GHrPODxRgmKzln0eYJBx" alt=""><figcaption></figcaption></figure></div>

#### 8. Routing Rules

Define rules to route incoming calls to specific agents based on SIP header values, webhook response fields, or caller ID patterns. Rules are evaluated top-to-bottom with first-match-wins logic. A **Default Agent** is configured as a fallback when no rule matches.

#### 9. Call Tester

A built-in tool to validate your configuration without placing a real call, or to make an actual test call through the full pipeline:

**Validate mode** — Enter a DDI and simulated SIP headers, then click **Validate** to dry-run the routing logic and confirm the correct agent is selected.

**Test Call mode** — Enter your personal phone number and VoiceB.ai will call you, bridging you through the full incoming-call pipeline (SIP headers → pre-call webhook → routing rules → AI agent) so you can experience the end-to-end flow.

***

### Use Cases

#### Enterprise Contact Center Augmentation

Large contact centers (like the Vodafone España example shown in the interface) can offload frontline call handling to AI agents while keeping their existing carrier contracts, PBX infrastructure, and phone numbers untouched. The AI handles qualification, routing, FAQ resolution, and data collection — then transfers to human agents with full context via the pre-transfer webhook.

#### Dynamic CRM-Driven Conversations

Using the pre-call webhook, a financial services company can look up the caller's account, check their portfolio status, and inject personalized data into the AI agent's prompt before the caller even hears "hello." This eliminates the generic IVR experience.

#### Multi-Agent Routing by Business Unit

A multinational can use SIP header-based routing rules to direct calls to different AI agents depending on the department, product line, or language — all on a single trunk, with no changes to the PBX routing logic.

#### Gradual AI Adoption

Organizations can start by routing a single DDI to an AI agent, test performance, and progressively add more numbers and agents without disrupting existing telephony workflows.

***

### Competitive Advantages vs. Other AI Voice Platforms

#### VoiceB.ai vs. Bland AI

Bland AI offers a comprehensive SIP integration with a setup wizard, auto-discovery, failover, and number porting. However, the approach differs significantly:

**Architecture philosophy** — Bland expects you to either route calls to their infrastructure or port your numbers entirely. VoiceB.ai's model is the opposite: your PBX stays in control, and VoiceB.ai acts as an intelligent termination point. This is crucial for regulated enterprises that cannot port numbers or relinquish telephony control.

**Pre-call webhook with templated payloads** — VoiceB.ai's pre-call webhook lets you define a JSON payload template with `{{placeholder}}` tokens mapped to SIP headers. This means the AI agent has full CRM context before the conversation starts. Bland's approach focuses on pathway and persona configuration per number but doesn't offer the same depth of real-time, header-driven context injection.

**SIP header management** — VoiceB.ai provides a granular SIP header configuration table with direction (inbound/outbound/both), behavior (routing vs. passthrough), required/optional flags, and forwarding targets. This level of header control is essential for enterprise PBX integrations where headers carry call metadata like original caller ID, lead IDs, and interaction IDs. Bland supports custom headers but with less configurability.

**Pre-transfer webhook** — Before handing off to a human agent, VoiceB.ai fires a webhook with conversation data and supports HMAC-SHA256 signature verification for security. This allows the receiving agent's CRM screen to be pre-populated. Bland relies on standard call transfer without this middleware step.

**No-code UI** — VoiceB.ai's entire configuration is done through a visual dashboard with no API calls required. Bland's wizard is also user-friendly, but many advanced features (failover, codecs, alerts) require API calls.

#### VoiceB.ai vs. ElevenLabs SIP Trunking

ElevenLabs is primarily a voice synthesis platform that added SIP trunking to enable its AI agents to handle phone calls. The SIP integration is functional but comparatively basic:

**Configuration depth** — ElevenLabs' SIP setup covers transport type, media encryption, outbound address, optional custom headers, and optional digest auth. That's the full extent. VoiceB.ai offers all of this plus IP ACL management, pre-call webhooks with templated payloads, pre-transfer webhooks with HMAC signing, SIP header routing behavior configuration, call recording controls, disaster recovery URLs, routing rules, and a built-in call tester.

**Routing intelligence** — ElevenLabs assigns one agent per phone number and suggests using your PBX for conditional routing. VoiceB.ai has a native routing rules engine that evaluates SIP headers, webhook response fields, and caller ID patterns to dynamically select agents — moving routing intelligence from the PBX into the AI layer.

**Enterprise security** — VoiceB.ai provides dual-layer security (IP ACL + SIP digest auth), HMAC-SHA256 webhook signing, and configurable TLS/SRTP — all from the dashboard. ElevenLabs supports digest auth or IP-based ACL, and notes that for IP whitelisting you need to contact their support team. Their static IP infrastructure requires whitelisting an entire /24 block, which is less precise.

**Call testing** — VoiceB.ai includes a built-in Call Tester with both a dry-run validation mode (to verify routing logic without placing a call) and a live test call mode. ElevenLabs' troubleshooting guidance is documentation-based rather than tooling-based.

**Transfer workflow** — VoiceB.ai's Client Integration section with pre-transfer webhooks, SBC address configuration, and conversation data forwarding creates a complete transfer workflow. ElevenLabs does not offer an equivalent pre-transfer data handoff mechanism.

***

### Summary Comparison Table

| Capability                       | VoiceB.ai                     | Bland AI                     | ElevenLabs              |
| -------------------------------- | ----------------------------- | ---------------------------- | ----------------------- |
| Dedicated SIP trunk per customer | ✅                             | ✅                            | ✅                       |
| IP Access Control (ACL)          | ✅ Dashboard                   | ✅ API/Wizard                 | ⚠️ Contact support      |
| SIP digest authentication        | ✅                             | ✅                            | ✅                       |
| TLS + SRTP encryption            | ✅ Toggle                      | ✅ Required                   | ✅ Configurable          |
| Pre-call webhook (CRM context)   | ✅ With templated payloads     | ❌                            | ❌                       |
| SIP header capture & routing     | ✅ Full config table           | ⚠️ Basic                     | ⚠️ Custom headers only  |
| Routing rules engine             | ✅ Header/webhook/caller-based | ❌ Per-number only            | ❌ Per-number only       |
| Pre-transfer webhook             | ✅ With HMAC signing           | ❌                            | ❌                       |
| Client SBC address config        | ✅                             | ❌                            | ✅ (outbound address)    |
| Call recording controls          | ✅ Multiple modes              | ⚠️ Platform-level            | ⚠️ Additional config    |
| Disaster recovery / failover     | ✅ TwiML fallback URL          | ✅ Multi-server failover      | ❌                       |
| Built-in call tester             | ✅ Validate + live test        | ✅ SIP trace viewer           | ❌                       |
| Number porting                   | Not required (BYO trunk)      | ✅ Full porting support       | Not required            |
| No-code setup                    | ✅ Full dashboard              | ⚠️ Wizard + API for advanced | ✅ Basic dashboard       |
| Agent linking / multi-agent      | ✅ Per-trunk linking           | ✅ Per-number pathway         | ✅ Per-number assignment |

***

### Getting Started

1. Navigate to **SIP Trunk** in the VoiceB.ai sidebar.
2. Review your **Trunk Information** — note the Termination URI; this is what your SBC will send SIP INVITEs to.
3. Configure **Security** — add your SBC's IP addresses to the ACL.
4. Set up the **Twilio Trunk Configuration** — enable Secure Trunking, configure recording preferences.
5. Enter your **Client Integration** details — SBC address for transfers and the pre-transfer webhook URL.
6. Configure a **Pre-Call Webhook** — point it at your CRM endpoint and define the payload template.
7. Assign **Phone Numbers (DDIs)** and link your AI agents.
8. Define **SIP Headers** to capture and forward.
9. Set up **Routing Rules** if you need conditional agent selection.
10. Use the **Call Tester** to validate everything end-to-end.

***

*The SIP Trunk feature is available for enterprise customers. For provisioning, contact your VoiceB.ai account representative or reach out at* [*https://voiceb.ai/solutions/appointment-booking*](https://voiceb.ai/solutions/appointment-booking)

***


# Tools — Connect Your Agent to External Actions (Beta)

By Álvaro Iglesias. July 13th 2026

This guide explains how to use **Tools** to extend your VoiceB agent beyond conversation. Tools let your agent perform **real-time actions during a call** — checking data against your systems, validating a customer record, creating a lead, or triggering any HTTP endpoint — all while the conversation keeps flowing naturally.

The result: agents that don't just talk, but **execute** — querying, validating, and writing data in real time, exactly when the conversation requires it.

{% hint style="info" %}
**VoiceB tip:** Tools are **reusable at the organization level**. Build a tool once (e.g. "Check coverage by ZIP code") and attach it to as many agents as you need. Any changes you make are automatically **published to the agent's provider** — no manual sync required.
{% endhint %}

***

#### What is a Tool?

A Tool is an **HTTP action** your agent can call autonomously during a conversation. The AI decides *when* to use it based on the tool's description and the context of the call.

Typical use cases:

* **Real-time data checks** → verify if a caller is an existing customer before pitching.
* **Coverage / availability lookups** → check service availability by address or ZIP.
* **Lead creation** → push qualified lead data to your CRM mid-call.
* **Appointment booking** → query calendar availability and confirm a slot.
* **Order or ticket status** → retrieve live information from your backend.

***

#### Attaching Tools to an Agent

From the agent view, scroll to the **Tools** section:

<div data-with-frame="true"><figure><img src="/files/sSmODulTYjRnJWEvEhCM" alt=""><figcaption></figcaption></figure></div>

> 📷 **Image** — Tools (Beta) section on the agent

* Click **+** to create a new tool or attach an existing org tool.
* Click **Import from provider** to bring in tools already configured on the agent's provider — no need to rebuild them manually.

**Import from provider**

If your agent already has tools defined on the provider side, use **Import from provider** to pull them into VoiceB in one click. The import brings the full tool definition — name, description, request method, URL, headers, and parameters — and registers it as a reusable org tool.

From that moment, **VoiceB becomes the source of truth**: edit the tool in VoiceB and the changes are published back to the provider automatically.

{% hint style="info" %}
**Tip:** Import is the recommended path when migrating an existing agent into VoiceB — it guarantees the tool definitions in both systems start perfectly in sync.
{% endhint %}

***

#### Creating a New Tool

When you create a tool, you define **what it does**, **how it's called**, and **what data it sends and receives**.

**Name & Description**

<div data-with-frame="true"><figure><img src="/files/bKwknRuSc0PvbQyPHD9o" alt=""><figcaption></figcaption></figure></div>

> 📷 **Image** — New Tool modal: Name and Description fields

The description (up to **1,200 characters**) is the most important field: it's what the AI reads to decide **when to trigger the tool**.

Example: "Checks if the caller is an existing customer. Call this tool immediately after collecting and confirming the caller's phone number, before offering any product or plan. Send the phone number in international format (e.g. +34612345678). Returns `is_customer` (true/false) and `customer_segment`. If `is_customer` is true, redirect to customer service."

{% hint style="info" %}
**Tip:** Write the description as an **instruction to the agent**, not as internal documentation. State *when* to call it, *what inputs it needs*, and *what it returns*. Vague descriptions lead to tools being triggered too early, too late, or never.
{% endhint %}

**Request**

<div data-with-frame="true"><figure><img src="/files/qWzVl9XEuogn2lBOcq6J" alt=""><figcaption></figcaption></figure></div>

> 📷 **Image** — Request (Method + Webhook URL) and Headers

Define the HTTP call:

* **Method:** `GET`, `POST`, `PUT`, `PATCH`, or `DELETE`.
* **Webhook URL:** the endpoint of your API or middleware (e.g. Make, n8n, or a direct customer endpoint) — for example `https://api.example.com/v1/customers/lookup`.

**Headers**

Click **+ Add header** to add any headers your endpoint requires — typically authentication and content type. Each header has a key, a mode (**Fixed** or **Variable**), and a value. You can inject stored values with `{{ }}` syntax.

Common examples:

* `Authorization` → `Bearer {{API_KEY}}`
* `Content-Type` → `application/json`

**Parameters**

<div data-with-frame="true"><figure><img src="/files/SSSepFZGXIeeUaazlojT" alt=""><figcaption></figcaption></figure></div>

> 📷 **Image** — Parameters: name, location, type, fill mode, description, allowed values, Required

Parameters define the **data the agent sends** with each call. For each parameter you configure:

* **Location:** `Body`, `Query`, or `Path` — where the value is placed in the request.
* **Type:** `string`, `number`, `integer`, `boolean`, `object`, or `array`.
* **Fill mode** — how the value is populated:
  * **AI-filled** → the agent extracts the value from the conversation.
  * **Variable** → the value is injected from a declared variable using `{{variable}}` syntax.
  * **Fixed** → a constant value you set.
* **Description:** what the field means. The AI uses this to extract and format the value correctly.
* **Allowed values (comma-separated):** restrict the values the AI can send. Use it whenever your backend expects an enum (e.g. `Electricity, Gas`) — it prevents free-text values your CRM won't recognize.
* **Required:** mark it if the tool must never be called without this value. Applies to **AI-filled** and **Variable** parameters (a Variable may be empty if it wasn't provided and has no default). Fixed values are always present.

Example → a customer lookup might send `phone_number` (string, Body, AI-filled, required) and `product_interest` (string, Body, AI-filled, optional, allowed values `Electricity, Gas`).

**Object and array parameters**

<div data-with-frame="true"><figure><img src="/files/5l69RaU43YCKrBnFVfZT" alt=""><figcaption></figcaption></figure></div>

> 📷 **Image** — An `object` parameter with Fields, plus the Variables section

* **Object** types expose a **Fields** list (**+ Add field**). Each field is configured just like a parameter — name, type, fill mode, value, and Required. Use objects to group related data, e.g. a `contact` object holding `lead_id`, `name`, and `phone` populated from variables (`{{lead_id}}`, `{{name}}`, `{{phone}}`).
* **Array** types let you set the **item shape** — the type of each element (`string`, `number`, `integer`, `boolean`, or `object`).

{% hint style="info" %}
**Tip:** Give each parameter a clear, self-explanatory name and description. The AI fills parameter values from the conversation — the better it understands what a field means, the more accurately it extracts and formats the data.
{% endhint %}

**Variables**

Variables let you inject **dynamic values** into the tool — such as data passed via Click to Call Variables or values collected earlier in the call. Declare each variable (e.g. `lead_id`, `name`, `phone`), optionally give it a **default value**, and reference it anywhere in the tool — parameter values, object fields, or header values — with `{{variable}}`. Use them to avoid asking the caller for information you already have.

***

{% hint style="danger" %}
**Alert – Test before going live** Always validate a new tool with a **test call** before attaching it to a production agent: - Confirm the endpoint responds within a conversational latency budget (ideally **< 2 seconds**) — slow tools create awkward silences. - Verify required parameters are actually collected in the script **before** the tool can be triggered. - Check how the agent behaves when the endpoint fails or returns empty — define a fallback path in the script.
{% endhint %}

***

#### Best Practices

* **One tool, one job.** Keep each tool focused on a single action. Composite logic belongs in your middleware, not in the tool definition.
* **Align script and tools.** Reference the tool explicitly in the agent's script ("After confirming the phone number, check if the caller is an existing customer") so triggering is deterministic, not left to chance.
* **Mark critical fields as Required.** This prevents the agent from firing the tool with incomplete data.
* **Reuse org tools.** Centralize shared actions (customer lookup, CRM push) at org level instead of duplicating them per agent.
* **Use Variables for known data.** Map `{{lead_id}}`, `{{name}}`, `{{phone}}` from Click to Call instead of asking the caller again.
* **Handle the "no answer" path.** Every tool call can fail — script what the agent says while waiting and what it does if no result comes back.

***

#### Final Checklist

| :x: Don't                                            | :white\_check\_mark: Do                                              |
| ---------------------------------------------------- | -------------------------------------------------------------------- |
| Write vague descriptions like "checks customer data" | Write instruction-style descriptions: when to call, inputs, outputs  |
| Guess the method or endpoint URL                     | Confirm the correct method and endpoint with your API owner          |
| Send requests without authentication                 | Configure auth headers (`Authorization`, `Content-Type`)             |
| Leave parameters untyped or in the wrong location    | Type every parameter and place it correctly (Body / Query / Path)    |
| Let the agent fire the tool with incomplete data     | Mark critical fields as **Required**                                 |
| Ask the caller for data you already have             | Map **Variables** from Click to Call or earlier in the conversation  |
| Accept slow endpoints that create dead air           | Keep endpoint latency **under 2 seconds**                            |
| Assume the tool always responds                      | Define a **fallback path** in the script for errors or empty results |
| Attach a new tool directly to production             | Validate with a **test call** first                                  |

By [Álvaro Iglesias](https://www.linkedin.com/in/aiglesiasreina/). July 13th 2026.


# How to enable Two-Step Verification (2FA)

By Alex Bisbe. July 13th 20262

Two-step verification (also called 2FA or MFA) adds a second security layer to your VoiceB account. After entering your password, you'll also enter a 6-digit code from an authenticator app on your phone — so even if someone gets your password, they can't sign in without your device.

Setup takes about 2 minutes.

***

### Before you start

You'll need an **authenticator app** on your phone. If you already use one, skip ahead. If not, see Which authenticator app should I use? below and install one first.

***

### Enable Two-Step Verification

<div data-with-frame="true"><figure><img src="/files/YSsITWFAqmx1OmmiJtBF" alt="" width="375"><figcaption></figcaption></figure></div>

1. In the VoiceB dashboard, click your **name/profile icon** at the bottom left.
2. In the menu that opens, click **Manage account**.
3. In the Account window, select **Security** from the left sidebar.
4. Next to **Two-step verification**, click **+ Add two-step verification**.
5. A **QR code** will appear. Open your authenticator app and scan it:

   * Google Authenticator: tap **+** → **Scan a QR code**
   * Microsoft Authenticator: tap **+** → **Other account** → scan

   If you can't scan (for example, you're on the same device), most apps let you enter a setup key manually instead — look for a "manual entry" or "enter code" option.
6. Your authenticator app will now show a **6-digit code** that refreshes every 30 seconds. Type that code into the dashboard and confirm.
7. **Save your backup codes.** During setup you'll be shown a set of one-time backup codes. Store these somewhere safe (a password manager is ideal). They're the only way back in if you ever lose access to your phone.

<div data-with-frame="true"><figure><img src="/files/C88b6Tqcyi0umGfYKNUi" alt="" width="563"><figcaption></figcaption></figure></div>

<div data-with-frame="true"><figure><img src="/files/SZUGALHAe8MzKkbFzRm2" alt="" width="563"><figcaption></figcaption></figure></div>

You're done when the **Two-step verification** row shows **Authenticator application — Default**.

From now on, every time you sign in you'll enter your password and then a fresh 6-digit code from your app.

***

### How to remove Two-Step Verification

If you need to disable 2FA or switch to a new device:

1. Go to **Manage account → Security**.
2. Next to **Authenticator application**, click the **...** (three-dot) menu.
3. Choose **Remove** and confirm.

You can then set it up again from scratch with your new device.

***

### Which authenticator app should I use?

Any standard TOTP (time-based one-time password) app works with VoiceB. These are the main options:

* **Google Authenticator** *(recommended)* — free, simple, iOS and Android. Backs up codes to your Google account so you don't lose them if you change phones.
* **Microsoft Authenticator** — free, iOS and Android. Good choice if your company already uses Microsoft 365; also supports cloud backup.
* **Authy** — free, iOS, Android and desktop. Strongest multi-device support, so codes sync across your phone, tablet and computer.
* **1Password / Bitwarden** — if your team already uses one of these password managers, it can generate the codes and store your password together in one place.

> **Note:** Avoid Duo as your authenticator app if possible. It generates codes on a slightly different timing that can cause "invalid code" errors during setup. If you use Duo, enter the code quickly after it appears.

#### Installing Google Authenticator (if you don't have one)

1. Open the **App Store** (iPhone) or **Google Play Store** (Android).
2. Search for **Google Authenticator** and install it.
3. Open the app and sign in with your **Google/Gmail account** (this enables cloud backup of your codes).
4. Return to step 5 above to scan the QR code.

***

### Troubleshooting

* **"Invalid code" when entering the 6-digit code** — codes expire every 30 seconds. Wait for a fresh one and enter it right away. Also check that your phone's clock is set to update automatically; a wrong clock is the most common cause.
* **I lost my phone / changed devices** — use one of your saved **backup codes** to sign in, then go to Security and remove and re-add two-step verification with your new device. No backup codes? Contact VoiceB support and we'll reset it after verifying your identity.
* **The QR code won't scan** — use your app's manual-entry option and type in the setup key shown on screen.

***

### Need more help?

Reach out to the VoiceB team and we'll be glad to help.


# How to Connect Your VoiceB.ai Agent with HubSpot

Integrate your VoiceB.ai agent with HubSpot to automatically push call outcomes, lead data, and conversation metadata to your CRM — in real time, with no manual work.

### Prerequisites

Before you start, make sure you have:

* An active agent configured in VoiceB.ai
* A Data Collection Schema assigned to that agent (see: Create a Data Collection Schema)
* A HubSpot account with sufficient permissions to install integrations

***

### Step 1 — Activate the HubSpot Integration

Go to the main menu and click on Integrations. You will see the integrations currently active for your organization. If HubSpot is not already listed, activate it here. Once activated, HubSpot will be available to assign to any of your agents.

<figure><img src="/files/AQGARJsbn8gslPQ9zmiL" alt=""><figcaption></figcaption></figure>

***

### Step 2 — Verify Your Data Collection Schema

Go to [Data Collection](https://docs.voiceb.ai/create-a-data-collection-schema) in the main menu and confirm that the agent you want to connect has a schema configured. The fields defined in your schema (for example: name, phone, email, address) are what will appear in the HubSpot field mapping later. If you have not created a schema yet, do that first before proceeding.

<figure><img src="/files/QfHyMqVm5JunfuhzEA8d" alt=""><figcaption></figcaption></figure>

***

### Step 3 — Connect HubSpot to Your Agent

Navigate to your agent (for example, Angela) and scroll down to the Integrations section. Click the + button and select HubSpot from the list. This option will only appear if HubSpot has been activated in Step 1.

<figure><img src="/files/nuTDxqxp2dofjvO0Oyro" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/3wkkiJalcX4EehlDg8wq" alt="" width="563"><figcaption></figcaption></figure>

You will be redirected to a HubSpot login screen. Sign in with your HubSpot credentials and accept the permissions to allow VoiceB.ai to push data to your CRM. Once authorized, you will be returned to the agent configuration screen.

<figure><img src="/files/vvIMbLv8fwoFZBsnGU8d" alt=""><figcaption></figcaption></figure>

***

### Step 4 — Map Your Fields

After connecting, the HubSpot configuration panel will open automatically. This is where you tell VoiceB.ai which field in HubSpot should receive each piece of data from the call.

The top section shows the default VoiceB.ai call fields:

* Id — the unique call identifier
* Start / End — call start and end timestamps
* From / To — caller and destination numbers
* Agent — the VoiceB agent that handled the call
* Type — call type
* Status — call status
* Transcript — full call transcript (map this or set to Ignore)
* Summary — AI-generated summary (map this to Message or Ignore)
* [Tags](https://docs.voiceb.ai/tags-custom-ai-classification-for-calls-and-leads) — VoiceB classification tags (map or Ignore)

Below these default fields, you will see all the fields from your Data Collection Schema — for example Name, Phone, Email, Address. Map each one to the corresponding HubSpot contact property. For example, map VoiceB's email field to HubSpot's Email property.

<figure><img src="/files/avFcDws5P3HD5ktXx2CK" alt=""><figcaption></figcaption></figure>

{% hint style="warning" %}
Any field left unmapped will be ignored and will not be written to HubSpot.
{% endhint %}

***

### Step 5 — Configure Constants (Optional)

Below the field mapping, you will find the Constants section. Constants let you inject a fixed value into a HubSpot field on every call, regardless of what the agent collected. This is useful for tagging records with a source, a pipeline name, or an owner that should always be the same.

To add a constant, click Add Constant, select the HubSpot field from the dropdown, and enter the fixed value you want to inject. You can add as many constants as needed.

<figure><img src="/files/9Hl0DBkz4ONVHgXwu2hR" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
Example: set HubSpot field "Lead Source" to constant value "VoiceB" so every contact created by the agent is tagged correctly in your CRM.
{% endhint %}

***

### Step 6 — Set Trigger Conditions

The Trigger Conditions section controls which call outcomes cause a record to be pushed to HubSpot. This is one of the most important settings for keeping your CRM clean.

You have two quick options at the top:

* All [outcomes](https://docs.voiceb.ai/understanding-status-and-outcomes-in-voiceb) — every call, regardless of result, pushes a record to HubSpot
* Valid only — only calls with a sales-relevant outcome push a record

Below that, you can select individual outcomes manually. The available outcomes are: Sold, Reschedule, Talk to Agent, Pending, Customer Care, Not Interested, No Answered, Busy, Voicemail, Dropped, Failed, Unreachable, Not Found.

<figure><img src="/files/wDbnJu7laVDx0Wxofijw" alt=""><figcaption></figcaption></figure>

For most sales use cases, we recommend selecting Valid only, which covers Sold, Reschedule, Talk to Agent, and Pending. This prevents invalid calls, voicemails, and no-answers from polluting your CRM pipeline.

<figure><img src="/files/Rr8xMIOuxfOJkIvTzZei" alt=""><figcaption></figcaption></figure>

If you only want closed deals in HubSpot, select Sold only.

***

### Step 7 — Save

Click the Save button. Changes are applied immediately. From this point forward, every call that matches your trigger conditions will push data to HubSpot in real time — no delays, no manual imports.

***

### Deleting the Integration

If you need to remove the HubSpot integration from an agent, scroll to the bottom of the configuration panel and click the red Delete Integration button. This removes the connection for that agent only and does not affect other agents or your organization-level HubSpot activation.

***

### Need Help?

If you run into login issues, field mapping problems, or want to extend the integration to HubSpot workflows or deal creation, contact support via the Help Center.


# How to Connect Your VoiceB.ai Agent with Salesforce

Easily integrate your AI Agent with Salesforce to view all call data, outcomes, and collected lead info directly in your CRM — in real time.

This guide explains how to connect your VoiceB agent to Salesforce and configure advanced object mapping, multi-object creation, constants, and deduplication logic.

VoiceB is not a basic webhook integration. It allows you to define how your CRM architecture behaves directly from the VoiceB dashboard.

***

## Step 1 — Connect Salesforce

1. Go to **Integrations → Salesforce**
2. Open an agent and scroll down, now within the agent view Go to **Integrations → + → Salesforce** Click **Login**
3. Complete the OAuth authentication process
4. Grant the required permissions

Once connected, VoiceB will retrieve all available objects and fields from your Salesforce instance.

📌 If you need help creating a Salesforce External Client App, follow this guide:\
👉 *How to Register a Local External Client App in Salesforce*

<div data-with-frame="true"><figure><img src="/files/ucWojRXzPu6NoqYbhs43" alt=""><figcaption></figcaption></figure></div>

<div data-with-frame="true"><figure><img src="/files/q6CAMFiDnJgbNwnELoon" alt=""><figcaption></figcaption></figure></div>

<div data-with-frame="true"><figure><img src="/files/fMtp52QLtznF9RwSNF7H" alt="" width="318"><figcaption></figcaption></figure></div>

***

## Step 2 — Select the Main Object

VoiceB now allows you to choose which Salesforce object will receive the primary data.

You are no longer restricted to Leads.

Available options include:

* Lead
* Contact
* Account
* Opportunity
* Custom objects

This allows full flexibility depending on your sales architecture.

***

## Step 3 — Field Mapping Configuration

Each Salesforce field can be configured individually.

<div data-with-frame="true"><figure><img src="/files/xZonIVTVq0b3ihWGsNvs" alt=""><figcaption></figcaption></figure></div>

> Salesforce Object Configuration Interface in VoiceB Dashboard

For every field you can define:

### Source Type

You can choose between:

#### Variable

Dynamic data captured by the VoiceB agent.

Examples:

* `full_name`
* <kbd>email</kbd>
* `company_name`
* `summary`
* `from` (phone number)

#### Constant

A fixed value that will always be sent to Salesforce.

Example use cases:

* `LeadSource` = "Meta"
* `Campaign` = "Winter 2026"
* `Country` = "Spain"

This allows you to enforce CRM business rules directly from the integration layer without modifying the agent logic.

***

## Step 4 — Multi-Object Creation & Related Objects

VoiceB supports sending data to multiple linked Salesforce objects within a single execution.

You can:

* Create a Lead + Opportunity
* Create a Contact linked to an Account
* Create custom objects linked via relationship fields
* Associate objects using Salesforce relational logic

To configure this:

1. Click **Add Related Object**
2. Select the additional Salesforce object
3. Configure its field mapping
4. Define the relationship field

This enables structured CRM flows instead of dumping all data into one object.

***

## Step 5 — Deduplication Strategy

VoiceB allows configuring deduplication per object.

You can choose:

#### Option A — Always Create

A new record will be created every time.

#### Option B — Deduplicate by Field

Select a Salesforce field (e.g., Email or Phone).

If a matching record exists:\
→ It will be updated.

If no record exists:\
→ A new record will be created.

This gives you full control over whether VoiceB behaves as:

* A lead generator
* A CRM updater
* A strict record creator

Deduplication is configured at the object level.

***

## Step 6 — Optional Fields & Advanced Mapping

You can:

* Add optional Salesforce fields
* Map custom fields
* Remove unnecessary fields
* Adjust mapping without changing agent prompts

All mapping is managed directly from the VoiceB dashboard.

No development is required.

***

## How the Data Flow Works

1. Agent finishes conversation
2. Structured JSON outcome is generated
3. VoiceB evaluates mapping configuration
4. Objects are created or updated in Salesforce
5. Related objects are linked
6. Deduplication logic is applied

Everything happens automatically.

***

## What Changed (Compared to Previous Version)

The integration is no longer limited to:

* Fixed Lead creation
* Simple webhook behavior
* One-object mapping

VoiceB now acts as a CRM-native orchestration layer.

You control:

* Which objects receive data
* How objects relate
* Which values are constant
* How duplicates are handled

This allows enterprise-grade CRM automation without additional middleware.

***

#### Need Help?

Contact Support if:

* You experience login issues
* Field mapping doesn’t appear
* You want to extend the integration to workflows or deal creation

***

By integrating Salesforce with your VoiceB.ai Agent, your sales and support teams get full visibility of the conversation, call results, and lead data — **automatically**.


# How to register a local External Client App in Salesforce

### :white\_check\_mark: Prerequisites

* You have enough permissions to create connected apps

### Steps

* Go to setup > **Plaftorm Tools** > **Apps** > **External** **Client** **Apps** > **External** **Client** **App** **Manager**
* Click on **New External Client App**
* Fill the Basic Information as you wish

<figure><img src="/files/TR5d2fOqO1nOWqDgvUF8" alt=""><figcaption></figcaption></figure>

### API Section

* Check **Enable OAuth**
* Set **Callback URL** as <https://api.nango.dev/oauth/callback>
* Select the following **OAuth Scopes**:
  * Access the identity URL service (id, profile, email, address, phone)
  * Manage user data via APIs (api)
  * Perform requests at any time (refresh\_token, offline\_access)

<figure><img src="/files/sZZu4uKkQaiMNXexGmX8" alt=""><figcaption></figcaption></figure>

### Security Section

* Ensure the that bellow options are checked
  * Require secret for Web Server Flow
  * Require secret for Refresh Token Flow
  * Require Proof Key for Code Exchange (PKCE) extension for Supported Authorization Flows

<figure><img src="/files/TtPKHJyHIWI3IclZ6ozb" alt=""><figcaption></figcaption></figure>

* Click on **Create**

### Providing credentials to VoiceB

After clicking create it may take a while meanwhile Salesforce generate the Consumer Key and Consumer Secret.

You will find them under OAuth Settings section of your created app. By clicking on Consumer Key and Secret button, Salesforce will request validation sending a code to your email before showing them.

Since Salesforce integration is currently in beta, you'll need to contact VoiceB in order to provide the credentials.

<figure><img src="/files/n4nwxkFqXhSURC0pGBJE" alt=""><figcaption></figcaption></figure>


# Blog

The AI Sales Blog Insights, case studies, and updates from the future of Autonomous Sales Voice Agents

Browse all blog articles below.

<table data-view="cards"><thead><tr><th>Article</th><th data-card-target data-type="content-ref">Open</th><th data-hidden data-card-cover data-type="image">Cover image</th></tr></thead><tbody><tr><td>How Aticco turned website form fills into booked space tours with autonomous agents — 93% of bookings</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/EiM5GpFSNFRx7z0W1go4">/spaces/ggIr2RvhGm49D4vbjurT/pages/EiM5GpFSNFRx7z0W1go4</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/EiM5GpFSNFRx7z0W1go4">https://docs.voiceb.ai/~gitbook/ogimage/EiM5GpFSNFRx7z0W1go4</a></td></tr><tr><td>Your Voice Agent Sounds Great. So Why Isn't It Selling?</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/b9i8wGzputblCtYWt2RT">/spaces/ggIr2RvhGm49D4vbjurT/pages/b9i8wGzputblCtYWt2RT</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/b9i8wGzputblCtYWt2RT">https://docs.voiceb.ai/~gitbook/ogimage/b9i8wGzputblCtYWt2RT</a></td></tr><tr><td>The Rise of Autonomous Revenue Systems: Why 2027 Will Change Enterprise Sales Forever</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/Co71nfu9sGKDtCKgNoaH">/spaces/ggIr2RvhGm49D4vbjurT/pages/Co71nfu9sGKDtCKgNoaH</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/Co71nfu9sGKDtCKgNoaH">https://docs.voiceb.ai/~gitbook/ogimage/Co71nfu9sGKDtCKgNoaH</a></td></tr><tr><td>What Happens When Lead Capacity Has No Ceiling</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/XrjoDWlTHyqXemjQDD5L">/spaces/ggIr2RvhGm49D4vbjurT/pages/XrjoDWlTHyqXemjQDD5L</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/XrjoDWlTHyqXemjQDD5L">https://docs.voiceb.ai/~gitbook/ogimage/XrjoDWlTHyqXemjQDD5L</a></td></tr><tr><td>How BIG School closes 38.5% of sales autonomously with AI voice agents</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/aRabLzzFbYA5vo5hhzb1">/spaces/ggIr2RvhGm49D4vbjurT/pages/aRabLzzFbYA5vo5hhzb1</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/aRabLzzFbYA5vo5hhzb1">https://docs.voiceb.ai/~gitbook/ogimage/aRabLzzFbYA5vo5hhzb1</a></td></tr><tr><td>From click to checkout — guided by AI voice, closed without a human</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/99yXUw8krK2rY63OvhXl">/spaces/ggIr2RvhGm49D4vbjurT/pages/99yXUw8krK2rY63OvhXl</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/99yXUw8krK2rY63OvhXl">https://docs.voiceb.ai/~gitbook/ogimage/99yXUw8krK2rY63OvhXl</a></td></tr><tr><td>Video — VoiceB Platform Overview: Turning Inbound Calls into Autonomous Revenue</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/dt43TeZmO6vORc5bfvhg">/spaces/ggIr2RvhGm49D4vbjurT/pages/dt43TeZmO6vORc5bfvhg</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/dt43TeZmO6vORc5bfvhg">https://docs.voiceb.ai/~gitbook/ogimage/dt43TeZmO6vORc5bfvhg</a></td></tr><tr><td>The Real Advantage of Autonomous Inbound Sales Is Not Cost Reduction</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/TkHbJ0vM9NLr0QDeS2SI">/spaces/ggIr2RvhGm49D4vbjurT/pages/TkHbJ0vM9NLr0QDeS2SI</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/TkHbJ0vM9NLr0QDeS2SI">https://docs.voiceb.ai/~gitbook/ogimage/TkHbJ0vM9NLr0QDeS2SI</a></td></tr><tr><td>What is an Autonomous Sales Voice Agent (ASVA)?</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/RtawIDGirVhHrHHwm9Lj">/spaces/ggIr2RvhGm49D4vbjurT/pages/RtawIDGirVhHrHHwm9Lj</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/RtawIDGirVhHrHHwm9Lj">https://docs.voiceb.ai/~gitbook/ogimage/RtawIDGirVhHrHHwm9Lj</a></td></tr><tr><td>Where VoiceB Wins — The Right Conditions for Autonomous Sales</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/xlWmYvCnDWItjSLipzLo">/spaces/ggIr2RvhGm49D4vbjurT/pages/xlWmYvCnDWItjSLipzLo</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/xlWmYvCnDWItjSLipzLo">https://docs.voiceb.ai/~gitbook/ogimage/xlWmYvCnDWItjSLipzLo</a></td></tr><tr><td>Autonomy Levels (L1–L4) — VoiceB's Progressive AI Sales Framework</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/uR6AePf9g3dofT8vpfM5">/spaces/ggIr2RvhGm49D4vbjurT/pages/uR6AePf9g3dofT8vpfM5</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/uR6AePf9g3dofT8vpfM5">https://docs.voiceb.ai/~gitbook/ogimage/uR6AePf9g3dofT8vpfM5</a></td></tr><tr><td>AI for Inbound Sales — How VoiceB Transforms Revenue from Inbound Demand</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/i5MKNuTZBg43iigGRZvl">/spaces/ggIr2RvhGm49D4vbjurT/pages/i5MKNuTZBg43iigGRZvl</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/i5MKNuTZBg43iigGRZvl">https://docs.voiceb.ai/~gitbook/ogimage/i5MKNuTZBg43iigGRZvl</a></td></tr><tr><td>Call Center AI for Sales — Replacing and Augmenting Inbound Sales Teams</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/CBBta8ozmMIAERHE0vL4">/spaces/ggIr2RvhGm49D4vbjurT/pages/CBBta8ozmMIAERHE0vL4</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/CBBta8ozmMIAERHE0vL4">https://docs.voiceb.ai/~gitbook/ogimage/CBBta8ozmMIAERHE0vL4</a></td></tr><tr><td>The Future of Digital Sales: Assisted Voice Agents That Guide, Not Replace</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/0TXKM5tdjeJGF1mlxrZ8">/spaces/ggIr2RvhGm49D4vbjurT/pages/0TXKM5tdjeJGF1mlxrZ8</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/0TXKM5tdjeJGF1mlxrZ8">https://docs.voiceb.ai/~gitbook/ogimage/0TXKM5tdjeJGF1mlxrZ8</a></td></tr><tr><td>Holaluz Sales Autonomy Project</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/eKLZz3usXu8DYPXp7APZ">/spaces/ggIr2RvhGm49D4vbjurT/pages/eKLZz3usXu8DYPXp7APZ</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/eKLZz3usXu8DYPXp7APZ">https://docs.voiceb.ai/~gitbook/ogimage/eKLZz3usXu8DYPXp7APZ</a></td></tr><tr><td>Inbound Sales Is Being Replaced — Not Optimized</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/eVc0boRbX8U16KZu4CJh">/spaces/ggIr2RvhGm49D4vbjurT/pages/eVc0boRbX8U16KZu4CJh</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/eVc0boRbX8U16KZu4CJh">https://docs.voiceb.ai/~gitbook/ogimage/eVc0boRbX8U16KZu4CJh</a></td></tr><tr><td>VoiceB vs ElevenLabs, Bland, Retell, Decagon, and Salesforce Agentforce</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/oTjy6mLkdQziRbCW05v9">/spaces/ggIr2RvhGm49D4vbjurT/pages/oTjy6mLkdQziRbCW05v9</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/oTjy6mLkdQziRbCW05v9">https://docs.voiceb.ai/~gitbook/ogimage/oTjy6mLkdQziRbCW05v9</a></td></tr><tr><td>The Future of Landing Pages Is SuperVoice</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/2kVWxFAgbbGMtxns8bJ2">/spaces/ggIr2RvhGm49D4vbjurT/pages/2kVWxFAgbbGMtxns8bJ2</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/2kVWxFAgbbGMtxns8bJ2">https://docs.voiceb.ai/~gitbook/ogimage/2kVWxFAgbbGMtxns8bJ2</a></td></tr><tr><td>How Telecom Operators Use SuperVoice to Boost Digital Sales</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/OZqGk0BkKLxmru5mjxqE">/spaces/ggIr2RvhGm49D4vbjurT/pages/OZqGk0BkKLxmru5mjxqE</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/OZqGk0BkKLxmru5mjxqE">https://docs.voiceb.ai/~gitbook/ogimage/OZqGk0BkKLxmru5mjxqE</a></td></tr><tr><td>Business Case: Vodafone Spain × VoiceB.ai</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/Lt26kjNhmqbxbAlZx47o">/spaces/ggIr2RvhGm49D4vbjurT/pages/Lt26kjNhmqbxbAlZx47o</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/Lt26kjNhmqbxbAlZx47o">https://docs.voiceb.ai/~gitbook/ogimage/Lt26kjNhmqbxbAlZx47o</a></td></tr><tr><td>SuperVoice video demo for automotive</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/CV4RZ3UqKKfquHqCnUGy">/spaces/ggIr2RvhGm49D4vbjurT/pages/CV4RZ3UqKKfquHqCnUGy</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/CV4RZ3UqKKfquHqCnUGy">https://docs.voiceb.ai/~gitbook/ogimage/CV4RZ3UqKKfquHqCnUGy</a></td></tr><tr><td>Beyond the Form: Why the Future of B2C Inbound is Invisible, Instant, and Conversational</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/BtvecoEXxmwM2vv3drWK">/spaces/ggIr2RvhGm49D4vbjurT/pages/BtvecoEXxmwM2vv3drWK</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/BtvecoEXxmwM2vv3drWK">https://docs.voiceb.ai/~gitbook/ogimage/BtvecoEXxmwM2vv3drWK</a></td></tr><tr><td>Business Case: SHV Energy / Primagas Spain</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/S2FKQgO3kS5d2b5n5mi0">/spaces/ggIr2RvhGm49D4vbjurT/pages/S2FKQgO3kS5d2b5n5mi0</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/S2FKQgO3kS5d2b5n5mi0">https://docs.voiceb.ai/~gitbook/ogimage/S2FKQgO3kS5d2b5n5mi0</a></td></tr><tr><td>How VoiceB SuperQualification Transforms Inbound Sales</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/lLJuDq1BueiQujdB0e3n">/spaces/ggIr2RvhGm49D4vbjurT/pages/lLJuDq1BueiQujdB0e3n</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/lLJuDq1BueiQujdB0e3n">https://docs.voiceb.ai/~gitbook/ogimage/lLJuDq1BueiQujdB0e3n</a></td></tr><tr><td>Why Voice Still Matters in Digital Sales—and How AI Is Changing the Game</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/vBZgkTlLvOUlh4AwCoyp">/spaces/ggIr2RvhGm49D4vbjurT/pages/vBZgkTlLvOUlh4AwCoyp</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/vBZgkTlLvOUlh4AwCoyp">https://docs.voiceb.ai/~gitbook/ogimage/vBZgkTlLvOUlh4AwCoyp</a></td></tr><tr><td>Improve Your Lead Generation from Existing Traffic with AI Voice + WebRTC</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/EMFwyx633ffF13bmFyxO">/spaces/ggIr2RvhGm49D4vbjurT/pages/EMFwyx633ffF13bmFyxO</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/EMFwyx633ffF13bmFyxO">https://docs.voiceb.ai/~gitbook/ogimage/EMFwyx633ffF13bmFyxO</a></td></tr><tr><td>Video Case Study: How We Boosted Conversion by 46.7%</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/4WUsFEakmI3J25CMrnYr">/spaces/ggIr2RvhGm49D4vbjurT/pages/4WUsFEakmI3J25CMrnYr</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/4WUsFEakmI3J25CMrnYr">https://docs.voiceb.ai/~gitbook/ogimage/4WUsFEakmI3J25CMrnYr</a></td></tr><tr><td>Conversion Rates That Redefine Inbound Sales</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/CPBIsnjF8tJKgIZD0rK7">/spaces/ggIr2RvhGm49D4vbjurT/pages/CPBIsnjF8tJKgIZD0rK7</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/CPBIsnjF8tJKgIZD0rK7">https://docs.voiceb.ai/~gitbook/ogimage/CPBIsnjF8tJKgIZD0rK7</a></td></tr><tr><td>Widening the Funnel Without Adding Agents</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/VvEWwoWi5tH5mzguqUUL">/spaces/ggIr2RvhGm49D4vbjurT/pages/VvEWwoWi5tH5mzguqUUL</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/VvEWwoWi5tH5mzguqUUL">https://docs.voiceb.ai/~gitbook/ogimage/VvEWwoWi5tH5mzguqUUL</a></td></tr><tr><td>Top AI Voice Sales Solutions in 2025: What You Should Know</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/bO68ap6u3eYB33WH2zoa">/spaces/ggIr2RvhGm49D4vbjurT/pages/bO68ap6u3eYB33WH2zoa</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/bO68ap6u3eYB33WH2zoa">https://docs.voiceb.ai/~gitbook/ogimage/bO68ap6u3eYB33WH2zoa</a></td></tr><tr><td>Growth Is Changing. VoiceB.ai Is Already There</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/hK0NbFMLEbpjeHz75Aw0">/spaces/ggIr2RvhGm49D4vbjurT/pages/hK0NbFMLEbpjeHz75Aw0</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/hK0NbFMLEbpjeHz75Aw0">https://docs.voiceb.ai/~gitbook/ogimage/hK0NbFMLEbpjeHz75Aw0</a></td></tr><tr><td>How SHV Energy, Transformed Inbound Sales with AI Voice Agents</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/JD5Iai16pTXuOFracqS4">/spaces/ggIr2RvhGm49D4vbjurT/pages/JD5Iai16pTXuOFracqS4</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/JD5Iai16pTXuOFracqS4">https://docs.voiceb.ai/~gitbook/ogimage/JD5Iai16pTXuOFracqS4</a></td></tr><tr><td>SuperVoice Product Video Demo</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/Ap5Mh5Fbr2cvoWAcCAUy">/spaces/ggIr2RvhGm49D4vbjurT/pages/Ap5Mh5Fbr2cvoWAcCAUy</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/Ap5Mh5Fbr2cvoWAcCAUy">https://docs.voiceb.ai/~gitbook/ogimage/Ap5Mh5Fbr2cvoWAcCAUy</a></td></tr><tr><td>SuperQualification Video Demo for Insurance companies</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/w60d4B5xjw5RlRgEQ9ky">/spaces/ggIr2RvhGm49D4vbjurT/pages/w60d4B5xjw5RlRgEQ9ky</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/w60d4B5xjw5RlRgEQ9ky">https://docs.voiceb.ai/~gitbook/ogimage/w60d4B5xjw5RlRgEQ9ky</a></td></tr><tr><td>SuperQualification Video Demo for Online Banks</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/tP1xS6c8T9Lu4mbWgjKz">/spaces/ggIr2RvhGm49D4vbjurT/pages/tP1xS6c8T9Lu4mbWgjKz</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/tP1xS6c8T9Lu4mbWgjKz">https://docs.voiceb.ai/~gitbook/ogimage/tP1xS6c8T9Lu4mbWgjKz</a></td></tr><tr><td>SuperQualification Video Demo for Telco companies</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/UQXXF7HdRevTumn8RNGf">/spaces/ggIr2RvhGm49D4vbjurT/pages/UQXXF7HdRevTumn8RNGf</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/UQXXF7HdRevTumn8RNGf">https://docs.voiceb.ai/~gitbook/ogimage/UQXXF7HdRevTumn8RNGf</a></td></tr><tr><td>SuperVoice+ Video Demo for Utilities companies</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/pRxyLORIqs0TB2ViT2MY">/spaces/ggIr2RvhGm49D4vbjurT/pages/pRxyLORIqs0TB2ViT2MY</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/pRxyLORIqs0TB2ViT2MY">https://docs.voiceb.ai/~gitbook/ogimage/pRxyLORIqs0TB2ViT2MY</a></td></tr><tr><td>Introducing Balanced Voice Routing: Reliability, Speed, and Scale for Voice AI</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/8vJLG5ec0604krJ5dbDk">/spaces/ggIr2RvhGm49D4vbjurT/pages/8vJLG5ec0604krJ5dbDk</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/8vJLG5ec0604krJ5dbDk">https://docs.voiceb.ai/~gitbook/ogimage/8vJLG5ec0604krJ5dbDk</a></td></tr><tr><td>Voice AI Best Practices for Customer Experience</td><td><a href="/spaces/ggIr2RvhGm49D4vbjurT/pages/kDMoD6fv93gGEWWjgtDr">/spaces/ggIr2RvhGm49D4vbjurT/pages/kDMoD6fv93gGEWWjgtDr</a></td><td><a href="https://docs.voiceb.ai/~gitbook/ogimage/kDMoD6fv93gGEWWjgtDr">https://docs.voiceb.ai/~gitbook/ogimage/kDMoD6fv93gGEWWjgtDr</a></td></tr></tbody></table>


# How Aticco turned website form fills into booked space tours with autonomous agents — 93% of booking

By Alex Bisbe. July 31st 2026

***

## <sub>Aticco — Flexible Workspace (Spain & Portugal)</sub>

**Vertical:** Flexible workspace / commercial real estate \
**Deployment:** Click-to-call, outbound-on-inbound-lead \
**Autonomy level:** L4 (autonomous booking) \
**Data source:** One month of consecutive production call data

***

<figure><img src="/files/TdO07rYoydciaWlU1Apa" alt=""><figcaption></figcaption></figure>

### At a glance

|                                             |                  |
| ------------------------------------------- | ---------------- |
| Booking rate on connected conversations     | **27.4%**        |
| Tours booked with zero human involvement    | **93.3%**        |
| Calls resolved without an Aticco employee   | **96.8%**        |
| Calls that reached a human at Aticco        | **3.2%**         |
| Tours booked outside Aticco's service hours | **21.7%**        |
| Calls billed at zero                        | **88.8%**        |
| Median time to book a tour                  | **Under 2m 30s** |

***

### The client

[Aticco](https://aticco.com/) operates flexible workspaces across Barcelona, Madrid, Valencia and Lisbon. Its product range spans Flex and Fix memberships, private offices, custom-built offices, meeting rooms and event spaces — sold across four cities, in seven languages, at wildly different price points.

It is, in other words, exactly the kind of business where a phone call closes and a form does not.

### The challenge

Aticco's demand generation works. Its website captures leads continuously, in Spanish, English, Catalan, French, German, Italian and Portuguese, from prospects who are comparing three or four coworking operators at once.

Three structural problems sat between that demand and a signed membership:

**1. The response-time gap.** Aticco's customer service team is available for roughly a quarter of the week. Lead flow does not respect that schedule. A prospect who fills in a form on Friday afternoon, Saturday morning or during a Tuesday lunch peak waits — and while they wait, they visit a competitor's site.

**2. The qualification problem.** A form field marked "Flex, Barcelona" tells you almost nothing. Does this person want a hot desk two days a week, or a private office for a team of six? Which building fits their commute? Do they need a gym, a rooftop, parking, an auditorium? Building-level detail is what actually decides the sale, and a form cannot ask follow-up questions.

**3. The bottleneck was human capacity, not demand.** Every hour a community or sales team member spent dialling numbers that never answered was an hour not spent giving tours — which is where Aticco's sales actually close.

<figure><img src="/files/iCl5vDDGAt4ZfCXceWbU" alt=""><figcaption></figcaption></figure>

### The deployment

VoiceB deployed **Rosa**, an autonomous sales voice agent, on Aticco's click-to-call flow. The sequence:

1. A prospect submits a form on aticco.com.
2. Rosa calls them back immediately, referencing the specific product and city they enquired about.
3. Rosa qualifies: usage pattern, city, preferred building, team size, timing, budget sensitivity.
4. Rosa answers building-level questions — which spaces have a gym, which have a rooftop pool, which have parking, what is included in a membership.
5. Rosa checks live tour availability, proposes concrete slots, and books the tour directly into Aticco's calendar.
6. The structured outcome, tags and full transcript are written back for the space manager who will host the tour.

No human is involved in any of it unless the conversation requires one.

{% hint style="info" %}
**A note on "close."** In this deployment, the outcome VoiceB is paid for is a **booked tour**, not a signed membership. Aticco's memberships close in person, with a space manager, at the building. Rosa's job is to produce qualified, calendar-confirmed tours — the metrics below measure exactly that, and nothing more.
{% endhint %}

### Results

#### Only 3.2% of calls ever reached an Aticco employee

**96.8% of all telephone work was absorbed entirely by the autonomous agent.** Transfers to a human happened on 3.2% of calls — and only when the conversation genuinely required one.

#### 27.4% of conversations ended with a tour on the calendar

Of the calls where a prospect actually picked up and had a conversation, **27.4% ended with a confirmed tour.** Measured at the lead level rather than the call level, **close to 30% of every lead the agent successfully reached ended up booked.**

**93.3% of those bookings were made end-to-end by Rosa with zero human involvement.** Median booking conversation: **under two and a half minutes** — qualification, building recommendation, objection handling, live calendar check and confirmation, in the time it takes to make a coffee.

#### One in five tours was booked when Aticco's team was off the clock

**21.7% of all tours were booked outside Aticco's customer service hours.** Weekend traffic accounted for roughly one call in six, and weekend conversations converted at **22.5%** — in line with the weekday average.

Under the previous model, that demand either waited until Monday or went to a competitor. It is the most invisible category of loss in a headcount-limited sales team: you never see the leads you failed to call back.

#### Price objections did not kill the deal

An explicit price objection came up in **15% of conversations. 45.5% of those still ended in a booked tour** — well above the overall average.

This is the difference between an agent that reads a script and an agent that sells. When a prospect pushed back on monthly cost, Rosa reframed to the building's included services, offered the tour as the low-commitment next step, and closed the appointment rather than the objection.

#### Nearly a third of leads wanted something different from what they had filled in

In **28% of conversations**, the agent detected that the product the prospect had enquired about was not the product they actually needed — someone asking about a Flex membership who really needed a private office, or someone who had selected the wrong city.

**Those re-qualified conversations converted at 45.2%**, again far above average. A form cannot do this. It records the guess the prospect made in twenty seconds and passes it downstream as fact.

#### The agent switched language mid-call

When a prospect asked to continue in English, Rosa switched immediately, mid-conversation, without dropping context or restarting qualification. **Those calls converted at 46.2%.**

For an operator whose website serves seven languages and whose members are disproportionately international founders and expats, this removes an entire class of lost lead.

#### Callbacks to the agent's number converted at 45.2%

Inbound callbacks — prospects ringing the agent's number *back* — made up a small but disproportionately valuable slice of volume. **45.2% of those conversations ended in a booked tour**, well above the outbound average.

A prospect who returns your call is a prospect who has decided to take you seriously. Because Rosa answers that line instantly, at any hour, with the full context of the prior conversation already loaded, none of that intent leaks.

### What Aticco did not pay for

Outcome-based pricing is what makes this economically rational, and the data shows why.

**88.8% of calls were billed at zero.**

Most instructive: **one in five leads was never reachable at all** — disconnected numbers, mistyped international prefixes, permanently unavailable lines. Chasing them consumed **a third of all dialling activity.** In a traditional model, that third is salaried time. Under VoiceB's model it costs the client nothing, and it costs the client's team no attention whatsoever.

Aticco pays for booked tours and for transfers to a human. Not for minutes, not for seats, and not for the third of all attempts that reached a dead line.

### Why this pattern generalises

Flexible workspace shares a shape with the verticals VoiceB has already proven in energy, telecom and insurance:

→ **Speed decides the winner.** The prospect is comparing operators in a single browsing session. First qualified human contact wins disproportionately.&#x20;

→ **The offer is complex but bounded.** A finite set of buildings, products and amenities. Complex enough to defeat a form. Structured enough for an autonomous agent to master completely.&#x20;

→ **Demand arrives 24/7 and the sales team does not.** Over a fifth of bookings landed outside working hours.&#x20;

→ **The conversion event is an appointment, not a signature.** Any business whose real close happens in person — property viewings, test drives, clinic consultations, campus visits — can put an autonomous agent on the step immediately before it.

### What comes next

Two extensions are on the roadmap for this deployment:

**Closed-loop ad platform feedback.** Feeding VoiceB-confirmed tour bookings back to Meta and Google as conversion signals, so Aticco's acquisition campaigns optimise toward prospects who actually show up to a building — not toward form submissions.

**Post-tour outcome capture.** Closing the loop between booked tour, attended tour and signed membership, to give Aticco true cost-per-member visibility across its whole portfolio.

***

{% hint style="success" %}
**Want the same on your funnel?** If you generate 1,000+ inbound leads a month and your sales conversation happens on the phone, [book a demo](https://cal.com/team/voiceb-demo/voiceb-demo). Every engagement starts with a controlled, KPI-based pilot with agreed success criteria.&#x20;
{% endhint %}

*All figures are drawn from one month of consecutive production call data. Conversation-level metrics exclude voicemail, busy signals, unreachable numbers and sub-10-second drops. Absolute call and lead volumes are withheld at the client's discretion. Individual prospect details have been removed.*


# Your Voice Agent Sounds Great. So Why Isn't It Selling?

By Alex Bisbe, Co-founder & CEO, VoiceB.ai

In the last eighteen months, thousands of companies built AI voice agents. The tools made it easy: ElevenLabs gave everyone broadcast-quality voices, Retell and Bland made it possible to stand up a working agent in days instead of months. The demos were stunning. Executives heard an AI handle objections in a natural voice with sub-second latency, and budgets got approved on the spot.

> Then the pilot ended, and somebody in finance asked the only question that mattered:
>
> Did it sell?
>
> And nobody could answer.

There were transcripts. There were minutes consumed. There were sentiment scores, maybe a dashboard with call volumes. But there was no number anyone could put in front of a CFO that said: this many calls produced this many sales, this much pipeline, this cost per acquisition. The pilot didn't fail loudly. It just quietly produced no business case, and the project stalled.

If this sounds familiar, the problem is probably not your voice agent. And it's almost certainly not the platform you built it on. The problem is structural — and once you see it, you can't unsee it.

### Voice is solved. Outcomes aren't.

Let's give credit where it's due. The voice AI platforms are genuinely excellent at what they do. ElevenLabs made voice quality a commodity — we use their speech technology ourselves, and it's superb. Retell, Bland, and Vapi solved real-time orchestration: latency, interruption handling, turn-taking. Five years ago, each of these was a research problem. Today they're an API call.<br>

**That's exactly the point. The conversation layer is solved. Any competent team can ship an agent that sounds human in a week.**

But the conversation layer was never designed to answer commercial questions. It can transcribe a call perfectly and still have no idea what the call was. Was it a sale? A warm lead that needs a human? A wrong number? An existing customer asking about their bill? A competitor doing reconnaissance?

\
A transcript is not an outcome. A sentiment score is not a revenue event. And a per-minute invoice tells you exactly one thing: how long people talked. It tells you nothing about whether any of it was worth paying for.

\
This is the missing layer. Between a great-sounding agent and an actual revenue system sits a piece of infrastructure most voice deployments simply don't have: deterministic outcome classification, connected to billing, the CRM, and the ad platforms that generated the leads in the first place.

### The blind funnel

Here's why this matters beyond reporting. Without outcome classification, you can't even diagnose why your pilot underperformed.

A real example from our own production data. SHV Energy's Primagas brand has run autonomous sales agents with us for over a year, with a 67.8% autonomous close rate across twelve months. But there were specific periods where performance dropped sharply. If all we'd had were transcripts and minutes, the obvious conclusion would have been: the agent is failing. Tune the prompts. Change the voice. Maybe kill the project.

Because every call carried a deterministic outcome label, we could see what was actually happening: the agent's close rate on valid calls hadn't moved. What had changed was the ratio of valid to invalid traffic. A specific batch of Meta Ads campaigns was sending low-quality leads — wrong audience, wrong intent. The agent was fine. The leads were wrong.

That diagnosis took minutes, because the data existed. Without it, the same situation looks like AI failure, and that's exactly how most stalled voice pilots get read. The company concludes "voice AI doesn't work for us" when the truth was a fixable traffic problem nobody could see.

This is worth saying directly to everyone whose pilot disappointed: your team probably didn't fail. You were flying blind, because the infrastructure to see the funnel didn't exist in your stack.

#### Twelve months of outcome-classified traffic

Primagas / SHV Energy, vs prior period

<figure><img src="/files/LtrMyWEST1fiJglzONw0" alt=""><figcaption><p>Primagas - SHV Energy live production. All figures deterministic and auditable.</p></figcaption></figure>

{% hint style="success" %}
The same outcome data that diagnosed the traffic-quality problem now prevents it: every classified call feeds the diagnosis, the CRM, and the ad platform.&#x20;
{% endhint %}

### What the missing layer actually is

At VoiceB we call it the [Outcome Engine](https://voiceb.ai/outcome-engine), and it does three things the voice layer can't.

First, it classifies every call deterministically. Not a probabilistic sentiment guess — an auditable classification. Every call gets a Status (Valid or Invalid), then an Outcome. A Sold outcome means the customer accepted the offer and provided the required data. A Talk to Agent outcome means warm intent that needs human follow-up. Invalid outcomes — not interested, voicemail, dropped, customer care, unreachable — are identified with 96% accuracy and separated from the sales flow entirely. On top of that, custom Tags apply your own business logic in plain language: high intent, price driven, SME, existing customer.

Second, it makes the CRM trustworthy. Outcomes push to Salesforce or HubSpot in real time, structured and validated — and only the outcomes you specify trigger CRM writes. No more polluting your pipeline with voicemails and wrong numbers. Your sales team sees only sales-ready demand, with full context attached.

Third, it closes the loop with your ad spend. This is the part almost nobody has. Every classified outcome is a conversion signal. Fed back to Meta and Google, it means your campaigns stop optimizing for form fills and clicks and start optimizing for what actually sold. The same outcome data that diagnosed the Primagas traffic problem now prevents it: the ad platform learns which audiences produce Sold outcomes and reallocates budget automatically.

A lead arrives from a Meta ad. The agent qualifies it, closes it, writes the confirmed sale to the CRM, and tells Meta the lead converted — all within minutes, without a human touching anything. That's not a voice bot. That's a revenue system.

#### What outcome classification reveals in real traffic

Holaluz, live production

<figure><img src="/files/2Nz7d6ClDvyk7CzH4ojn" alt=""><figcaption><p>Holaluz, live production. All figures deterministic and auditable.</p></figcaption></figure>

{% hint style="success" %}
Every segment above is a deterministic, auditable classification — not an estimate. Invalid calls never reach a human agent and are never billed, on any plan.&#x20;
{% endhint %}

### The pricing model is the proof

Here's a simple test for any voice AI vendor: look at how they charge you.

If a vendor charges per minute or per call, they're telling you something important — they don't know what happened on the call. They can't distinguish a closed sale from a wrong number, so they bill both identically, and you carry all the risk.

#### The pricing test

<figure><img src="/files/igH7sBX1b8tOmMpE661p" alt=""><figcaption></figcaption></figure>

{% hint style="success" %}
&#x20;A vendor that can classify outcomes can afford to be paid on them.
{% endhint %}

We charge €0 for every invalid call. Not interested, voicemail, dropped, busy, unreachable — all free, on every plan. We only get paid when a real revenue event occurs: a qualified handoff to your team, or a sale closed autonomously. We can offer this for exactly one reason: the Outcome Engine classifies every call, so we know which ones created value.

The pricing model isn't a marketing gimmick. It's the architecture made visible. A vendor who can classify outcomes can afford to be paid on them. A vendor who can't, can't.

The results in production: at Holaluz, 30.4% of valid inbound calls now close autonomously, with 78% of total inbound traffic auto-filtered before a human is ever involved. At Primagas, the twelve-month numbers came alongside a 153% increase in sales and a 37% reduction in cost per acquisition — because for the first time, every euro of ad spend could be traced to an actual outcome.

### If your pilot stalled, start here

I've sold conversion infrastructure to this exact buyer for fifteen years. At my previous company, Whisbi, we put human agents on video for AT\&T, Vodafone, Toyota, and thirty other enterprise brands — and we proved that live conversion of inbound demand works at industrial scale. The one thing we could never do was close the loop: prove the sale, attribute it, and feed it back to the ad platform automatically. That's the layer VoiceB was built to be.

So if you've built a voice agent that sounds great but isn't moving the revenue line, don't start by rebuilding the agent. Start by asking what happened on the calls you already ran.

We'll show you. Bring a sample of your agent's calls — or your call center's — and in twenty minutes we'll run an outcome classification on that traffic: how much was invalid, what was transferable, what was closeable. No demo, no pitch deck. Just your funnel, visible for the first time.<br>

[Book your outcome audit](https://cal.com/team/voiceb-demo/voiceb-demo) — or [read the technical docs](https://docs.voiceb.ai) to see how classification works under the hood.

***

VoiceB.ai is the Autonomous Sales Voice Agent (ASVA) platform. Live in production with Vodafone España, Holaluz, and SHV Energy/Primagas.

<br>


# The Rise of Autonomous Revenue Systems: Why 2027 Will Change Enterprise Sales Forever

By Alex Bisbe. May 20th 2026

For the last decade, enterprise software focused on one thing: helping humans work more efficiently.

> CRM platforms helped sales reps organize pipelines.\
> Chat tools improved collaboration.\
> Contact center software optimized agent productivity.\
> AI copilots helped employees write emails faster.

But 2027 will mark a fundamental shift in enterprise technology.

Companies are no longer looking for software that assists humans.

They are looking for systems that execute **outcomes** autonomously.

And nowhere is this transformation happening faster than in revenue generation.

***

## From Software Tools to Autonomous Revenue Infrastructure

The first generation of AI in business was mostly assistive:

* AI chatbots
* AI copilots
* AI-generated summaries
* AI-powered recommendations
* AI routing systems

Useful? Yes.

Transformational? Not really.

Most of these tools still required humans to perform the core business action:

* qualify the lead
* answer objections
* guide the customer
* close the sale
* transfer the call
* collect the information
* complete the onboarding

The next wave is fundamentally different.

In 2027, enterprises will increasingly deploy what we call:

### Autonomous Revenue Systems

These are AI systems capable of:

* handling inbound demand
* qualifying customers
* answering objections
* collecting data
* guiding purchases
* booking appointments
* transferring only high-intent opportunities
* completing transactions
* operating 24/7 at unlimited scale

This is not automation.

This is operational replacement.

***

## Why This Trend Is Exploding Now

The idea of AI-powered sales is not new.

What changed is the technology stack finally became viable at enterprise scale.

Several breakthroughs converged simultaneously.

### 1. Real-Time Voice AI Became Natural

Latency dropped dramatically.

Modern AI voice conversations now feel fluid enough for real commercial interactions.

Customers are increasingly comfortable speaking with AI systems when:

* the experience is fast
* the responses are accurate
* the outcome is useful

According to Gartner, by 2026, 40% of enterprise applications will include agentic AI capabilities, up from less than 5% in 2025 \[1].

This changes everything for industries historically dependent on massive contact center operations.

***

### 2. AI Can Now Execute Workflows

Modern AI agents are no longer isolated chat interfaces.

They can:

* access APIs
* retrieve pricing
* check availability
* validate identities
* update CRMs
* trigger workflows
* qualify customers dynamically

The result:\
AI is moving from conversation to execution.

McKinsey estimates that generative AI could automate activities representing 60–70% of employee time in many business workflows \[2].

***

### 3. Enterprises Need Unlimited Capacity

Traditional sales organizations scale linearly with headcount.

More leads require:

* more agents
* more supervisors
* more training
* more office infrastructure
* more operational complexity

AI changes the equation completely.

An autonomous revenue system can scale infinitely without proportional labor costs.

Deloitte predicts that enterprises implementing AI-driven customer operations will significantly outperform competitors in response time, operational scalability, and customer acquisition efficiency \[3].

This is especially critical in industries with:

* seasonal spikes
* advertising bursts
* large inbound volumes
* long opening hours
* multilingual demand

The bottleneck is no longer human capacity.

***

## The Biggest Shift: Revenue Becomes Software

Historically, companies purchased software licenses.

Now they increasingly want to purchase outcomes.

This is why the future belongs to platforms measured by:

* conversion rate
* sales generated
* qualified leads
* appointments booked
* CPA reduction
* revenue captured
* retention improvements

Not by:

* seats
* users
* licenses
* dashboards

The market is moving toward outcome-based AI infrastructure.

This is a major platform transition similar to:

* on-premise → cloud
* desktop → mobile
* software → autonomous systems

PwC estimates AI could contribute up to $15.7 trillion to the global economy by 2030, with enterprise operational transformation being one of the largest drivers \[4].

***

## Industries That Will Adopt Autonomous Revenue Systems First

Some sectors are perfectly positioned for rapid adoption.

### Telecommunications

Telcos manage massive inbound demand:

* plan changes
* upgrades
* coverage questions
* pricing comparisons
* retention conversations

Most interactions follow structured sales logic and repetitive objection handling.

AI systems are uniquely effective in this environment.

***

### Insurance

Insurance sales are highly process-driven:

* qualification
* policy explanation
* pricing collection
* onboarding
* verification

Autonomous AI can drastically reduce operational costs while increasing availability and responsiveness.

***

### Banking & Financial Services

Banks increasingly need:

* instant lead qualification
* financing simulations
* account onboarding
* product recommendations
* customer verification

AI agents can operate continuously while handling large-scale inbound demand.

***

### Energy & Utilities

Energy switching and contract acquisition involve:

* repetitive qualification
* pricing explanation
* data collection
* customer onboarding

This is an ideal environment for autonomous conversational systems.

***

### Healthcare Intake & Qualification

Healthcare providers face enormous operational pressure managing:

* appointment qualification
* patient intake
* insurance verification
* routing
* pre-screening

AI agents can dramatically reduce operational bottlenecks.

***

## Why Voice Will Dominate Over Text

For years, many companies believed chat would replace voice.

The opposite is beginning to happen in high-value sales environments.

Why?

Because voice creates:

* higher trust
* better emotional connection
* faster qualification
* better objection handling
* more natural communication
* significantly higher conversion rates

Humans still close more effectively through conversation than through forms or chat interfaces.

According to Salesforce research, customers continue to prefer voice interactions for complex or high-value transactions because voice creates higher trust and faster resolution \[5].

AI voice systems are now reaching the quality threshold where enterprises can operationalize this at scale.

The result:\
Voice is evolving from a support channel into a primary revenue channel.

***

## The Companies That Win Will Own the Full Revenue Loop

The biggest winners in 2027 will not be generic AI vendors.

They will be platforms capable of orchestrating the full customer acquisition flow:

* inbound capture
* qualification
* conversation
* conversion
* transfer
* onboarding
* CRM synchronization
* outcome optimization

This requires much more than an LLM.

It requires:

* orchestration
* telephony infrastructure
* workflow execution
* latency optimization
* analytics
* integrations
* memory systems
* outcome tracking
* enterprise reliability

The moat is operational, not just technological.

***

## How VoiceB Fits Into This Future

At [VoiceB.ai](https://voiceb.ai/?utm_source=chatgpt.com), we believe the future of enterprise sales is autonomous.

VoiceB was built around a simple idea:

> Companies should not be limited by human sales capacity.

Our platform enables enterprises to deploy AI voice agents capable of:

* qualifying inbound leads
* answering objections
* guiding users through complex purchase processes
* transferring only high-intent opportunities
* automating repetitive sales conversations
* operating 24/7 at enterprise scale

Unlike traditional contact center software or AI copilots, VoiceB focuses on outcomes:

* more conversions
* lower CPA
* higher scalability
* faster response times
* operational efficiency

We are not building another chatbot.

We are building autonomous revenue infrastructure.

And as enterprises move toward AI-driven operations in 2027 and beyond, we believe this category will become one of the most important transformations in modern business.

***

## Final Thoughts

The next generation of enterprise software will not simply help humans work better.

It will execute business operations autonomously.

Autonomous Revenue Systems represent one of the largest B2B opportunities of the next decade because they directly impact:

* revenue growth
* operational scalability
* customer acquisition economics
* enterprise efficiency

The companies that adapt early will gain a structural advantage:

* faster response times
* unlimited lead handling
* lower operational costs
* better customer experience
* higher market share

The shift has already started.

By 2027, it will become mainstream.

***

## Sources

\[1] Gartner — *Gartner Predicts 40% of Enterprise Applications Will Feature Agentic AI by 2026*\
<https://www.gartner.com/en/newsroom/press-releases/2025-08-26-gartner-predicts-40-percent-of-enterprise-apps-will-feature-task-specific-ai-agents-by-2026-up-from-less-than-5-percent-in-2025>

\[2] McKinsey & Company — *The Economic Potential of Generative AI*\
<https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier>

\[3] Deloitte — *The State of Generative AI in the Enterprise*\
<https://www2.deloitte.com/us/en/insights/focus/cognitive-technologies/generative-ai-in-the-enterprise.html>

\[4] PwC — *Sizing the Prize: What’s the Real Value of AI for Your Business?*\
<https://www.pwc.com/gx/en/issues/artificial-intelligence/publications/artificial-intelligence-study.html>

\[5] Salesforce — *State of the Connected Customer Report*\
<https://www.salesforce.com/resources/research-reports/state-of-the-connected-customer/>


# What Happens When Lead Capacity Has No Ceiling

By Alex Bisbe. April 15th, 2026.

Most companies evaluate AI voice agents as a way to reduce contact center costs. Fewer agents. Lower cost per call. Better CPA.

These are real benefits. But they are the **least interesting** thing that happens when you deploy autonomous inbound sales.

The real transformation is structural. It changes how your entire demand engine works — from which leads you buy, to which channels you invest in, to how your marketing team thinks about the top of the funnel.

This article is about what we've observed across our enterprise clients — Vodafone Spain, Holaluz, SHV Energy / Primagas, BIG School, and others — as they've moved from fixed-capacity contact centers to unlimited autonomous lead handling.<br>

<div data-with-frame="true"><figure><img src="/files/lNEW2yj01fJWjCXJiEsc" alt=""><figcaption></figcaption></figure></div>

***

### The Constraint Nobody Talks About

Every B2C company with a contact center operates under an invisible ceiling: the number of calls their team can handle per hour.

This ceiling shapes every upstream decision. Marketing only buys the leads the contact center can absorb. Campaigns are throttled during peak hours and paused on weekends. Quality filters are set high — not because lower-quality leads are worthless, but because every bad call burns an expensive human minute. Off-hours traffic is a dead zone.

The result: most companies are not capacity-constrained by their market. They are capacity-constrained by their contact center.

They generate less demand than they could, from fewer sources than they should, at a higher cost per lead than necessary — all because there is a human bottleneck between the ad click and the sale.

***

### Remove the Ceiling. Everything Changes.

When VoiceB replaces or augments the first layer of inbound sales, the capacity constraint disappears. Every call is answered in under 4 seconds. There is no queue. There is no overflow. There are no off-hours.

What happens next is not incremental. It is a structural shift in how the company acquires revenue.

#### 1. You Can Buy Cheaper Leads

With a fixed-capacity contact center, you need every lead to be high quality. You pay a premium for that — expensive lead sources, aggressive quality filters, narrow targeting. Because if a bad lead gets through, it wastes a seat.

When filtering is free, the economics invert. You can open cheaper, higher-volume lead sources that were previously uneconomical. A lead that converts at 5% instead of 15% is unprofitable when a human handles it. When AI handles it, the marginal cost approaches zero — and 5% of a large number is a lot of sales.

**SHV Energy / Primagas** saw this firsthand. After deploying VoiceB, they grew lead volume by **251%** while reducing cost per acquisition by **37%**. More leads. Cheaper leads. More sales. The capacity ceiling was the bottleneck — not the market.

> *"VoiceB has provided more than a decade of lead conversion know-how combined with a powerful AI platform that has delivered real, measurable results — increasing our sales by 153%, leads by 251%, and reducing CPA by 37%."* — **Jordi Alonso**, Marketing Director,  SHV Energy Group

#### 2. You Open More Lead Sources

When every lead gets answered instantly, you stop being selective about where demand comes from. You can layer in sources that were previously too risky for a human-staffed center:

→ Low-CPM off-peak campaigns (evenings, weekends, late night — when ad costs are lowest and your competitors' agents are offline) → Broader paid media targeting with lower intent filters → Affiliate and partner traffic that you previously rejected for quality reasons → SEO and organic traffic that generates calls outside business hours → Landing page voice widgets that convert visitors who would otherwise abandon a form

**Holaluz** deployed VoiceB as the first touchpoint for all inbound traffic. The AI filters **78% of calls automatically** — invalid, non-sales, customer care — so human agents only handle sales-ready demand. The result: Holaluz can accept traffic from any source, at any time, knowing that noise will be separated from signal before a human is involved.

**30.4% of valid inbound calls are now closed autonomously by AI.** Another 21% are transferred to humans with full context. The contact center is no longer the bottleneck — it is the closer of last resort.

#### 3. You Reduce Agency and Platform Dependency

A fixed-capacity sales team makes you dependent on whoever controls the lead flow. Your media agency decides which keywords to bid on. Your Meta Ads account manager controls audience segmentation. Your Google Ads strategy is constrained by what your agents can absorb.

When capacity is unlimited, **you** control the volume dial. You can diversify across platforms, test new channels without risking overflow, and negotiate with agencies from a position of strength. If one source underperforms, you replace it — the AI handles the transition without missing a call.

This is not theoretical. We've seen clients shift budget across Meta, Google, organic, and affiliate channels in real time because the constraint was never the handling — it was the decision to try.

#### 4. No Human Threshold

Traditional contact centers have hard limits: X agents × Y hours × Z calls per hour = maximum capacity. Go above that and you get queues, abandoned calls, and lost revenue — precisely when your acquisition spend is highest.

**Vodafone Spain** handles **40,000+ monthly inbound calls**. After deploying VoiceB's AI agent "Rosa" on their inbound flow, over **70% of traffic is handled without human intervention**. Short transfers (low-value calls reaching agents) dropped by 19 percentage points. Agent workload was reduced by the equivalent of 26 FTEs — while maintaining identical sales performance.

The critical insight: Vodafone didn't deploy VoiceB to cut costs. They deployed it to **remove the ceiling** on how much inbound traffic their sales infrastructure can process. The cost savings are a side effect. The strategic value is that their acquisition capacity is now decoupled from headcount.

#### 5. Lead Qualification Stops Being a Gate — TOFU Redesign

This is the most underappreciated consequence and the biggest strategic unlock.

In a traditional model, the top of the funnel is designed to filter. Marketing qualifies before the call. Forms ask screening questions. Lead scoring models discard anything below a threshold. The entire TOFU exists to protect a scarce resource: human agent time.

When agent time is no longer the constraint, **the purpose of TOFU changes**. You stop filtering and start capturing. The qualification threshold drops to near zero — any lead with any signal gets handled. The AI qualifies during the conversation, not before it.

**BIG School**, an education company running high-volume enrollment campaigns, uses VoiceB to handle every inbound inquiry. The AI qualifies student interest, collects enrollment data, and routes only decision-ready conversations to human advisors. The funnel doesn't filter at the top — it captures everything and lets the AI sort in real time.

This changes how you design landing pages, how you write ad copy, and how you think about conversion rate optimization. When every visitor who shows intent gets a real-time conversation instead of a form, the definition of a "qualified lead" becomes irrelevant. Every lead is attended. Every signal is captured. Qualification happens in the conversation, not before it.

***

### The Compounding Effect

The companies that remove the capacity ceiling first will build a structural advantage that compounds over time.

Here is the loop:

**Unlimited capacity → more leads captured → better conversion data → lower CPA → more acquisition budget → more leads captured.**

The player that starts this loop earliest generates more data, which improves targeting, which lowers cost per acquisition, which frees budget for more volume. Late movers face the same market with less data, higher costs, and a capacity-constrained infrastructure.

This is not a technology advantage. It is a **market capture** advantage.

***

### It Works Across Models — Not Just Full Autonomy

A common misconception: this only works if the AI closes every sale end-to-end.

That's not how it works in production.

VoiceB deploys across a spectrum of autonomy levels — from pure qualification (L1) to fully autonomous closing (L4). Most enterprise clients combine multiple levels:

→ **Full autonomous close** for simple, repeatable products (residential energy, basic mobile plans, subscription enrollments) → **Hybrid model** where AI qualifies and collects data, then transfers with full context for human closing → **Assisted flow** where AI handles objections and guides the conversation, escalating only edge cases

**We have converted 100% of our enterprise POCs** because the methodology adapts to what each client needs. The capacity unlock happens at every level — whether AI closes the sale or simply ensures that every human agent spends 100% of their time on closeable conversations.

More volume. More sales. Lower CPA. Regardless of autonomy level.

***

### Production Data — Not Theory

| Client                    | Sector    | Capacity Unlock                                                                                         |
| ------------------------- | --------- | ------------------------------------------------------------------------------------------------------- |
| **SHV Energy / Primagas** | Energy    | +251% leads processed, +153% sales, −37% CPA over 12 months                                             |
| **Holaluz**               | Energy    | 78% of traffic auto-filtered, 30.4% autonomous close rate, 100% lead coverage 24/7                      |
| **Vodafone Spain**        | Telco     | 40,000+ monthly calls handled, 70%+ without human intervention, −26 FTE equivalent workload             |
| **BIG School**            | Education | All inbound enrollment inquiries handled by AI, human advisors handle only decision-ready conversations |

All results are from live production environments — not pilots or controlled tests.

***

### Video — See the Platform in Action

**VoiceB Platform Overview:** How inbound calls become autonomous revenue — qualification, routing, closing, CRM write-back, and structured outcome data in a single system.

{% embed url="<https://www.loom.com/share/42f9f906163b44f3bc1d320762887c8a>" %}

**SuperQualification for Telco:** Autonomous sales agent designed for telecom acquisition funnels — digital-assisted checkout, phone-based data collection, and qualified human transfer for edge cases.

{% embed url="<https://www.loom.com/share/29b5bb70047c411bb0c033dd4862c82c?sid=304fd66f-8043-404e-af9d-fcb6a6d5d28e>" %}

***

### The Question Is Not "Should We Deploy AI?"

The question is: **how much demand are you leaving on the table because your contact center has a ceiling?**

Every lead that goes unanswered because agents are busy. Every campaign that isn't running at 11pm. Every cheap lead source you rejected because the conversion rate was "too low for human agents." Every weekend when your competitors' customers are looking for alternatives and your phones are off.

That is the real cost of a capacity-constrained sales infrastructure. And it compounds every month you wait.

***

**VoiceB.ai** — Autonomous Sales Voice Agents for enterprise inbound.

→ [Read the SHV Energy / Primagas Case Study](https://docs.voiceb.ai/blog/business-case-shv-energy-primagas-spain)&#x20;

→ [Read the Holaluz Sales Autonomy Project](https://docs.voiceb.ai/blog/holaluz-sales-autonomy-project)&#x20;

→ [Read the Vodafone Spain Business Case](https://docs.voiceb.ai/blog/business-case-vodafone-spain-voiceb.ai)&#x20;

→ [Book a Working Session](https://cal.com/team/voiceb-demo/voiceb-demo)

Production data: SHV Energy / Primagas (+251% lead growth, +153% sales, −37% CPA) · Holaluz (30.4% autonomous close rate, 78% auto-filtered) · Vodafone Spain (40,000+ monthly calls, 70%+ AI-handled, −26 FTE workload equivalent).


# How BIG School closes 38.5% of sales autonomously with AI voice agents

By Alex Bisbe. April 8th, 2026

***

**Vertical:** Education — online master's degrees&#x20;

**Agent:** Virginia (VoiceB autonomous agent)&#x20;

**Period:** March 2026 (31 days)&#x20;

**Product:** Máster de Desarrollo con Inteligencia Artificial

**Geography**: Spain and LATAM

***

### The challenge

[BIG School](https://thebigschool.com/) generates inbound demand at scale through paid acquisition campaigns. Every click-to-call lead represents a prospect who expressed interest moments earlier — a high-intent window that closes fast if no one picks up.

The traditional problem: contact centers are finite. Overflow means lost leads. Off-hours means lost leads. High agent turnover means inconsistent quality. And every call — whether it ends in a sale or a hang-up — consumes an agent's time.

VoiceB deployed Virginia, an autonomous AI voice agent, to handle 100% of inbound click-to-call traffic for the Máster de Desarrollo con IA. No queue. No wait time. No off-hours gap.

***

### Results — March 2026

38.5% autonomous close rate. 72.9% of traffic filtered before reaching a human agent. 100% conversion once the payment page was reached.

**38.5% autonomous close rate** — of all valid sales interactions, more than one in three were closed end-to-end by the AI agent, with no human involvement at any stage of the conversation.

**72.9% of all inbound traffic auto-filtered** — nearly three out of four calls were handled, classified, and resolved by the agent without consuming a single second of human agent time.

**100% conversion at payment page** — every call where Virginia successfully guided a lead to the payment page resulted in a sale.

**Average call duration on sold calls: 3m 33s** — a complete qualification, objection handling, and purchase guidance flow in under four minutes.

<div data-with-frame="true"><figure><img src="/files/cxs5OromiSvrRpw18PlB" alt=""><figcaption></figcaption></figure></div>

<div data-with-frame="true"><figure><img src="/files/RcXJDUyZnFzqpvHmQI9j" alt=""><figcaption></figcaption></figure></div>

***

### How the autonomous close works

Virginia handles the full conversation from the moment the call connects:

1. Greets the prospect and identifies the product they enquired about
2. Answers questions on curriculum, pricing, payment options, access terms, and bonifications (including FUNDAE for employed learners)
3. Handles objections — payment from abroad, installment eligibility, cryptocurrency, VAT exemptions
4. Guides the prospect to the payment page and explains the checkout process
5. Classifies the outcome and writes structured data to the CRM

No script-reading. No hold music. No "let me check with a colleague." The agent knows the product and closes the conversation.

***

### What the agent detected in converted calls

Beyond the sale, Virginia tagged every call with structured intent signals. Among calls that converted to a sale:

* **75.7%** reached the payment page, guided there by the agent
* **40.5%** expressed explicit purchase intent mid-conversation
* **24.3%** requested additional product information before committing
* **13.5%** asked about FUNDAE corporate training bonification
* **5.4%** asked to speak to a human agent — and converted anyway

<div data-with-frame="true"><figure><img src="/files/EdNEMMYNHHATiRHFU0JR" alt=""><figcaption></figcaption></figure></div>

This data feeds back into script optimization. Every call is a labeled training signal for the next version of the agent.

***

### Benchmarks

BIG School's first-month autonomous close rate of 38.5% exceeds VoiceB's published energy benchmark of 30.4% (Holaluz).&#x20;

Auto-filtered traffic reached 72.9%, compared to 78% for Holaluz after 12 months of optimization. \
\
For a first month of production — before any script optimization cycle — this is a strong baseline.

<div data-with-frame="true"><figure><img src="/files/38YWqvQlHS0B8YbYUOdA" alt=""><figcaption></figcaption></figure></div>

***

### What comes next

The March results establish the baseline. The optimization levers available from here:

* **Script refinement** — using call-level outcome data to identify which objections preceded drop-offs and which topics correlated with conversion
* **Traffic quality analysis** — aligning ad targeting with the lead profiles that convert at the highest rate
* **FUNDAE flow** — nearly one in seven converted calls raised bonification questions; a dedicated script block for FUNDAE-eligible leads could meaningfully accelerate those conversions
* **Sustained traffic volume** — consistent daily lead flow unlocks statistically significant optimization cycles

***

*VoiceB is an Autonomous Sales Voice Agent (ASVA) platform. Virginia is a VoiceB-configured AI agent. All outcomes are classified by the VoiceB Outcome Engine — deterministic, auditable, and directly tied to performance-based pricing. Results from March 2026, real production calls only, test calls excluded.*

*Want to see how VoiceB performs in your vertical?* [*Book a demo*](https://cal.com/team/voiceb-demo/voiceb-demo)


# From click to checkout — guided by AI voice, closed without a human

Product · SuperVoice · March 2026 · 7 min read

***

Most companies spend a fortune driving traffic to their website. The landing page converts at 3%. The form sits there. The visitor has a question. The chat widget says "we'll reply in a few hours." They leave.

This is the gap VoiceB was built to close — not with a chatbot, not with a callback form, but with an AI voice agent that appears on the page, talks to the visitor in real time, guides them through the purchase, fills in the form on their behalf, collects consent, and hands a confirmed sale to the CRM. No human involved. No wait time. No drop-off.

> **"60–70% of users abandon landing page forms. Phone-based sales close at 3–4× the rate of email or chat. VoiceB puts a voice agent directly in that gap."**

***

### Two ways to deploy — same result

VoiceB's guided checkout experience can be embedded on any page in two modes, depending on whether your infrastructure is web-first or telephony-first:

* **WebRTC widget (VoIP)** — instant in-browser voice, no phone number needed on either side, works on desktop and mobile, zero latency for the user.
* **Click-to-Call widget** — traditional PSTN call initiated from the page, works on any device, familiar for users, with full SIP trunk support for enterprise.

Both are embedded as a single JavaScript snippet — one line of code on the landing page. No IT project. No custom integration. Both modes give the agent the same capabilities: it knows which page the user is on, it can open new URLs in the user's browser, it fills forms in real time, and it collects dual consent before submitting anything.

***

### The journey, step by step

Here is exactly what happens from the moment a visitor lands on your page to the moment a confirmed sale appears in your CRM.

#### 1. The widget activates

The widget can be configured to appear after a scroll depth threshold, a time delay, on exit intent, or immediately on high-intent pages. The button is a single tap — no form to fill, no callback to wait for. The agent picks up in under one second.

#### 2. The agent knows where it is

This is one of the most powerful aspects of the VoiceB widget. The agent receives the current page URL and page title at the moment the call starts. It knows whether the user is on the homepage, a product landing page, a tariff comparison page, or the checkout step. The script adapts accordingly — no generic greeting, no "how can I help you today."

#### 3. Qualify, explain, handle objections

The agent captures what the user needs: their interest, their situation, their hesitations. It handles objections using client-trained scripts with configurable guardrails. It never invents pricing. It queries live pricing or coverage APIs where available. For energy clients, it validates the CUPS supply point ID in real time during the conversation.

#### 4. Open the checkout page

Once the user has committed to proceeding, the agent opens the right URL in the user's browser — the checkout page, the sign-up flow, or the contract page. This happens programmatically, via the widget, without the user having to navigate. They stay in the conversation while the page loads alongside them.

#### 5. Fill the form on their behalf

The agent fills in the form fields it has already collected during the conversation: name, email, address, tariff selection, supply point, identity information. The user sees the form being completed in real time. They can correct anything — they remain in control throughout.

#### 6. Dual consent — the standard, done right

Before any form is submitted, the agent collects two layers of consent explicitly, with confirmation in the widget and a voice record of acceptance:

* **Terms & Conditions and Privacy Policy** — confirmed via widget checkbox and verbally during the call.
* **Consent to voice recording** — required under EU consumer protection law for distance contracts and compliant with the Spanish Real Decreto on certified sales confirmation.
* **GDPR data processing consent** — collected with the agent stating the data controller identity, purpose, and retention period. Full audit trail preserved.

This is not a checkbox the user clicks once and forgets. The agent reads the consent statement, the user confirms verbally, the widget records acceptance with a timestamp, and the audio is stored as part of the interaction audit trail. This structure is compliant with EU AI Act transparency requirements and fully auditable by regulators or enterprise legal teams.

#### 7. Sale confirmed, CRM written

The form is submitted. The sale is written to the CRM in real time — HubSpot, Salesforce, or Bitrix via webhook. The agent sends a confirmation email or WhatsApp trigger. The interaction is tagged in the Outcome Engine as an Autonomous Sale. The human contact center never touched it.

***

### What this looks like in production

This is not a prototype. The flow above is live in production with enterprise energy and telco clients. The numbers below come from 12 months of real deployment data.

| Client                    | Result                                                             |
| ------------------------- | ------------------------------------------------------------------ |
| **Primagas / SHV Energy** | 67.8% autonomous close rate over 12 months of production data      |
| **Holaluz (energy)**      | 30.4% of valid inbound calls closed autonomously by AI agent       |
| **Response time**         | Under 4 seconds to first agent response — zero wait, zero overflow |

{% embed url="<https://www.loom.com/share/d429f1be47fa41deab4e95fd634ae0dc>" %}

**Watch the demo — guided checkout in action: VoiceB SuperVoice · WebRTC widget · form fill + dual consent**

***

### Why this matters beyond the demo

The form abandonment problem is not a UX problem. It is a trust and friction problem. Users leave forms because they have questions nobody is answering, because the process feels uncertain, and because filling in their personal data on a cold page with no confirmation feels risky.

A voice agent resolves all three. It answers questions in real time. It guides the user step by step. It fills in what it already knows — so the user is reviewing, not data-entering. And it collects consent with a spoken confirmation that the user experiences as a real, human-grade commitment, not a checkbox.

> **"The autonomous close rate is more powerful than cost reduction as a headline metric. '30.4% of your sales happen without a human' changes how boards think about contact center investment."**

This also closes the loop with ad platforms. Every sale completed via the VoiceB widget generates a structured, labeled conversion event — Autonomous Sale — that can be sent directly to Meta Ads Conversions API or Google Ads offline conversions. Ad campaigns start optimizing for actual sales, not clicks. The entire acquisition cycle tightens.

***

### Technical requirements — genuinely minimal

Deploying the guided checkout widget requires exactly one JavaScript snippet added to the landing page. VoiceB handles the rest.

| Requirement              | Detail                                                                                                                         |
| ------------------------ | ------------------------------------------------------------------------------------------------------------------------------ |
| **Client-side**          | One-line JS embed. Works on any CMS, landing page builder, or custom web stack.                                                |
| **Data residency**       | All conversation data processed and stored on VoiceB EU infrastructure. GDPR-compliant by default.                             |
| **IT project required?** | No. For clients with public-facing pricing or coverage APIs, VoiceB connects directly without backend work on the client side. |
| **CRM integration**      | HubSpot and Salesforce natively. Bitrix and custom systems via webhook. Sale records appear instantly post-call.               |
| **Channels**             | WebRTC (VoIP, no phone number needed) or Click-to-Call (PSTN/SIP trunk). Both embedded identically.                            |

***

### Who is this for?

This capability is built for companies that:

* Run Meta or Google Ads campaigns linking to landing pages with a call-to-action to call or be called.
* Have a sign-up or checkout flow with more than 3 fields — the drop-off risk zone.
* Operate in telco, energy, or insurance, where the sales conversation is repeatable and the product is explainable by phone.
* Receive 1,000+ inbound leads per month and cannot scale headcount to match demand.

If your website has a "call us" button or a "request a callback" form, VoiceB replaces that with a real-time AI agent that does not wait, does not overflow, and does not miss a lead because it is 11pm on a Sunday.

***

### Get started

Book a 30-minute demo and we will run the full guided checkout flow on your actual landing page URL — not a sandbox, not a generic script.

[Book a demo →](https://cal.com/team/voiceb-demo/voiceb-demo)


# Video — VoiceB Platform Overview: Turning Inbound Calls into Autonomous Revenue

{% embed url="<https://www.loom.com/share/42f9f906163b44f3bc1d320762887c8a>" %}

Inbound sales teams face a structural inefficiency: thousands of calls arrive every month, but real buying intent is only discovered after spending time on each conversation. VoiceB introduces an **autonomous revenue layer** that qualifies, guides, and closes sales in real time.

### From Calls to Structured Revenue Intelligence

VoiceB codifies every interaction automatically, generating:

* Recordings, transcripts, and AI summaries
* Outcome classification and custom tags
* Structured data to optimize campaigns and conversion

This turns voice conversations into **actionable revenue intelligence**.

### Agent Studio: Build and Optimize Autonomous Agents

Inside Agent Studio, teams can:

* Configure scripts and guardrails
* Monitor performance metrics
* Manage agent versions and experiments
* Deploy agents across campaigns and markets

Agents are fully configurable by language, voice provider, LLM model, and campaign type.

### Memory That Removes Friction

VoiceB includes configurable conversation memory. When prospects call again, the agent verifies identity and continues the process with full context — eliminating repetition and improving conversion probability.

### Enterprise Integration by Design

VoiceB integrates directly into existing commercial infrastructure:

* Native CRM sync (e.g. Salesforce, HubSpot)
* SIP trunk and phone number management
* Real-time outcome data pushed into workflows

This ensures autonomous conversations translate into **measurable revenue impact**.

### Autonomous Sales, Not Just Automation

VoiceB enables organizations to:

* Close transactions autonomously
* Scale inbound capacity without increasing headcount
* Route only high-intent conversations to human teams

The result is a shift from managing calls to **operating a scalable revenue engine**.


# The Real Advantage of Autonomous Inbound Sales Is Not Cost Reduction

It's market capture. And the window is closing.

Most conversations about AI in sales start and end with cost reduction. Fewer agents. Lower cost per call. Better CPA.

These are real benefits. But they're the wrong frame.

The companies that will win with autonomous sales AI are not the ones that cut costs fastest. They're the ones that **capture the most demand before their competitors can respond.**

This article is about that advantage.

***

### The Market Is Not Infinite — Leads Are a Finite Resource

Every product category has a finite addressable demand at any given moment. There are only so many people in Spain looking to switch energy provider this month. Only so many families shopping for a new insurance policy. Only so many SMEs evaluating a new telecom contract.

That demand is not evenly distributed over time — it spikes around campaigns, price changes, competitor moves, and seasonal patterns. And it is not patient.

A lead that doesn't hear from you within minutes either goes cold, calls a competitor, or stops looking entirely.

> The inbound sales game is not about who has the best product. It is about who captures the most intent, fastest, at scale.

### Speed Is Not a Nice-to-Have — It Is the Product

There is substantial research on lead conversion and response time. The pattern is consistent: conversion probability drops sharply after the first 5 minutes of inbound intent.

A person who clicked your ad is thinking about your product right now. They have high intent, low friction, and an open decision. Every minute they wait, that window closes — they get distracted, they call someone else, they talk themselves out of it.

VoiceB responds in under 4 seconds. Not because fast is better. Because **first is often everything.**

<div data-with-frame="true"><figure><img src="/files/Q2T95Dd9h802K96KW4QN" alt=""><figcaption></figcaption></figure></div>

FIG. 01Conversion Probability vs. Response Time

Inbound lead conversion collapses within minutes of intent.

→After 5 minutes, conversion probability drops by **\~80%**. VoiceB responds in **under 4 seconds** — at 2pm on a Tuesday or 11pm on a Sunday.

### Beyond Cost: The Experience Advantage

The cost reduction narrative also ignores something important: **AI-powered inbound sales delivers a measurably better customer experience** than the human alternative in high-volume environments.

Consider what most inbound sales calls actually look like today: the customer waits on hold for 4–12 minutes, gets connected to an agent handling 3 other calls simultaneously, and receives a script delivered 200 times that day. Product knowledge varies depending on who answers.

Now consider VoiceB:

| Experience factor    | Traditional call center     | VoiceB                           |
| -------------------- | --------------------------- | -------------------------------- |
| Response time        | 4–12 min queue              | <4 seconds                       |
| Availability         | Business hours only         | 24/7, no gaps                    |
| Product knowledge    | Variable by agent           | Perfect, every call              |
| Memory of past calls | Rarely                      | Always                           |
| Consistency          | Varies by agent, time, mood | Identical quality at call 10,000 |
| Concurrent capacity  | Limited by headcount        | Unlimited                        |

This is not a downgrade from human sales. In high-volume inbound environments, it is a significant upgrade — for the customer and for conversion rates.

### The Strategic Picture: First-Mover Compounding

Traditional sales has a hard capacity ceiling. Human teams scale linearly with headcount, and headcount takes time and money to scale. When your campaign generates 3× the normal lead volume, agents are overwhelmed, queues spike, leads abandon, and conversion collapses exactly when acquisition spend is highest.

**Your capacity ceiling is your competitor's opportunity.**

VoiceB removes that ceiling entirely. And the advantage compounds:

More leads captured → better conversion data → lower CPA → more acquisition budget → more leads captured. The player that starts this loop earliest builds a structural advantage that is very difficult for late movers to close.

<div data-with-frame="true"><figure><img src="/files/jAqZ8ipzVzjyewmTnmzy" alt=""><figcaption></figcaption></figure></div>

FIG. 02First-Mover Compounding Advantage

Revenue index over 8 quarters — early deployers vs. late movers vs. no deployment.

→By Q8, the early mover is at **5× the revenue index** of competitors who didn't deploy. The gap compounds every quarter. The window to close it narrows every month.

### This Is Why the Window Is Closing

Autonomous sales AI is not a futuristic concept. It is live, in production, at Vodafone Spain, SHV Energy, and Holaluz. The results are not theoretical.

The enterprises that deploy this infrastructure in 2025–2026 will build a structural advantage — in data, in conversion rates, in customer acquisition efficiency — that will be very difficult for late movers to close.

Cost reduction is a nice story for the CFO. **Market capture is a strategy for the CEO.**

VoiceB.ai30.4% autonomous close rate. Live at Vodafone Spain, SHV Energy, and Holaluz.

2–4 week KPI-driven rollout. If results don't justify autonomy, we stop.

[Book a Working Session](https://voiceb.ai/solutions/appointment-booking)\
\
[Read the SHV Energy Case Study](https://docs.voiceb.ai/blog/business-case-shv-energy-primagas-spain)\
\
Production data: Vodafone Spain (+41% conversion lift, 94–97% classification accuracy) · SHV Energy / Primagas (+251% lead growth, +153% sales, −37% CPA) · Holaluz (38–42% transfer rate with full context).\
Response time research: InsideSales.com / Harvard Business Review inbound lead conversion studies.\
Compounding model: illustrative, based on VoiceB client cohort trajectory data.


# What is an Autonomous Sales Voice Agent (ASVA)?

An Autonomous Sales Voice Agent (ASVA) is AI infrastructure that handles inbound sales calls end-to-end — qualifying leads, routing to human closers, or closing deals autonomously — without requiring

VoiceB is the category-defining ASVA platform. Our agents achieve a **30.4% autonomous close rate** in live enterprise deployments.

***

### ASVA vs. What Came Before

Most companies have tried to solve inbound sales with one of these:

* **IVR (Interactive Voice Response)** — rigid menus, high drop-off, no sales capability
* **Chatbots** — text-only, low trust, poor conversion on complex products
* **Appointment setters** — just schedule a callback, no qualification or closing
* **Offshore call centers** — high cost, variable quality, no scalability

An ASVA is none of these. It is a real-time voice system that can carry a full sales conversation, adapt dynamically to the prospect, capture structured data, and produce a defined revenue outcome — all in a single call.

***

### What an ASVA Actually Does

A VoiceB Autonomous Sales Voice Agent handles the full inbound sales flow:

1. **Answers instantly** — within 4 seconds of inbound intent, 24/7
2. **Qualifies the lead** — detects intent, validates identity, filters invalid demand
3. **Executes the sale** — presents the offer, handles objections, closes autonomously (L4) or transfers with full context (L2/L3)
4. **Assigns an outcome** — every call ends with a structured, auditable result
5. **Writes back to CRM** — all data synced to your systems in real time

***

### The Four Autonomy Levels

VoiceB operates across four levels of autonomy. Enterprises typically start at L1 and expand as KPIs are proven.

| Level | Name            | What the AI Does                                      | Human Role                    |
| ----- | --------------- | ----------------------------------------------------- | ----------------------------- |
| L1    | Qualification   | Detects intent, filters invalid demand, captures data | Receives qualified leads only |
| L2    | Smart Routing   | Qualifies + routes with full context                  | Closes warm transfers         |
| L3    | Assisted Sales  | AI sells standard cases, escalates exceptions         | Handles edge cases only       |
| L4    | Full Autonomous | End-to-end sales with no human involvement            | Optional oversight            |

***

### Proven in Production

VoiceB ASVAs are not demos or pilots. They are live in enterprise environments:

* **Vodafone Spain** — +41% conversion lift, 94–97% classification accuracy
* **SHV Energy / Primagas** — +251% lead growth, -37% CPA
* **Holaluz** — 38–42% of valid calls transferred with full context and structured summary

> "VoiceB's AI voice infrastructure is guiding users end-to-end — completing forms via live voice, increasing digital sales, and reducing operational costs." — Victor M. Sanchez, Managing Director at GAIN

***

### Who ASVA Is For

ASVA is designed for enterprises generating **1,000+ inbound leads per month** via phone, form, or ad click — particularly in:

* Telecom
* Energy & Utilities
* Insurance
* Banking
* Automotive

If your sales team is spending time on calls that never convert, or if you're losing revenue outside business hours, ASVA directly addresses those gaps.

***

### Get Started

VoiceB runs KPI-driven pilots in 2–4 weeks. If results don't justify autonomy, we stop.

[Book a Working Session →](https://voiceb.ai/solutions/appointment-booking)


# Where VoiceB Wins — The Right Conditions for Autonomous Sales

VoiceB is not a universal solution. It is purpose-built for a specific problem: high-volume inbound sales environments where conversion speed, qualification accuracy, and 24/7 availability determine r

Here is where VoiceB consistently delivers the strongest results.

***

### The Ideal Deployment Profile

VoiceB performs best when three conditions are met:

**1. High inbound volume** 1,000+ leads per month from paid acquisition, form fills, or direct calls. Below this threshold, the cost of human qualification is manageable. Above it, the economics of ASVA become compelling.

**2. Repeatable sales conversation** Products and services where the sales conversation follows a structured path — a fixed set of qualifying questions, a defined offer, a predictable objection set. Telco contracts, energy tariffs, insurance policies, and financial products fit this profile.

**3. Speed-to-contact matters** Research consistently shows conversion drops dramatically after the first 5 minutes of inbound intent. If your agents are calling back leads 20–60 minutes later, VoiceB recovers that revenue by engaging within 4 seconds.

***

### Industries Where VoiceB Has Proven ROI

#### Telecom

Vodafone Spain uses VoiceB to handle inbound sales calls with 94–97% classification accuracy. The system filters invalid demand, qualifies genuine prospects, and routes them to closers with full context — resulting in a +41% conversion lift.

#### Energy & Utilities

SHV Energy / Primagas deployed VoiceB for inbound lead activation. Result: +251% lead growth, +153% sales increase, -37% cost per acquisition. The agent handles CUPS validation, identity verification, and tariff presentation autonomously.

#### Insurance

High-volume inbound insurance sales benefit from L3 autonomy — the AI handles standard product presentation and qualification, escalating only complex edge cases to human agents.

#### Banking & Financial Services

Identity validation, product eligibility screening, and appointment routing for mortgage, loan, and investment products.

#### Automotive

Lead qualification from digital campaigns, test drive scheduling, and pre-sales intent scoring for dealership networks.

***

### Where VoiceB Does Not Win

Being honest about fit:

* **Low-volume, relationship-driven sales** — if you're doing 50 calls/month on complex enterprise deals, ASVA is not the right tool
* **Highly regulated environments with no AI approval pathway** — some sectors require explicit regulatory clearance before AI can close sales
* **Outbound cold calling** — VoiceB is built for inbound. We do not support cold-list outbound
* **Pure customer support** — ASVA is a revenue tool, not a support deflection tool

***

### The VoiceB Competitive Advantage

Compared to generic AI voice platforms (Bland AI, Retell, ElevenLabs):

| Capability                  | Generic Platforms | VoiceB |
| --------------------------- | ----------------- | ------ |
| Sales-first design          | No                | Yes    |
| Outcome Engine              | No                | Yes    |
| Enterprise vertical scripts | No                | Yes    |
| Memory across calls         | No                | Yes    |
| CRM write-back              | Limited           | Full   |
| Outcome-based pricing       | No                | Yes    |
| Production close rate data  | No                | 30.4%  |

[See how VoiceB compares to Bland, Retell, and ElevenLabs →](https://docs.voiceb.ai/blog/voiceb-vs-elevenlabs-bland-retell-decagon-and-salesforce-agentforce)

***

[Book a Working Session →](https://voiceb.ai/solutions/appointment-booking)


# Autonomy Levels (L1–L4) — VoiceB's Progressive AI Sales Framework

VoiceB operates across four levels of sales autonomy.

The framework is designed so enterprises can start with low-risk qualification and expand to full autonomous closing as KPIs are proven.

**Humans participate only when the system decides they're needed.**

***

### Why Levels Matter

Most AI voice vendors offer a single mode: answer the call, qualify, transfer. VoiceB is the only platform that defines a structured progression from qualification to fully autonomous closing — and gives you control over where on that spectrum you operate.

This matters because:

* Different products within the same company may warrant different autonomy levels
* Regulated environments may require human closing even with AI qualification
* Trust in the system is built incrementally, not assumed from day one

***

### Level 1 — Qualification

**What the AI does:** Answers every inbound call instantly. Detects intent. Validates identity. Filters invalid demand (duplicates, existing customers, out-of-scope requests, support noise). Captures structured data according to your schema.

**What humans receive:** Only calls that meet your qualification criteria, with a full structured summary.

**Business impact:** Eliminates 40–60% of calls that were previously consuming agent time with zero revenue potential.

**Best for:** Any enterprise starting with AI voice. Zero risk — humans still close every sale.

***

### Level 2 — Smart Routing

**What the AI does:** Everything in L1, plus real-time routing logic. High-value prospects transferred live to the right closer based on product interest, geography, or value tier. Low-intent qualified leads scheduled for callback.

**What humans receive:** Warm transfers with full context — the agent knows exactly who they're speaking to and what was discussed.

**Business impact:** +20–40% improvement in agent close rate on transferred calls, because they're receiving pre-sold, contextualised prospects.

**Best for:** Enterprises with segmented sales teams or tiered product lines.

***

### Level 3 — Assisted Autonomous Sales

**What the AI does:** Everything in L1–L2, plus autonomous handling of standard sales cases. The AI presents the offer, addresses common objections, and closes the sale for products that follow a predictable pattern. Complex or unusual cases are escalated.

**What humans receive:** Only exceptions — edge cases, complaints, complex configurations, regulatory queries.

**Business impact:** 20–40% reduction in agent headcount requirements for standard product lines. Higher agent satisfaction (they handle the interesting cases, not repetitive scripts).

**Best for:** Products with well-defined sales scripts and predictable objection sets — energy tariffs, standard insurance policies, basic telecom contracts.

***

### Level 4 — Full Autonomous Sales

**What the AI does:** End-to-end sales execution. The AI qualifies, presents, handles objections, captures payment or contract details, validates identity, and closes — entirely without human involvement.

**What humans receive:** A completed sale record in the CRM, ready for fulfilment.

**Business impact:** 24/7 revenue generation. No queue. No headcount constraint. Every inbound lead handled at peak intent.

**Proven results:** VoiceB achieves a **30.4% autonomous close rate** at L4 in live enterprise deployments.

**Best for:** High-volume, standardised products where regulatory environment permits autonomous closing.

***

### Choosing Your Starting Level

| Starting condition                             | Recommended level |
| ---------------------------------------------- | ----------------- |
| First AI voice deployment, risk-averse         | L1                |
| Existing human sales team, want to optimise    | L2                |
| Proven product script, want to scale           | L3                |
| High volume, standardised product, 24/7 needed | L4                |

Most VoiceB clients start at L1 or L2 and expand within 90 days based on KPI results.

***

[Book a Working Session →](https://voiceb.ai/solutions/appointment-booking)

***

***

##


# AI for Inbound Sales — How VoiceB Transforms Revenue from Inbound Demand

Inbound sales is fundamentally different from outbound. The lead has already raised their hand. They clicked your ad, filled your form, or called your number. The conversion window is minutes, not day

***

### The Inbound Sales Problem

Enterprise inbound sales operations face three structural problems that AI directly solves:

**Problem 1: Speed-to-contact** The average B2C company responds to inbound leads in 47 hours. Studies show conversion probability drops by 80% after the first 5 minutes. Every minute of delay is lost revenue.

VoiceB responds within 4 seconds of inbound intent — 24/7, including weekends and public holidays.

**Problem 2: Qualification waste** In most high-volume inbound environments, 40–60% of calls are invalid — duplicates, existing customers, support requests, or out-of-scope inquiries. Human agents spend significant time filtering these before reaching a genuine sales conversation.

VoiceB filters invalid demand automatically, so agents only speak with people worth speaking to.

**Problem 3: Capacity constraints** You cannot instantly scale a human sales team. When campaign spend increases, lead volume spikes, or business hours end, revenue opportunity is lost.

VoiceB has no capacity limit. It handles unlimited concurrent inbound calls.

***

### How AI Changes the Inbound Sales Model

Traditional inbound sales model:

```
Ad Click → Form / Call → Queue → Agent → Qualification → Close
         [Revenue lost here: delay, filtering, capacity]
```

VoiceB inbound sales model:

```
Ad Click → VoiceB triggers instantly → Qualify + Close/Route → CRM outcome
         [Revenue recovered: speed, accuracy, 24/7 availability]
```

***

### What AI Can and Cannot Do in Inbound Sales

**AI does well:**

* High-repetition qualification questions
* Standard objection handling
* Identity and eligibility verification
* Data capture and CRM write-back
* 24/7 availability without degradation
* Consistent script execution across thousands of calls

**AI escalates to humans:**

* Complex product configurations
* Regulatory edge cases
* High-emotion situations requiring empathy
* Complaints and disputes

VoiceB's L1–L4 framework gives you precise control over where that boundary sits.

***

### Inbound Sales Metrics That Change with VoiceB

| Metric                 | Before VoiceB     | After VoiceB           |
| ---------------------- | ----------------- | ---------------------- |
| Speed to first contact | 20–60 min         | <4 seconds             |
| Invalid call filtering | Manual            | Automatic (40–60%)     |
| After-hours coverage   | None              | 24/7                   |
| Cost per acquisition   | High              | -37% (SHV Energy case) |
| Conversion rate        | Baseline          | +41% (Vodafone case)   |
| Lead capacity          | Headcount-limited | Unlimited              |

***

### The Outcome-Based Difference

VoiceB doesn't charge per minute or per seat. You pay per outcome — per qualified lead or per AI-closed sale. This means our incentives are aligned with yours: we only win when your inbound sales operation wins.

[Read the SHV Energy case study →](https://docs.voiceb.ai/blog/business-case-shv-energy-primagas-spain)

[Book a Working Session →](https://voiceb.ai/solutions/appointment-booking)


# Call Center AI for Sales — Replacing and Augmenting Inbound Sales Teams

Traditional inbound call centers are expensive, inconsistent, and capacity-constrained. AI is not replacing call centers overnight — but it is fundamentally changing how the economics work.

VoiceB is built specifically for the revenue side of this transformation.

***

### The Call Center Sales Problem at Scale

Enterprise inbound call centers handling sales face a structural cost problem:

* **Fixed headcount costs** regardless of call volume fluctuation
* **Training and attrition** — average call center agent turnover is 30–45% annually
* **Quality inconsistency** — script adherence, tone, and conversion rate vary significantly across agents
* **Coverage gaps** — evenings, weekends, and holidays generate leads with nobody to answer them
* **Supervision overhead** — quality monitoring, coaching, compliance recording

VoiceB addresses every one of these — not by eliminating the call center, but by restructuring what humans do within it.

***

### The VoiceB Model: AI Handles Volume, Humans Handle Value

The most effective deployment model is not AI vs. humans — it is AI + humans operating at their respective strengths.

**VoiceB handles:**

* 100% of initial inbound call answering (instant, 24/7)
* Invalid demand filtering (40–60% of volume in most enterprises)
* Standard qualification and data capture
* Routine sales closing for standardised products (L3/L4)
* After-hours lead activation

**Human agents handle:**

* Complex product configurations
* High-value prospect closing (warm transfers from VoiceB with full context)
* Relationship management for enterprise accounts
* Exceptions and escalations

The result: agents spend 100% of their time on calls that are already qualified, contextualised, and worth their attention.

***

### What This Does to Call Center Economics

| Cost driver                     | Traditional model    | VoiceB model                          |
| ------------------------------- | -------------------- | ------------------------------------- |
| Cost per qualified conversation | £8–25                | £1–7                                  |
| After-hours coverage            | Overtime or none     | Included                              |
| Scaling to handle volume spikes | Hire + train (weeks) | Instant                               |
| Agent attrition impact          | High                 | Reduced (agents do higher-value work) |
| Quality consistency             | Variable             | 94–97% classification accuracy        |

***

### Compliance and Auditability

Enterprise call centers operate in regulated environments. VoiceB is built for this:

* Full call recording and transcription
* Structured outcome logging with audit trail
* GDPR-compliant data handling
* SOC 2 Type II compliance
* Identity validation flows built into the conversation

Every AI decision is explainable and auditable — critical for financial services, insurance, and telecom.

***

### Implementation: How Fast Can You Deploy?

VoiceB integrates with your existing telephony infrastructure. No rip-and-replace required.

* **SIP trunk integration** — works with your existing phone system
* **CRM connectors** — HubSpot, Salesforce, and custom endpoints
* **Script setup** — your sales script, configured and tested in 2–4 weeks
* **KPI baseline** — we define success metrics before go-live, not after

Most enterprises are live with their first VoiceB deployment within 4 weeks.

***

### Enterprise References

* **Vodafone Spain** — SIP-level integration, live production, +41% conversion lift
* **SHV Energy / Primagas** — Full inbound sales automation, +251% lead growth
* **Holaluz** — Inbound qualification and routing, 38–42% transfer rate with full context

[Book a Working Session →](https://voiceb.ai/solutions/appointment-booking)&#x20;

[Read enterprise case studies →](https://docs.voiceb.ai/blog/business-case-shv-energy-primagas-spain)


# The Future of Digital Sales: Assisted Voice Agents That Guide, Not Replace

By Vanessa Goslow & Alex Bisbe. February 2026.

Digital sales teams face a different structural problem than call centers.

It’s not qualification.

It’s **friction inside the online buying journey**.

Most digital acquisition funnels break in the same places:

* Users get confused selecting plans
* Forms feel long and complex
* Price comparisons create hesitation
* Checkout flows lose momentum
* Users abandon before completion

The result is predictable:

* Conversion drops
* CPA increases
* Media efficiency declines
* Sales volatility grows

> Marketing drives traffic.\
> Product builds the funnel.\
> But nobody owns the **live moment of decision**.

That’s where Assisted Voice Agents change the game.

***

### The Shift: From Qualification to Digital Sales Guidance

The future of digital sales is not replacing humans with robots.

It’s deploying **assisted voice agents that guide users in real time through an online process**.

Not collecting personal data.\
Not replacing the web.\
Not forcing a call center interaction.

Instead:

* The agent integrates as a widget on the website.
* It can open pages for the user.
* It can direct navigation.
* It explains options.
* It handles objections.
* It helps the user complete forms — while the user types their own data.

This solves two major problems simultaneously:

1. Conversion friction
2. Data protection & compliance constraints

The user remains in control of their personal information.\
The AI guides. The user inputs.

***

### How Assisted Digital Sales Works in Practice

{% embed url="<https://www.loom.com/share/659df31a983040adadc28ed349c1c3b6>" %}

#### 1️⃣ Web-Guided Assisted Checkout

The voice agent acts as a real-time digital sales assistant embedded directly into the online journey.

It:

* Explains plans in natural language
* Compares options clearly
* Recommends the right configuration
* Opens the correct pages automatically
* Guides the user field-by-field through the form
* Confirms completion before moving forward

Crucially:

The agent **does not collect personal data**.\
The user enters all sensitive information directly into the website.

This maintains full compliance while eliminating friction.

It’s like having a top-performing sales rep sitting next to the customer — without removing control from the user.

***

#### 2️⃣ Assisted Objection Handling During the Flow

Digital journeys lose sales at hesitation points:

“Is this permanent?”\
“Can I cancel?”\
“Is there a better plan?”\
“Why is this more expensive?”

Instead of abandoning the page, the user can simply ask.

The agent:

* Clarifies pricing
* Explains conditions
* Handles common objections
* Reinforces value
* Keeps momentum

This reduces abandonment and stabilizes conversion.

***

#### 3️⃣ Human Intervention Only When Required

If a user explicitly requests human support or falls outside standard flows:

* The agent provides context
* Shares the journey stage
* Transfers only when necessary

Humans become escalation, not default.

***

### Why This Model Is Structurally Superior

Traditional digital funnels assume:

<kbd>Traffic → Static Pages → Conversion</kbd>

Assisted digital sales introduces:

<kbd>Traffic → Interactive Guidance → Assisted Completion → Conversion</kbd>

{% hint style="success" %}

#### The impact:

#### Higher completion rates

#### Lower CPA

#### Reduced dependency on large call center teams

#### More predictable revenue performance

{% endhint %}

This is not another chatbot.

This is **real-time voice-guided digital sales execution**.

***

### Compliance by Design

One of the biggest concerns in AI-driven sales is data protection.

The solution is simple:

> #### The AI does not collect personal data. It does not store sensitive information. It guides users to fill official fields directly in the company’s own environment.

This architecture aligns with strict data protection requirements while preserving performance.

***

### The Strategic Implication for Digital Sales Leaders

Digital sales teams have optimized ads.\
They’ve optimized landing pages.\
They’ve optimized UX.

But the last unoptimized layer is **real-time guidance during the decision moment**.

Assisted Voice Agents fill that gap.

They:

* Convert more of the same traffic
* Stabilize performance
* Reduce funnel leakage
* Enable scalable autonomy

The future of digital sales is not more traffic.

It’s **smarter conversion execution**.

And that future is assisted, voice-driven, and embedded directly inside the web journey.

—\
Vanessa Goslow & Alex Bisbe


# Holaluz Sales Autonomy Project

By Alex Bisbe. April 30th, 2026.

### AI as the First Sales Touchpoint

<div data-with-frame="true"><figure><img src="/files/tFPpMgZ8LEI2Luo24PHG" alt=""><figcaption></figcaption></figure></div>

### Company Overview

Holaluz-Clidom, S.A. is a Spanish green technology company founded in 2010 with a clear mission: **decarbonize the economy through 100% renewable energy**.

The company is widely recognized for its **“Rooftop Revolution”** strategy — transforming residential rooftops into distributed green power plants through:

* Solar panel installations
* Home battery systems
* Integrated renewable energy contracts

#### Public Listing

Holaluz is publicly listed on the Spanish growth market:

* Exchange: BME Growth
* IPO Date: November 29, 2019
* Ticker: **HLZ**

#### Scale & Market Position (Early 2024)

* **335,000+ electricity and maintenance contracts**
* **15,000+ solar installations under management**

Holaluz is not a traditional utility.\
It operates as a vertically integrated green energy platform.

***

## Sales Autonomy Transformation

### AI as the First Sales Touchpoint

As a high-growth B2C energy player, Holaluz faces large inbound traffic volumes across:

* Electricity contracts
* Solar installations
* Maintenance services
* Customer inquiries

The challenge was structural:\
How to ensure sales agents spend time only on reachable, high-intent prospects?

The answer: **AI becomes the first commercial gatekeeper.**

<figure><img src="/files/MTmQJ200w3FIYwvR1rH3" alt=""><figcaption></figcaption></figure>

***

### The Challenge

Inbound traffic included:

* High-intent buyers
* Customer care requests
* Low-intent prospects
* Unreachable contacts
* Voicemail / busy

Historically, all calls reached human agents.

This created:

* Queue congestion
* Agent time lost on non-sales traffic
* Missed revenue outside business hours
* Inefficient cost structure

The system was reactive.\
Holaluz needed a scalable revenue architecture.

***

### The Solution: AI-Led Revenue Orchestration

VoiceB was deployed as the **first autonomous sales layer**.

Every inbound interaction is now handled by AI first.

The AI:

1. Engages in under 4 seconds
2. Qualifies commercial intent
3. Filters non-sales traffic
4. Transfers only high-intent prospects
5. Closes simple products autonomously

Human intervention is triggered only when economically justified.

***

### Operational Results

* **78% of inbound calls were handled automatically by AI**\
  → Cost saving - No hiring required\
  → Lower structural CPA

```mermaid
pie title Inbound calls outcome
  "Handled automatically by AI (78%)" : 78
  "Processed further with human agents (22%)" : 22
```

Now the key revenue metric:

* **30.4% of valid sales were CLOSED autonomously by AI**
  * → Unlocks unlimited capacity to manage lead volume
* **21% of valid sales required human intervention**

```mermaid
pie title Valid sales outcome
  "Closed autonomously by AI (30.4%)" : 30.4
  " Not closed / other (21%)" : 21
  "Required human intervention (48.6%)" : 48.6
```

Meaning:

AI is not just filtering.\
**AI is generating revenue directly.**

Humans now focus only on the conversations that truly need them.\
Capacity expands.\
CPA drops.\
Total sales scale without adding headcount.

This is autonomous sales infrastructure.

***

### Business Impact

#### 24/7 Revenue Activation

* <4 second response time
* No dependency on human schedules

#### 100% Lead Coverage

* Zero missed calls
* No backlog accumulation
* No leakage outside business hours

#### Reduced Operational Load

* Significant reduction in non-commercial interactions
* Increased productivity per FTE
* Higher sales efficiency

#### Structured Revenue Routing

Holaluz moved from reactive call handling to:

* Intent-based routing
* AI-driven qualification
* Structured revenue signals
* Predictable commercial activation

***

### Executive Perspective

> "Our objective with VoiceB is clear: no customer should wait in a queue, and our agents should only speak with people who are reachable and genuinely ready to buy. \
> \
> VoiceB attends every inbound call instantly, qualifies the demand in seconds, and transfers only sales-ready customers to our team. \
> \
> For simpler products, it closes the sale fully autonomously, without needing an agent at all. The result is a better experience for the customer and a much more efficient use of our sales capacity."

<div align="left" data-with-frame="true"><figure><img src="/files/Oh5MfOTLA4dM6P1fZswU" alt="" width="100"><figcaption></figcaption></figure></div>

**Oriol Pou**\
Head of Direct Sales at Holaluz

***

### Conclusion

Holaluz transitioned from:

**Call Center Management → Autonomous Revenue Execution**

AI now controls the front door of sales.\
Every inbound lead is qualified.\
Every interaction is structured.\
Every human minute is monetized.

This is what sales autonomy looks like at scale in a listed energy company.


# Inbound Sales Is Being Replaced — Not Optimized

By Alex Bisbe. January, 20th 2026.

Let’s be clear: **inbound sales as we know it is ending**.\
Not because companies want change—but because **the economics no longer work**.

For two decades, enterprises have tried to *optimize* inbound sales:

* Better IVRs
* Better scripts
* Better training
* Better dashboards

None of it fixed the core problem.

**Humans don’t scale. Software does.**

What’s happening now isn’t incremental improvement. It’s **replacement**.

***

### The uncomfortable truth about inbound sales

Inbound sales sits at the intersection of three structural failures:

1. **Cost inflation**
   * Hiring, training, attrition
   * Idle time, peak-hour overload
   * Variable quality per agent
2. **Experience degradation**
   * Long wait times
   * Repeated questions
   * Inconsistent answers
3. **Conversion ceilings**
   * Scripts are forgotten
   * Objections are mishandled
   * Follow-ups are missed

Enterprises have accepted this as “normal”.

It isn’t.\
It’s a legacy constraint.

***

### Why IVRs, Chatbots, and AI Voice Fell Short

The market didn’t fail to *try*.\
It failed by optimizing the wrong layer.

<figure><img src="/files/ZK5mq0KdionWGJLF4SND" alt=""><figcaption></figcaption></figure>

#### IVRs — Cost Deflection Engines

* Built to **avoid calls**, not close sales
* Rigid decision trees in a **probabilistic buying journey**
* Success measured in **cost reduction**, not revenue impact

**Result:** Friction, abandonment, zero persuasion.

***

#### Chatbots — Text in a Voice Moment

* Text-first in moments where **users want to speak**
* High drop-off, especially on mobile
* Effective for FAQs, **irrelevant for sales conversations**

**Result:** Engagement without conversion.

***

#### AI Voice / Copilots — Assistance, Not Autonomy

* Depend on humans to execute the sale
* Scale limited by agent availability
* Outcomes vary with training, fatigue, and context

**Result:** Better agents, same bottleneck.

***

#### The Core Problem

All of these approaches **optimize humans**.\
None of them **replace the sales function**.

That’s the line Autonomous Agents cross.

***

### The inflection point: AI that can sell

The breakthrough isn’t “AI that talks”.

It’s AI that can:

* Pitch
* Handle objections
* Qualify intent
* Collect structured data
* Route or close autonomously
* Learn from every interaction

This is not conversational UI.\
This is **sales execution**.

Voice + LLMs + real-time decisioning = **Autonomous Inbound Sales**.

***

### The autonomy spectrum (L0 → L4)

Inbound sales doesn’t flip overnight.\
It transitions through **levels of autonomy**.

#### L0 — Fully Human

* Traditional call center
* High cost, low scalability
* Zero leverage

#### L1 — AI Qualification

* AI answers, qualifies, routes
* Humans still sell
* Cost reduction begins

#### L2 — AI Pre-Sales

* AI pitches and handles first objections
* Humans close
* Conversion improves

#### L3 — Assisted Autonomous Sales

* AI sells end-to-end
* Humans handle edge cases only
* This is where economics flip

#### L4 — Fully Autonomous Sales

* AI handles 100% of inbound
* Humans manage strategy, not calls
* Call centers disappear

Most enterprises are already moving—whether they admit it or not.

***

### Why this time is different

Three things changed simultaneously:

1. **LLMs understand intent**, not keywords
2. **Voice synthesis is human-grade**, not robotic
3. **Inference costs collapsed**, enabling scale

This stack didn’t exist 24 months ago.

Now it does.

***

### What happens to humans?

They don’t disappear.\
They **move up the value chain**.

Humans become:

* Exception handlers
* Deal doctors
* Relationship owners
* Strategy and optimization layers

What disappears is **human-as-a-router**.

That role is economically obsolete.

***

### The enterprise reality check

Enterprises don’t adopt Autonomous Sales because it’s cool.\
They adopt it because:

* Inbound demand keeps growing
* CAC keeps rising
* Hiring does not scale
* Boards demand margin

Autonomy is not a vision.\
It’s a **forced move**.

***

### Where VoiceB.ai fits

VoiceB was built on one premise:

> *Inbound sales should be autonomous by default.*

Not assisted.\
Not augmented.\
**Autonomous.**

That means:

* AI that sells, not just answers
* Clear autonomy levels, not vague promises
* Enterprise-grade data control and compliance
* Measurable outcomes: conversion, cost, revenue

We don’t replace agents.\
We replace **the need for agents** in inbound.

***

### The bottom line

Inbound sales will not be optimized into the future.\
It will be **replaced by software**.

The only open question is:

* Who moves first
* And who keeps paying for legacy cost structures

Autonomous Inbound Sales isn’t coming.

It’s already happening.


# VoiceB vs ElevenLabs, Bland, Retell, Decagon, and Salesforce Agentforce

By Alex Bisbe. January 9th, 2026.

AI voice is exploding.\
But most platforms are solving the **wrong problem**.

They optimize for **voice quality**, **agent demos**, or **developer tooling**.\
VoiceB is built for one thing only: **turning inbound demand into revenue at scale**.

This article explains where VoiceB fits — and why enterprises running real inbound sales choose it over alternatives like ElevenLabs, Bland, Retell, Decagon, and Salesforce AgentForce.

***

### The Core Distinction: Voice vs Sales Autonomy

Most AI voice platforms answer one question:

> “Can an AI talk like a human?”

VoiceB answers a different one:

> “Can AI **replace or augment inbound sales at scale**, with measurable outcomes?”

That difference changes everything: product design, pricing, governance, and who actually gets value.

***

### ElevenLabs: Best-in-Class Voice — Not a Sales System

ElevenLabs is a **world-class text-to-speech engine**.\
It excels at realism, emotion, and multilingual voice synthesis.

But ElevenLabs is **not a sales platform**:

* No qualification logic
* No intent classification
* No routing or closing
* No outcomes or revenue attribution

And that’s intentional — it’s infrastructure.

**Important clarification:**\
VoiceB **uses ElevenLabs** (and similar providers) for voice quality where appropriate.

We don’t compete with them.\
We **build on top of them**.

With VoiceB, enterprises get:

* Best-in-class human voice
* Plus autonomous qualification, selling, routing, and closing

**Voice quality is table stakes. Sales autonomy is the differentiator.**

***

### Bland: Great Demos, Not Enterprise Sales Operations

Bland makes it easy to spin up AI callers quickly.\
It’s optimized for experimentation, demos, and lightweight use cases.

Where it breaks down for enterprise inbound sales:

* No autonomy levels (everything is “agent-like”)
* No deterministic outcomes
* No audit-grade decision trails
* Limited control over escalation, compliance, and scale

VoiceB is built for **production sales environments**:

* Telco
* Energy & utilities
* Insurance
* Banking
* Automotive

These are not demo environments. They are **revenue-critical systems**.

***

### Retell: Developer Flexibility Without a Sales Model

Retell provides developer-friendly building blocks to create voice agents.\
It’s flexible, but flexibility shifts the burden to the customer.

That means:

* You design the sales logic
* You define qualification
* You maintain performance
* You prove ROI

VoiceB ships with a **sales operating model**:

* Qualification-first
* Outcome-driven
* Autonomous by default
* Humans only where they add value

Enterprises don’t want to build sales logic from scratch.\
They want **conversion lift**.

***

### Decagon: Support Automation, Not Revenue Generation

Decagon is focused on **AI for customer support**.

That’s a different problem:

* Cost reduction
* Ticket deflection
* Case resolution

Inbound sales is the opposite:

* High intent
* Revenue risk
* Conversion sensitivity

VoiceB is not a support tool adapted for sales.\
It’s a **sales system designed for voice from day one**.

***

### Salesforce AgentForce: CRM Intelligence, Not Voice Autonomy

AgentForce sits inside the Salesforce ecosystem.\
It’s powerful for CRM workflows, copilots, and agent assistance.

But it’s not a voice-native sales engine:

* Voice is an add-on, not the core
* Latency, scale, and concurrency are constrained
* Autonomy is limited by CRM-driven flows

VoiceB operates **outside the CRM**, at the moment of intent:

* WebRTC
* Click-to-call
* Inbound phone
* VoIP

And pushes **structured outcomes back** into Salesforce, HubSpot, or any CRM.

That separation is intentional — and critical for scale.

***

### What VoiceB Actually Is

VoiceB is an **Autonomous Sales Voice Agent (ASVA) platform**.

One platform. Four autonomy levels:

* **L1 – Qualification**: Filter noise, capture intent
* **L2 – Intelligent Routing**: Send the right lead to the right closer
* **L3 – Assisted Autonomous Sales**: AI sells, humans step in only when needed
* **L4 – Full Autonomous Sales**: AI closes end-to-end (where compliant)

You don’t buy products.\
You choose **how autonomous your inbound sales should be**.

***

### Pricing Reflects This Difference

Most platforms price:

* Per minute
* Per agent
* Per seat
* Per API call

VoiceB prices on **outcomes**:

* Per qualified lead
* Per AI-closed sale
* Per transferred revenue opportunity

Because voice is not the value.\
**Conversion is.**

***

### Best of Both Worlds

To be clear:

* Yes, VoiceB sounds human
* Yes, we use ElevenLabs, Bland, and other best-in-class voice providers
* No, voice alone is not enough

With VoiceB, enterprises get:

* **Best-in-class voice quality**
* **Plus autonomous sales intelligence**
* **Plus enterprise governance**
* **Plus measurable revenue impact**

Voice is solved.\
Inbound sales is not.

That’s why VoiceB exists.


# The Future of Landing Pages Is SuperVoice

By Alex Bisbe. December 30th 2025.

Landing pages are broken.\
Not visually. **Operationally**.

We keep adding channels—chat, inbound numbers, click-to-call, callbacks, WhatsApp—thinking more options mean more conversions. In reality, we’re adding **friction**, **drop-off**, and **unqualified noise**.

The future is not another button.

The future is [**SuperVoice**.](https://voiceb.ai/supervoice)

<div data-with-frame="true"><figure><img src="/files/mmnI62Ohf7a4B1vixeFc" alt=""><figcaption></figcaption></figure></div>

***

### From High-Barrier Channels to Zero-Friction Voice

Traditional landing page conversion paths look like this:

* Chat → slow, text-heavy, low intent
* Inbound phone numbers → high friction, device switching
* Click-to-call → carrier dependent, low control
* Callbacks → delayed, low pickup rates
* WhatsApp → context loss, async chaos

Every one of these introduces **delay or effort** at the exact moment when intent is highest.

SuperVoice flips the model.

A **Voice over IP (WebRTC)** widget embedded directly into the landing page lets users speak instantly—no dialing, no app switching, no waiting.

One click. One conversation. Right now.

<figure><img src="/files/CVaaqlGo8kdgcwLKGISk" alt=""><figcaption></figcaption></figure>

***

### Why Voice Must Be AI-First (No Exceptions)

Lowering the barrier means **more conversations**.\
And let’s be honest: many of them will be **non-valid**.

That’s not a problem.\
That’s the point.

This model **only works with an AI agent**:

* AI can handle unlimited concurrent conversations
* AI can qualify, filter, and pre-sell in real time
* AI can absorb noise at near-zero marginal cost
* AI never burns sales team capacity

Humans can’t scale at the top of the funnel.\
AI can.

SuperVoice turns low-barrier voice into a **high-signal funnel**.

***

### Qualify, Pre-Sell, Then Decide: Transfer or Schedule

Once the AI agent qualifies the conversation, the flow becomes deterministic:

* **During working hours**\
  → Transfer the call in real time to a human agent\
  → VoIP → converted seamlessly into a regular phone call
* **Outside working hours**\
  → Schedule a call automatically\
  → Pass full context, intent, and data to the sales team

No cold handoffs.\
No “can you repeat everything?”.\
No wasted agent time.

Humans only speak when it makes commercial sense.

***

### Landing Pages Become Performance Machines

SuperVoice doesn’t replace your landing page.\
It **activates** it.

What changes:

* Higher engagement rates
* More conversations started
* Better-qualified leads
* Faster time-to-sale
* Happier sales teams

This is the fastest way to turn a static landing page into a **real-time sales interface**.

There is no form that can compete with a conversation.\
There is no chat that can compete with voice.

***

### Examples in Automotive and Telco

This model already works in high-consideration, high-volume industries:

* **Automotive**: car configurators, model selection, financing questions
* **Telco**: plans, devices, eligibility, promotions

In both cases, SuperVoice sits inside the digital journey, qualifies intent, and escalates only when value is clear.

<figure><img src="/files/bzYsK8DcLZQdmBuEwVj2" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/nn1Xgg2c4CXBvjTOmDzb" alt=""><figcaption></figcaption></figure>

The result: **more leads, better leads, lower cost per sale**.

***

### Bottom Line

The future of landing pages is not chat.\
It’s not phone numbers.\
It’s not callbacks.

It’s **AI voice, embedded, instant, and scalable**.

SuperVoice turns landing pages into **super-performance pages**—designed for how people actually buy.

If your landing page doesn’t talk back,\
it’s already outdated.

{% hint style="success" %}
**Book a demo** [**here**](https://voiceb.ai/solutions/appointment-booking)**.**
{% endhint %}


# How Telecom Operators Use SuperVoice to Boost Digital Sales

Live Demo Included. By Alex Bisbe, November 28th, 2025.

Telecom customers want speed, clarity, and zero friction when choosing a plan or device online. But the reality is brutal: complex plans, bundled offers, eligibility checks, financing options, and coverage concerns push users to abandon mid-journey. The result is low online conversion, inflated CPAs, and wasted traffic.

**SuperVoice fixes this.**

[SuperVoice](https://voiceb.ai/supervoice) embeds an AI voice agent directly into your digital buying flow, guiding users through plans, devices, offers, and checkout steps in real time. It transforms the website from a self-service catalog into a true digital sales channel.

Our team brings **15 years of execution in telecom digital sales** from Whisbi. With VoiceB.ai, we’re pushing the next evolution: AI-driven, voice-first sales at scale.

***

### The Demo: AI-Assisted Digital Journey for Telecom

This live demo shows how the AI agent steps in exactly where users drop off — comparing plans, clarifying device options, explaining promotions, and walking the customer toward purchase completion. It handles the complexity that typically kills digital sales.

{% embed url="<https://www.loom.com/share/6d38145b009f42a7ad975f2fa6b320d9>" %}

***

### Why Telecom Operators Are Deploying SuperVoice

The business impact is immediate and measurable:

**1. Higher online sales**\
The AI helps users choose the right plan or device, reducing confusion and boosting completions.

**2. Lower abandonment and friction**\
No more decision fatigue. The agent explains, recommends, and resolves doubts in seconds.

**3. Dramatically lower CPA**\
More conversions shift from call center and retail to low-cost digital journeys.

**4. Increased cross-sell and up-sell**\
The AI proactively recommends:

* Higher-speed fiber
* Device financing
* TV and streaming bundles
* Family/second-line deals
* Accessories and protection plans

**5. Unlimited 24/7 sales capacity**\
No queues, no callbacks, just instant voice guidance at the moment of highest intent.

***

### Real-Time Telecom Performance Dashboard

SuperVoice feeds every interaction into a complete dashboard so operators can track:

* Call volume
* Average call duration
* Completed orders
* Conversion rates
* Call recordings
* AI summaries
* Structured customer data
* Complete transcripts

<figure><img src="/files/TNiJX6ScJJskITZ7frb7" alt=""><figcaption></figcaption></figure>

It’s a fully transparent revenue engine powering telecom digital sales.

***

### Telecom KPIs We Improve

Across operators, SuperVoice consistently drives:

* **+18–35% increase** in digital conversion
* **30–60% reduction** in abandonment
* **2–4x uplift** in cross-sell and up-sell performance
* **40–70% reduction** in CPA
* **Instant scalability** during peaks (Black Friday, Christmas, campaigns, promotions)

At telecom scale, these gains are transformational.

***

### Trends: Hyper-Personalized, AI-Driven Digital Journeys

Telecom is shifting from generic funnels to **hyper-personalized digital sales**. SuperVoice accelerates this through memory and adaptive behavior.

The AI can recall:

* Previous interactions
* Preferred plans or devices
* User objections
* Fiber availability and address checks
* Promotions discussed
* Cart progress

This turns the experience into a continuous, personalized sales journey — not a fresh start every time.

The industry is moving from static websites to **dynamic, AI-led conversations** that feel tailored, intelligent, and human.

***

### A Proven Partner for Telecom

We already work with operators like **Vodafone**, backed by a team that has delivered telecom digital-sales innovation for over a decade. SuperVoice is the natural evolution: faster, more scalable, and conversion-obsessed.

If you’re ready to increase online sales and cut your CPA, we can deploy SuperVoice in your telecom funnels in a matter of hours.

Let’s redefine telecom digital sales with AI voice.


# Business Case: Vodafone Spain × VoiceB.ai

By Alex Bisbe. November 27th 2025.

### **Overview**

Vodafone Spain partnered with **VoiceB.ai**, an AI voice platform designed specifically for **inbound sales**, to automate lead qualification, routing, and conversion in its consumer acquisition funnel. The goal: replace manual screening with an intelligent voice layer capable of engaging, qualifying, and transferring calls in real time — without human intervention and without sacrificing sales performance.

<figure><img src="/files/NHAuLMnsd42wdzrCyyxK" alt=""><figcaption></figcaption></figure>

***

### **Challenge**

Vodafone’s inbound traffic exceeds **40,000 monthly calls**, a mix of high-intent prospects and existing customers. Traditional IVR and call-center screening created friction:

* **Long wait times** and repetitive menus led to drop-offs.
* **Low-value or invalid leads** consumed significant agent time.
* **Operational costs** scaled linearly with call volume.

Vodafone needed to **reduce human workload** and **preserve conversion**, ensuring high service quality and seamless user experience.<br>

<figure><img src="/files/8w7AWbu3KnILItq6Snp1" alt=""><figcaption></figcaption></figure>

***

### **Solution: VoiceB.ai**

VoiceB deployed **“Rosa,”** an AI voice agent trained on Vodafone’s real scripts, tone, and qualification logic. The system handles the full inbound flow:

* Detects and filters **invalid or existing-customer calls** automatically.
* Transfers only **qualified leads** to live agents in real time.
* Operates **24/7**, managing off-hour leads and re-engaging dropped calls.

**VoiceB replaces the traditional IVR**, creating a direct, conversational path from inquiry to sale.

<figure><img src="/files/NOqb8CI3t7KmNZipEYw3" alt=""><figcaption></figcaption></figure>

***

### **Results**

Over a four-week pilot, VoiceB delivered measurable improvements:

| KPI                          | Before (IVR + Agents) | After (AI + Agents) | Delta           |
| ---------------------------- | --------------------- | ------------------- | --------------- |
| Filtering efficiency         | 57%                   | **72%**             | +15 pts         |
| Dropped calls                | 13.8%                 | **<8%**             | −6 pts          |
| Conversion (vs total calls)  | Confidential          | **Confidential**    | +0.8 pts        |
| AI accuracy (classification) | -                     | **97.1%**           |                 |
| Short transfers (<2 min)     | 34.7%                 | **15.8%**           | −19 pts         |
| Agent workload reduction     | —                     | **−30%**            | ROI +€13K/month |

The AI maintained **identical sales performance** while **reducing agent workload by \~26 people** (equivalent). The cost savings and 24/7 coverage delivered a **positive ROI from the first month**.

***

#### **What VoiceB Delivers Today**

**Forms → Frictionless Conversations**\
Consumers don’t want to fill out forms or wait for callbacks.\
VoiceB replaces manual flows with natural, real-time conversations that qualify and route instantly.

**Voice as the Interface**\
The shift from *Click & Type* to *Speak & Solve* is here.\
VoiceB enables instant product explanations, plan recommendations, and live transfers.

**Zero-Wait Standard (24/7)**\
Calls are answered in **under 4 seconds**, day and night.\
Low-intent, invalid, and duplicate calls are filtered automatically.

**Frictionless Journeys**\
Voice-guided onboarding and structured data collection reduce drop-off and accelerate time-to-sale.

**Humans Elevated**\
AI handles volume and repetition.\
Human agents focus only on high-value conversations where persuasion and complexity matter.

***

#### **What’s Next — Hyper-Individualized Sales**

This is our next leap: **AI that remembers**.

**Hyper-Individualized Sales** means:

* VoiceB stores and recalls **previous interactions**, intent, objections, and context.
* Messaging and tone become **fully adaptive**, personalized to the caller’s history.
* Offers, scripts, and follow-ups are **context-aware** across sessions and channels.
* Sales journeys become continuous, not isolated — “Rosa” knows the user before the conversation even starts.

This transforms the AI agent from a reactive qualifier into a **persistent sales companion** capable of higher trust, higher relevance, and higher conversion.

***

### **Team & Vision**

VoiceB is founded by **industry veterans in conversational sales technology**.\
\
CEO [**Álex Bisbe**](https://www.linkedin.com/in/alex-bisbe-88571718/) previously led **Whisbi**, the pioneer in real-time video sales for enterprises like Vodafone, AT\&T, Verizon, Telefónica, Toyota and Hyundai.\
\
The team combines deep expertise in **telecom sales, AI orchestration, and voice infrastructure**, with a vision to **transform inbound contact channels into fully automated, revenue-generating systems.**

VoiceB’s mission is clear:

> “To make every inbound conversation instantly solvable — by voice.”

***

### **Next Steps**

With a proven pilot, Vodafone is now evaluating full deployment of **Rosa across 100% of inbound traffic**, replacing the IVR layer.\
Expected impact:

* **€13K+ monthly operational savings**
* **Consistent conversion performance**
* **Scalable 24/7 customer acquisition channel**


# SuperVoice video demo for automotive

From Configurator Assistance to Dealer-Grade Lead Control

Digital automotive funnels don’t fail because of traffic.\
They fail because **too many leads reach dealers without being worth their time**.

SuperVoice — powered by VoiceB — is built to fix exactly that.

***

### The real problem OEMs face today

Automotive brands already have:

* configurators
* forms
* WhatsApp
* chat
* CRM and AI agents (including Salesforce AI)

Yet dealers still complain about the same thing:

> “Most digital leads are not actionable.”

The issue is not *how many* leads you generate.\
It’s **who is allowed to reach the dealer**.

***

### What SuperVoice actually is

**SuperVoice is a real-time voice qualification layer** that sits on top of your digital journeys.

It can:

* guide users inside the configurator
* qualify intent before a form is submitted
* filter, route, or discard leads **before they hit the CRM**

This makes SuperVoice fundamentally different from asynchronous tools like WhatsApp, chat, or AI agents that react *after* the lead already exists.

***

### Example 1: SuperVoice inside the car configurator

The two video demos below show SuperVoice embedded directly in an automotive configurator.

In this setup, SuperVoice:

* engages users while they configure
* answers questions in real time
* compares versions, trims, and options
* handles objections and doubts
* collects structured intent data (model, urgency, financing interest)

**Outcome:**\
A fully qualified, dealer-ready lead — not just a configuration.

> The dealer receives context, intent, and priority — not a raw form.

{% embed url="<https://www.loom.com/share/9c1dbe349dda42a3a7536b5f44abb50a>" %}

*Video below – audio in Spanish 🇪🇸*

{% embed url="<https://www.loom.com/share/c7445a38248c4a0f9d8256c715d4be56>" %}

### The Lead Dashboard: Full Visibility on Every Automotive Interaction

SuperVoice feeds directly into a real-time dashboard that gives OEMs and dealers a clear view of performance. You see total calls, average duration, confirmed orders, and conversion rates in one place — the operational heartbeat of your configurator funnel.\
\
*Video below – audio in Spanish 🇪🇸*

{% embed url="<https://www.loom.com/share/b14930b6d3cc493392f94c6b4c594bd4>" %}

Every interaction is logged with status (Valid/Invalid) and outcomes such as Sold, Talk to Agent, Not Interested, Dropped, or Voicemail. Clicking a call reveals the recording, AI-generated summary, collected data (name, email, phone), and the full transcript.

It’s a streamlined control center that turns voice interactions into measurable, actionable sales opportunities.

***

### Beyond the configurator: VoiceB as a qualification layer

SuperVoice is not limited to configurators.

VoiceB can be integrated into:

* lead forms
* landing pages
* campaign traffic
* click-to-call flows
* inbound numbers

Anywhere a lead is generated, VoiceB can act as a **gatekeeper**.

***

### How VoiceB filters and routes leads

Instead of sending everything to the CRM, VoiceB applies real-time qualification logic:

#### 🔥 Hot leads

* High intent
* Clear model interest
* Urgency detected

**Action:**\
Transferred to a dealer or agent in real time during working hours.

***

#### 🌡 Warm leads

* Interest confirmed
* Timing unclear
* Needs follow-up

**Action:**\
Automatically scheduled for a call or follow-up at the right moment.

***

#### ❌ Invalid or low-quality leads

* No real intent
* Inconsistent answers
* Browsing behavior

**Action:**\
Discarded before they ever reach the dealer or CRM.

***

### Why dealers love this

Because VoiceB changes their reality:

* Fewer leads, higher quality
* No time wasted filtering noise
* Clear priority on every contact
* Better conversion per call

For the first time, **digital leads earn dealer trust**.

***

### Why this cannot be done with WhatsApp, chat, or async AI agents

Asynchronous tools:

* react after the lead exists
* cannot stop bad leads from entering the CRM
* push the filtering burden to dealers

VoiceB is different because it:

* operates **before** the lead is created
* qualifies in real time via voice
* decides who is allowed to exist as a lead

This is not a UX improvement.\
It’s **funnel control**.

***

### The real value for OEMs

With VoiceB:

* marketing generates credible demand
* dealers trust digital channels again
* CRM data quality improves automatically
* conversion increases without increasing traffic

OEMs can change tools, agencies, or CRMs.\
What they can’t afford is going back to a funnel dealers don’t trust.

***

### SuperVoice is just the visible part

The demos show how SuperVoice guides users through the configurator.

But the real product is VoiceB underneath:

* filtering
* qualifying
* routing
* protecting dealers from noise

That’s why VoiceB is not a channel.\
It’s **infrastructure for digital automotive sales**.<br>

***


# Beyond the Form: Why the Future of B2C Inbound is Invisible, Instant, and Conversational

By Alex Bisbe. November 19th 2025

**Category:** Sales Trends / AI Innovation\
**Reading Time:** 4 min

***

In the last decade, "Inbound Sales" meant a very specific, often tedious dance: a customer clicks an ad, lands on a static page, fills out a long form, downloads an eBook, and waits. And waits. Maybe, 24 hours later, a human SDR calls them back—usually when the customer is busy or has already moved on to a competitor.

This model is breaking.

As we look toward the next 5 to 10 years, the B2C landscape is shifting from **attracting traffic** to **eliminating friction**. The future of inbound isn't about getting a lead into a database; it’s about resolving their intent in real-time.

At **VoiceB.ai**, we see this shift happening every day. We are moving away from the "Click and Type" era into the "Speak and Solve" era. Here is what the future of B2C inbound sales looks like—and why the static web form is destined for extinction.

<figure><img src="/files/DRp7Sd0NhqpwymATMGBm" alt=""><figcaption></figcaption></figure>

#### 1. The End of "Contact Us" (The Rise of Voice First)

For years, the screen has been the primary interface for commerce. But typing is inherently high-friction. It requires focus, dexterity, and patience.

The future of inbound is **Conversational Commerce**, dominated by voice.

* **Current State:** A user reads a pricing page and types questions into a chatbot that replies with pre-set links.
* **The Future:** A user simply asks, "I need a plan that covers three users and includes international calling." An AI Voice Agent understands the intent, negotiates the best package, and processes the upgrade instantly.

This is where **VoiceB** is leading the charge. We are transitioning web experiences from visual browsing to verbal resolution. The goal is to let your customers speak to your brand as easily as they speak to a friend, with zero latency and 100% context.

#### 2. From Personalization to "Hyper-Individualization"

Today, "personalization" often just means an email subject line that says “Hello, \[Name].”

In the future, Generative AI will govern the entire inbound experience. When a prospect interacts with your brand, the AI won't just read a script; it will generate a unique interaction based on that specific user’s mood, history, and hesitation level.

If a customer sounds unsure, the AI agent adapts its tone to be more empathetic and educational. If a customer is in a rush, the AI switches to a transactional, efficient mode. This **dynamic adaptation**—which our **SuperVoice** technology is already pioneering—turns a generic sales pitch into a tailored consultation for a "segment of one."

#### 3. The Collapse of the Sales Funnel (Instant Gratification)

The traditional funnel (Awareness -> Interest -> Decision) is collapsing into a single moment. In an on-demand economy, waiting for a callback is the number one killer of conversion rates.

* **The "Zero-Wait" Standard:** Future inbound strategies will rely on AI agents that never sleep. Whether a lead comes in at 2:00 PM or 2:00 AM, the "first response time" must be under 4 seconds.
* **Predictive Qualification:** Before a human agent ever picks up the phone, AI will have already analyzed the lead. Tools like **SuperQualification** are the precursors to this future, ensuring that human sales teams stop wasting time on invalid numbers or low-intent leads and focus 100% on closing ready-to-buy customers.

#### 4. The "Phygital" Merger and Invisible Payments

We are heading toward an era where the line between online and offline dissolves. Inbound sales will happen in "Phygital" (Physical + Digital) environments.

Imagine a customer seeing a product on social media or in a virtual environment. Instead of clicking through three pages to checkout, they authorize the purchase via voice biometrics. The friction of the shopping cart disappears. The sale feels invisible.

For service-based B2C companies (insurance, telco, utilities), this means complex onboarding processes that used to take 20 minutes of typing will be handled by an AI Voice Agent in a 3-minute conversation while the customer is driving to work.

#### 5. The Human Role: Empathy and Complexity

Does this mean the human salesperson is obsolete? **Absolutely not.**

In the future of B2C inbound, the role of the human elevates. As AI handles the transactional, repetitive, and qualification tasks (the "grunt work"), humans will be freed to handle:

* High-ticket negotiations.
* Complex problem solving.
* Emotional connection and brand loyalty.

The winning companies will be those that master the **handoff**. The AI tees up the perfect shot, and the human swings the club. That is the core philosophy behind VoiceB.ai: **Augmentation, not replacement.**

#### Summary: The Shift in Paradigms

| Traditional Inbound                 | The Future of Inbound (VoiceB Era)      |
| ----------------------------------- | --------------------------------------- |
| **Focus:** Traffic & Forms          | **Focus:** Conversations & Intent       |
| **Speed:** "We'll call you back"    | **Speed:** "We're listening now"        |
| **Interface:** Keyboards & Screens  | **Interface:** Voice & Natural Language |
| **Data:** Static Cookies            | **Data:** Real-time Sentiment & Context |
| **Qualification:** Manual Filtering | **Qualification:** AI-Driven at Scale   |

#### Are You Ready to Listen?

The technology to power this future isn't science fiction; it is already here. The brands that cling to static forms and slow response times will be left behind by those that embrace the speed and humanity of voice AI.

At **VoiceB.ai**, we aren't just predicting this future—we are building the infrastructure for it.

[**Book a demo today**](https://voiceb.ai/solutions/appointment-booking) and hear what your inbound sales could sound like.


# Business Case: SHV Energy / Primagas Spain

How Primagas increased inbound sales by 153%, grew leads by 251%, and cut CPA by 37% with AI Voice Agents.

By Alex Bisbe · Updated March 2026

***

**Primagas**, part of **SHV Energy Group**, is a leading European energy company specializing in the distribution of Propane, Biopropane, and LNG (Liquefied Natural Gas) across rural and off-grid areas. Operating in Spain since 1996 and part of a global group active in 26 countries, Primagas manages high volumes of inbound demand from digital campaigns and web traffic.

***

### The Challenge

Primagas was generating thousands of inbound leads from Meta, Google Ads, and their website — but a traditional call center was the bottleneck between intent and revenue.

The core problems:

* **Capacity ceiling** — the call center could only handle a fixed number of contacts per day. Scaling leads meant scaling headcount. The economics didn't work.
* **Response latency** — leads generated outside business hours waited hours or were never called back. Intent decays fast.
* **No real-time filtering** — valid and invalid demand reached human agents equally, wasting capacity on contacts that would never convert.
* **Rising CPA** — acquisition costs were increasing with no scalable answer.

The team needed to break the fixed relationship between lead volume and headcount — and engage every inbound contact instantly, around the clock.

<figure><img src="/files/VJu0MV1K6ozMHj1FEaCJ" alt=""><figcaption></figcaption></figure>

***

### The Solution

Primagas deployed VoiceB.ai as the autonomous voice layer across their entire inbound sales funnel, integrating directly with their "Quiero Gas" web widget, click-to-call forms, and Meta and Google Ads lead flows.

**VoiceB's AI agent — Rosa — now handles every inbound contact automatically:**

* **Engages instantly** — responds in under 4 seconds via web widget or phone, 24/7, 365 days a year
* **Qualifies in real time** — location, gas needs, property type, and identity captured in a natural conversation
* **Filters invalid demand** — duplicates, out-of-scope requests, and existing customers classified and excluded automatically
* **Routes to Salesforce** — every valid contact, with full structured data and conversation context, lands directly in Salesforce for the Account Executive team

**Two paths from there — both autonomous in initiation:**

1. **SOLD** — Rosa closes the sale end-to-end on the call. The Account Executive reviews the structured output in Salesforce and sends the contract for signature. No human involved in the sales conversation.
2. **Talk to Agent** — Rosa qualifies, captures data, and routes to a human AE for complex cases. The AE receives full context, completes the conversation, and sends contract if there's interest.

The call center is gone. VoiceB is the first — and often only — point of contact for every inbound lead Primagas receives.

***

### Results — 10 Months in Production

*Measured against the same period before VoiceB deployment. Live data, not a pilot.*

| KPI                                 | Before VoiceB | After VoiceB | Change    |
| ----------------------------------- | ------------- | ------------ | --------- |
| Sales Volume                        | Baseline      | —            | **+153%** |
| Leads Generated                     | Baseline      | —            | **+251%** |
| Cost per Acquisition                | Baseline      | —            | **−37%**  |
| Autonomous close rate (valid calls) | 0%            | **67.8%**    | —         |
| Call center headcount               | Active        | **Closed**   | 100%      |

**Over 1,335 real production calls analyzed:**

* 280 sales closed autonomously by Rosa — no human in the conversation
* 133 transferred to Account Executives with full Salesforce context
* Peak month (December 2025): **87.7% of valid sales closed autonomously**
* Last 3 months (Dec 2025–Feb 2026): **74.6% autonomous close rate**

The causal chain is direct: closing the call center removed the capacity ceiling. Removing the ceiling allowed Primagas to scale lead investment aggressively. More leads at lower operational cost per contact drove +251% lead growth, +153% sales, and −37% CPA — simultaneously.

***

### What Made the Difference

**Unlimited capacity changed the investment logic.** With a call center, every additional lead had a marginal cost in agent time. Primagas had to be conservative with ad spend. With VoiceB, the marginal cost of an additional lead approaches zero. That changed how Primagas could think about Meta and Google investment — and they leaned in.

**Speed is conversion.** Every inbound contact is engaged in under 4 seconds. The previous model meant leads waited hours or were never called back. Rosa picks up in under 4 seconds, every time, including weekends, evenings, and campaign spikes.

**Salesforce integration made it enterprise-grade.** Every call outcome — sold, transferred, invalid — writes structured data directly to Salesforce. Account Executives work from a clean pipeline, not raw call logs. The AE's job is to review, validate, and send contracts — not to qualify from scratch.

**Autonomous close rate compound over time.** Starting at 47% in March 2025, Rosa's autonomous close rate on valid calls reached 87.7% by December 2025. The system improves as it handles more volume.

***

### Testimonial

> *"VoiceB has provided more than a decade of lead conversion know-how combined with a powerful AI platform that has delivered real, measurable results — increasing our sales by 153%, leads by 251%, and reducing CPA by 37%. Super happy with VoiceB's solutions and experienced team."*
>
> — **Jordi Alonso**, Marketing Director, Primagas (SHV Energy Group)

<div align="left" data-with-frame="true"><figure><img src="/files/f2b9Pl8yrfZTbCUndlSV" alt="" width="114"><figcaption></figcaption></figure></div>

***

### Conclusion

Primagas didn't optimize their call center. They closed it.

By replacing manual inbound qualification with VoiceB's autonomous voice agents, Primagas broke the fixed relationship between lead volume and operational cost. The result: unlimited lead capacity, a 67.8% autonomous close rate across 12 months of live production, and a leaner, faster, more profitable sales operation.

153% more sales. 251% more leads. 37% lower CPA. Zero call center.

*VoiceB is now the only first point of contact for every inbound lead Primagas receives.*

***

*Interested in similar results for your energy or utility business?* [*Book a demo →*](https://voiceb.ai/solutions/appointment-booking)

***


# How VoiceB SuperQualification Transforms Inbound Sales

Alex Bisbe / September 28th, 2025

One of the biggest challenges in inbound sales today is **lead quality**. Whether leads come from paid campaigns, organic traffic, or third-party providers, sales teams often waste valuable time handling calls that never had real purchase intent. This is a costly problem: agents spend hours on invalid leads while genuine prospects are left waiting or overlooked.

At **VoiceB.ai**, we built [**SuperQualification**](/blog/how-voiceb-superqualification-transforms-inbound-sales) to solve this exact issue. Our AI-powered system analyzes every inbound lead with **96% accuracy (and improving)**, instantly classifying whether it’s valid or invalid. From there, the system can either:

* Complete the sale directly with AI, or
* Execute a **live transfer to the right sales agent in real time.**

***

### Smarter Routing = Higher Conversions

VoiceB doesn’t just transfer calls randomly—it routes them intelligently. By using **prompt terms and contextual information from the conversation**, the system can decide which agent or department should receive the lead. For example, calls can be routed based on:

* Type of product or service requested
* Language or region
* Customer status (new vs. existing)
* Priority keywords that indicate higher intent

This ensures every valid lead is connected with the **best possible agent** for conversion, instead of being lost in generic call queues.

***

### Infrastructure That Others Don’t Offer

Unlike many platforms that only pass the call forward, VoiceB was designed from the ground up to **manage and track leads efficiently—not just calls.**

We’ve built an infrastructure using **SIP Trunks in Twilio** that guarantees:

* **Caller ID preservation**: the transfer is always sent with the **user’s phone number**, so agents immediately recognize who is calling.
* **Better tracking and identification**: every call can be tied back to its source campaign, traffic channel, or sales outcome.
* **Compliance and transparency**: full traceability for auditing and reporting.

This level of routing precision and lead attribution is unique in the market—other platforms often fail to keep the original caller’s phone number when transferring, which breaks tracking and complicates reporting.

***

### The Impact: From 100 Leads to Real Sales

The diagram above shows a **real-world performance example** with VoiceB:

* Start with **100 leads**.
* **50%** are immediately filtered out as invalid (support requests, existing customers, wrong numbers, etc.).
* Of the remaining valid leads, **12% convert directly with AI**, while **38% are routed to human agents**.
* Human agents then close another **10%** of total leads.

That means:

✅ **22% of all inbound leads convert into sales** (12% AI + 10% human).\
❌ Meanwhile, agents are spared from wasting time on the 50% of invalid calls.

{% hint style="info" %}
These numbers reflect **real client performance** with VoiceB. Results may vary depending on the **product, commercial strategy, lead source, and market conditions**—but the efficiency gains and quality filtering are consistent across industries.
{% endhint %}

<figure><img src="/files/BNWurQjmQreCh503v0qK" alt=""><figcaption></figcaption></figure>

***

### Why [SuperQualification](/blog/how-voiceb-superqualification-transforms-inbound-sales) Matters

* **Higher efficiency**: Sales reps talk only to real prospects.
* **Smart routing**: Every valid lead goes to the **right agent** in real time.
* **Better ROI**: Marketing dollars aren’t wasted on invalid traffic.
* **True attribution**: Transfers preserve the **customer’s original phone number**, ensuring full visibility.
* **Scalability**: The system adapts to high call volumes without adding headcount.

Inbound sales is bigger than eCommerce, yet many companies still rely on outdated IVRs and manual processes. With VoiceB [SuperQualification](/blog/how-voiceb-superqualification-transforms-inbound-sales), enterprises in telecom, banking, insurance, utilities, and automotive are already seeing **up to 40% higher conversion rates** compared to human-only teams.

***

If your company—or your clients—are struggling with lead quality, VoiceB [SuperQualification](/blog/how-voiceb-superqualification-transforms-inbound-sales) can help. We’d be happy to share more details and results from live deployments.

{% hint style="success" %}
[Book a meeting](https://voiceb.ai/solutions/appointment-booking)
{% endhint %}


# Why Voice Still Matters in Digital Sales—and How AI Is Changing the Game

By Alex Bisbe. September 11th, 2025

In a world dominated by clicks, forms, and one-click checkouts, it’s easy to assume that **voice sales are fading away**. After all, ecommerce platforms like Amazon, Alibaba, or Mercado Libre generate trillions in online transactions every year.

But if we look closer, a surprising truth emerges: **voice remains the dominant channel for some of the world’s biggest industries**—industries that together represent more GMV than ecommerce itself.

***

### The Numbers: Ecommerce vs. Voice-Driven Sales

* **Global Ecommerce GMV (2023):** \~$6.3 trillion (Statista)
* **Industries heavily reliant on phone/voice sales:** \~$10–12 trillion (Telecom, Insurance, Banking, Automotive, Home Services, Travel, etc.)

👉 Put simply: while ecommerce gets the headlines, **the bigger opportunity lies in voice-driven industries.**

<figure><img src="/files/6GEMS9y6uDavKrBeBBUq" alt=""><figcaption></figcaption></figure>

***

### Why These Industries Rely on Voice

#### 1. Complex & High-Value Products

Telecom bundles, insurance policies, mortgages, and cars aren’t impulse buys. Customers need **guidance to navigate complexity**—something far easier to provide in a conversation than through a static web form.

#### 2. Trust & Human Reassurance

When buying a life insurance policy, signing a loan, or booking an expensive trip, people want to **hear a voice**. A human (or human-like AI) builds trust that a webpage simply can’t.

#### 3. Regulatory & Compliance

In industries like banking and insurance, conversations must be recorded for compliance. Phone calls provide an **auditable trail of consent and disclosures**—a legal safeguard that “click-to-accept” isn’t always enough.

#### 4. Lead-Driven Acquisition

Home services, real estate, and B2B SaaS don’t sell in one click. They sell by **capturing a lead and converting it over the phone**. Calls consistently deliver conversion rates **10–15x higher than web forms**.

#### 5. Cross-Sell & Up-Sell Power

Voice allows dynamic upselling: “Would you like to add roadside assistance?” is more effective when asked conversationally than when buried in a checkout page.

#### 6. Demographics & Urgency

Older demographics still prefer calling. And in urgent scenarios—like a car breakdown or a heating failure—people don’t browse. They **pick up the phone**.

***

### The Untapped Opportunity

Ecommerce is efficient and optimized. Amazon already won that game.

But in **multi-trillion-dollar industries still dominated by phone interactions**, the sales process is **inefficient, expensive, and frustrating**—for both companies and customers.

* Call centers filter millions of non-sales calls manually.
* Sales agents spend time on low-value leads instead of high-intent buyers.
* Customer journeys get stuck in IVRs or abandoned mid-form.

This is where the real disruption is waiting to happen.

***

### Enter Voice AI: The Next Revolution in Sales

At VoiceB, we believe **voice is the next frontier for digital sales.**

Our AI voice agents—what we call [**SuperVoice**](https://voiceb.ai/supervoice)—are designed to:

* **Qualify leads in real time**: filter out invalid traffic before it reaches your reps.
* **Handle sales conversations end-to-end**: pitch, answer objections, collect data, and even close sales when appropriate.
* **Scale instantly**: unlike human agents, AI can take unlimited simultaneous calls.
* **Integrate seamlessly**: with your website (click-to-call, WebRTC), CRM, and phone system.
* **Boost conversion rates**: by widening the funnel without adding more human capacity.

***

### Why [VoiceB](https://voiceb.ai/)

Voice is where the **biggest sales volumes still live**—and where AI can deliver the **biggest performance gains**.

While ecommerce platforms perfected the digital cart, **we’re perfecting the digital conversation**.

If your business depends on inbound calls, leads, or phone-driven sales, [**VoiceB.ai** ](https://voiceb.ai/)**is your unfair advantage**:

* More conversions from the same traffic.
* [Lower cost per sale](https://voiceb.ai/pricing).
* Happier customers, happier sales teams.

***

✅ **The future of sales isn’t just clicks—it’s conversations.**\
And at VoiceB, we’re making sure those conversations work for you.


# Improve Your Lead Generation from Existing Traffic with AI Voice + WebRTC

By Alex Bisbe. September 8th, 2025.

Marketers often obsess over driving more traffic—bigger ad budgets, better SEO, new campaigns. But the real growth opportunity may be hiding in plain sight: **the visitors already landing on your website every day.**

The problem is not traffic. It’s **conversion.** Too many visitors bounce because the channel you’re offering—phone, form, or chat—doesn’t match the way they want to engage.

With [**VoiceB SuperVoice**](https://voiceb.ai/supervoice), you can unlock that hidden demand. By adding just two channels—**WebRTC AI Voice** and **Callback with AI + Humans**—you can capture more leads, close more sales, and do it without spending a single dollar more on ads.\
\
![](/files/sPSOohCMh1urmQ6QUQ6E)

***

### Results by Channel: What the Data Shows

Not all engagement channels perform the same. Benchmarks across telecom, utilities, and insurance landing pages show wide differences in both **lead volume per 1,000 visits** and **conversion rates**:

* **Phone visible**: \~10 leads, 15% conversion
* **Callback (AI + Human)**: \~20 leads, 20% conversion
* **Chat widget**: \~35 leads, 7.5% conversion
* **WebRTC AI Voice**: \~50 leads, 25% conversion

<figure><img src="/files/Hv0NsK8ApCNOrShP4P6T" alt=""><figcaption></figcaption></figure>

*(See chart: Inbound Sales Funnel Performance by Channel)*

{% hint style="success" %}
The impact is clear: **adding WebRTC AI Voice can increase lead volume by +43% compared to using only callbacks and chat**. For a site with 100,000 visits, that means an additional **15,000 qualified leads**, generated from traffic you already have.
{% endhint %}

***

### Why AI Unlocks the Potential of WebRTC

In the past, enabling real-time voice calls from a landing page was risky. Peaks of calls could overwhelm call centers, and many conversations were invalid or unqualified.

With[ **VoiceB SuperVoice**](https://voiceb.ai/supervoice), every inbound call is answered instantly by an AI agent that:

* **Filters invalid calls** (wrong numbers, test calls, irrelevant inquiries).
* **Handles FAQs and objections** using your own scripts, pricing, and product updates.
* **Escalates only qualified prospects** to human agents, with a transcript for full context.

The results are measurable:

* **+46.7% improvement in conversion rates** (15% → 22%).
* **62% fewer agent hours needed**, since only 38% of calls escalate to humans.

In other words, WebRTC no longer means “more calls than you can handle.” Now it means **more qualified sales opportunities**, handled automatically at scale.

***

### Generational Fit: Covering Every Segment with Two Channels

Generational research proves customers don’t all engage the same way:

* **Gen Z & Millennials**: digital-first, mobile-native, and less likely to dial a traditional phone number. They expect **instant, click-to-talk options like WebRTC voice** 【Zendesk CX Trends 2023】【Salesforce State of the Connected Customer】.
* **Gen X & Baby Boomers**: more comfortable with **callbacks and phone conversations**, valuing the reassurance of a real voice on the line 【Zendesk CX Trends 2023】.

👉 This is why VoiceB recommends **just two channels**:

* **WebRTC AI Voice** → frictionless for younger, digital buyers.
* **Callback with AI + Humans** → reassuring for older customers who still trust phone interactions.

Together, these two channels cover **every generation**, maximizing conversion from existing traffic without cluttering your landing page.

***

### The Future of Inbound Sales

The fastest way to grow isn’t necessarily more traffic. It’s **capturing more value from the traffic you already have**.

By adding **WebRTC AI Voice** to your landing page, you can increase lead volume by over **40%**, while callbacks ensure you don’t miss customers who still prefer phone-based engagement..

👉 Talk to VoiceB today and see how you can turn your existing web traffic into your best-performing sales funnel.

***

### Sources

* Google/Ipsos: The Role of Click to Call in the Path to Purchase
* [CallRail: Optimize Landing Pages with Click-to-Call](https://www.callrail.com/blog/optimize-landing-page-click-to-call-conversions)
* Forrester Consulting: The Total Economic Impact of Click-to-Call
* [Unbounce: Average Landing Page Conversion Rates](https://unbounce.com/average-conversion-rates-landing-pages)
* [Ruler Analytics (2025): Conversion Rate by Industry](https://www.ruleranalytics.com/blog/insight/conversion-rate-by-industry)
* Zendesk CX Trends 2023
* [Salesforce State of the Connected Customer](https://www.salesforce.com/resources/research-reports/state-of-the-connected-customer/)


# Video Case Study: How We Boosted Conversion by 46.7%

By Vanessa Goslow. September 5th, 2025

At VoiceB, we believe AI in the contact center should deliver measurable results—not just hype. While recent studies (like the MIT report noting that 95% of AI projects show no measurable ROI) highlight the industry’s challenges, our approach is different: we only charge for results, and our AI agents are laser-focused on driving sales.

In this video, we walk through a real example with a leading European energy company that partnered with us just two weeks ago. The early results are impressive:

✅ **Invalid leads automatically filtered out** – Nearly 50% of inbound calls were handled instantly by the AI, never reaching a human agent. This means huge savings in operational costs.

✅ **AI that closes deals on its own** – Our AI matched human conversion rates, closing around **10–15% of sales fully unassisted**, no human required.

✅ **AI-powered handoffs** – For the 30–40% of conversations that needed a human touch, the AI passed structured call summaries—including customer intent and objections—so agents could focus on simply closing the deal.

The outcome?

* **Before VoiceB:** 15% conversion rate.
* **After VoiceB:** 22% conversion rate.

That’s a **46.7% improvement in sales performance**, achieved without adding a single new human agent.

This case shows how VoiceB doesn’t just boost efficiency—it transforms sales results. And this is just the beginning.

👉 Watch the video above to see the breakdown, and let’s talk about how VoiceB can help your business achieve the same lift.

{% embed url="<https://www.loom.com/share/29b5bb70047c411bb0c033dd4862c82c?sid=c4daa7f8-58b5-4dc3-9b81-b30d660d9a6d>" %}

By [Vanessa Goslow](https://www.linkedin.com/in/vanessagoslow/)


# Conversion Rates That Redefine Inbound Sales

by [Alex Bisbe](https://www.linkedin.com/in/alex-bisbe-88571718/). August 26th, 2025

At **VoiceB.ai**, our mission is simple: help enterprises sell more without adding more agents. The way we do this is best explained in one metric — **conversion rates**.

***

### AI Conversion Rates That Match Humans

In our first live campaigns, VoiceB’s AI agents are delivering conversion rates that **match human performance**:

* **10–15% sales closed by AI** alone, fully unassisted.
* Example: In an energy campaign with a top Spanish utility, our AI closed **11.7% of all leads** — directly in line with the human average of 10–15%.

This means AI is not a supporting tool. It is a real sales closer.

***

### Smarter Funnel, Higher Overall Conversions

<div align="center"><figure><img src="/files/QxuojMbr4oaoavdyGhE2" alt="" width="563"><figcaption></figcaption></figure></div>

The strength of VoiceB lies in how it handles every single inbound call:

* **\~50% filtered** as invalid traffic (existing customers, hang-ups, fake numbers).
* **10–15% fully closed** by the AI agent.
* **35–40% qualified as Talk To Agent**, with summaries, objections, and intent captured for the human team.

This structure **widens the funnel**:\
Humans only spend time on leads with real intent, while AI handles filtering and closing. The result is **more total sales from the same lead volume**.

At **VoiceB.ai**, we don’t just promise results — we **track and deliver real outcomes**. Every call is logged with its status and conversion result, whether it’s **Sold**, **Talk to Agent**, or **Not Interested**. For example, in recent live campaigns you can see AI outcomes in action: one call marked **Sold** after 6 minutes, others flagged as **Talk to Agent** for human follow-up, and invalid leads instantly filtered out.&#x20;

<div align="center"><figure><img src="/files/WA30VRQt5QAuBmzbeybw" alt="" width="563"><figcaption></figcaption></figure></div>

This level of transparency ensures that enterprises always know exactly how many sales the AI is closing and how many qualified opportunities are being handed over to the team.

***

### Why Conversion Rates Matter

Traditional call centers face staffing bottlenecks. More calls mean more queues, missed opportunities, and frustrated customers. With AI:

* Every call is answered instantly — no missed opportunities.
* Invalid calls never drain human resources.
* Human agents get only the most promising leads, increasing their personal close rates.

The outcome: **higher conversion, no extra headcount.**

***

### The Bottom Line

VoiceB.ai proves that **AI can sell**. Not only does it achieve human-level conversion rates, but it also reshapes the funnel to ensure every lead is maximized.

👉 With 10–15% direct AI conversion and 35–40% qualified handoffs, enterprises can achieve **more sales with the same team size** — a true breakthrough in inbound sales.


# Widening the Funnel Without Adding Agents

VoiceB.ai’s First B2C Results: More Sales Without More Agents - Business Case

At VoiceB.ai, we are rethinking how enterprises approach inbound sales. Traditional call centers face a fundamental bottleneck: **human dimensioning.** Every lead that calls in needs to be picked up by an agent, creating constraints on staffing, queues, and ultimately on revenue growth.

AI changes the equation.

### Why VoiceB is Different

Unlike AI tools built for support or qualification only, VoiceB is a **sales-focused AI agent**. Our platform is designed to:

* **Filter invalid leads automatically.** Around **50% of inbound calls are discarded by the AI** as invalid (existing customers, hang-ups, test calls, fake numbers). These never reach a human agent, saving enormous operational cost.
* **Close sales independently.** Our AI achieves **10–15% pure conversion rates** — fully unassisted, with no human intervention. This is on par with what human agents deliver today.
* **Qualify and prepare the rest.** The remaining **35–40% of leads are flagged as&#x20;*****Talk To Agent***. For these, the AI provides a **structured call summary**, including customer intent, objections raised, and data collected. Human agents call back fully prepared, drastically increasing their chance of closing.

This three-step split — **50% filtered, 10–15% closed, 35–40% qualified for callback** — creates a more efficient, more scalable funnel.

<figure><img src="/files/4aofm0zFUtH9nIaNncf9" alt=""><figcaption></figcaption></figure>

### Wider Funnel, No More Agents

The beauty of this approach lies in **funnel expansion without staffing growth.**

* With WebRTC and VoIP, customers can connect with the AI instantly, anytime, without queuing.
* Because AI filters out half of the traffic, **call centers never get saturated.**
* More leads come in, but human teams only see the qualified fraction, supported by AI summaries.

In other words: **more leads = more sales, without hiring more agents.**

### Real-World Proof

In our first live campaigns with major enterprises:

* **Conversion rates match human agents.** For example, in an energy sales campaign with a Top Spanish utility company, AI agents closed **11.7% of leads** on their own — right in line with the 10–15% human average.
* **AI-qualified leads outperform expectations.** Around **40% of calls were marked as Talk To Agent**, with full call summaries sent to the CRM. Agents could prioritize callbacks and focus on deals most likely to close.
* **Invalid traffic drops out automatically.** Nearly half of all inbound calls never reach a human because AI detects non-sales intent immediately.

<figure><img src="/files/pBVWtX1iMCuXOL4A8L5u" alt=""><figcaption></figcaption></figure>

### VoiceB’s Unique Differentiators in AI Voice Sales 🚀

VoiceB stands out in the AI voice sales market through several key innovative approaches:

#### 1. Sales-Focused AI Strategy 🎯

* **Widening the Sales Funnel:** Unlike traditional lead qualification methods that discard potential leads, VoiceB facilitates customer connections and filters *within conversations*.
* **24/7 Service Availability:** Providing constant customer engagement without human resource limitations.
* **Rapid Response:** Guaranteeing customer service with calls answered in **less than four seconds**.

#### 2. Advanced Knowledge Base Management 📚

* **Automated Web Scraping:** Updating AI agent knowledge every 24 hours.
* **Real-Time Content Updates:** Scanning websites, social media, and optionally API integrations to keep information current.
* **95% Automatic Updates:** Minimizing manual intervention and keeping scripts always aligned with the latest product/pricing.

#### 3. Innovative Lead Handling Metrics 📊

* **Conversion Rates:** AI agents achieving **10–12% conversion rates**, replicating the full sales process at human levels.
* **Lead Classification:**
  * 50% invalid (filtered out)
  * 10% closed entirely by AI
  * 40% qualified for human agent follow-up

#### 4. Flexible Integration Capabilities 🔗

* **CRM Integrations:** Native integrations with Salesforce and HubSpot.
* **Webhook Support:** Easy integration with virtually any system.
* **Multiple Communication Channels:** WebRTC, click-to-call, callback, and public phone numbers.

#### 5. Technological Infrastructure 💻

* **Cloud Infrastructure:** Built on Vercel (scalable and secure).
* **Tech Stack:** Next.js, Clerk for login/security, and multiple AI voice providers (ElevenLabs, Bland.ai).
* **Provider Switching:** Ability to swap between AI voice providers instantly to avoid latency or quality issues.

#### 6. Unique Selling Proposition 🌟

VoiceB isn’t just about cost-saving — it’s about **increasing sales conversion dramatically** by:

* Reducing friction in customer interactions.
* Providing intelligent, context-aware AI agents.
* Enabling seamless, prepared handoffs between AI and human agents.

Our philosophy: **operate lean while delivering enterprise-grade solutions**, focusing on quality over quantity in both team size and customer interactions.

### The Strategic Advantage

VoiceB.ai is not just about cost reduction. It’s about **scaling revenue**. By widening the top of the funnel, engaging every possible lead, and letting AI handle the sorting and selling, enterprises can:

* Capture more opportunities without additional headcount.
* Reduce missed calls and queues to zero.
* Deliver consistent, immediate engagement 24/7.
* Equip human agents with better-prepared, higher-quality leads.

The result? **Higher sales, lower operational friction, and a customer journey that feels seamless.**

***

👉 At VoiceB.ai, we believe sales teams should not fear AI — they should embrace it. Our AI agents are already proving that they can **filter, sell, and qualify at scale**, unlocking a new model for digital-first sales.

***


# Top AI Voice Sales Solutions in 2025: What You Should Know

And Why VoiceB.ai Leads the Pack

<sup>Alex Bisbe    7/23/2025     4min read</sup>

In today’s fast-moving digital world, **AI-powered voice agents are no longer just a nice-to-have** — they’re becoming essential for B2B and B2C companies looking to scale sales without scaling headcount.

But not all voice AI is built for sales.

In this post, we’ll explore the top AI voice solutions in 2025 that deliver real sales results — and explain why **VoiceB.ai is uniquely positioned** to lead this emerging space.

***

### 🚀 Why AI Voice Agents Matter for Sales

Voice is still the **most effective channel for closing sales**, handling objections, and qualifying leads. But human call centers are:

* Expensive
* Inconsistent
* Impossible to scale instantly

AI voice agents offer a smarter path forward:

* Available 24/7
* Always pitch-perfect
* Infinitely scalable

But only a few platforms are truly optimized for **converting leads**, not just answering FAQs.

***

### 🎯 What Makes a Voice AI Good for Sales?

To work in real sales environments (both B2B and B2C), a voice AI solution needs:

* A **trained sales model** that knows how to pitch, qualify, and handle objections
* Ability to **guide users through online journeys** (like subscriptions, bookings, or purchases)
* **Real-time data capture and scoring** to trigger next steps or CRM actions
* Integration with **click-to-call**, VoIP buttons, and inbound phone numbers
* Tools for compliance: **transcripts, audits, call recordings**

That’s exactly what we built at VoiceB.ai. But let’s look at what’s out there.

***

### 🧠 Top AI Voice Sales Platforms (2025)

#### 1. **VoiceB.ai** – *The Specialist in Inbound Sales Automation*

* 100% AI-powered voice agents for **replacing or assisting human sales agents**
* Trained on **scripts + objections + data capture logic**
* Works with inbound traffic via **SuperVoice** and **SuperQualification**
* Proven conversion lifts in industries like **telco, insurance, energy**
* Performance-based pricing available

#### 2. **Invoca**

* Best known for **call tracking and conversation intelligence**
* Focuses on **analyzing agent performance**, not replacing them
* Great for **post-call insights**, media optimization
* Limited real-time sales automation

#### 3. **Replicant**

* Voice AI assistant for **customer support and basic sales flows**
* Can handle **outbound and inbound**, but better suited to FAQs and support logic
* More service-focused than sales-conversion driven

#### 4. **PolyAI**

* High-quality voice UX, excellent **multilingual support**
* Used by global hospitality and logistics brands
* Strong in **booking and reservation flows**, not pure B2C sales

#### 5. **Tenyx**

* A newer voice AI platform with focus on **sales and service**
* Customizable but still maturing
* Interesting for pilot projects and custom integrations

***

### 🧪 How They Compare

| Feature                       | VoiceB.ai                | Invoca            | Replicant       | PolyAI          | Tenyx       |
| ----------------------------- | ------------------------ | ----------------- | --------------- | --------------- | ----------- |
| **Sales Focus**               | 🎯 Core Mission          | ⚠️ Limited        | ❌ Support-first | ⚠️ Partial      | ✅ Moderate  |
| **Inbound Sales**             | ✅ Yes                    | ❌ No              | ✅ Partial       | ⚠️ Booking-only | ✅ Yes       |
| **Outbound Sales**            | ✅ (non-cold)             | ❌ No              | ✅ Yes           | ❌ No            | ✅ Yes       |
| **Agent Replacement**         | ✅ 100% AI                | ❌ No              | ⚠️ Semi-auto    | ⚠️ Partial      | ✅ Yes       |
| **Real-Time Data Collection** | ✅ Yes                    | ⚠️ Analytics only | ❌ No            | ❌ No            | ⚠️ Partial  |
| **Click-to-Call/VoIP**        | ✅ Built-in               | ❌ No              | ❌ No            | ❌ No            | ⚠️ Optional |
| **Compliance Tools**          | ✅ Recordings, audit logs | ✅ Yes             | ✅ Yes           | ✅ Yes           | ⚠️ TBD      |

***

### 🧩 When to Choose VoiceB.ai

VoiceB.ai is the **only AI voice platform designed specifically to convert inbound sales**.

We’re not a support bot. We don’t stop at data analysis. We **sell** — autonomously.

Choose VoiceB.ai if:

* You want to **replace or augment your inbound sales team**
* You want AI that can **pitch, qualify, collect**, and **close**
* You need scalable solutions for **telco, banking, energy, insurance, or B2C SaaS**
* You value **compliance, auditability**, and clear ROI

***

### 🔚 Final Thoughts

As AI transforms voice interactions, only a few platforms are ready to step into the real sales funnel.

VoiceB.ai isn’t just another AI assistant — it’s a sales rep that works 24/7, never gets tired, and knows your product by heart.

📞 Ready to activate your first AI sales agent?\
[Book a demo today.](https://cal.com/team/voiceb-demo/voiceb-demo)


# Growth Is Changing. VoiceB.ai Is Already There

<sup>Alex Bisbe    7/21/2025     5min read</sup>

In a recent episode of *20VC*, Kieran Flanagan, CMO of HubSpot, laid out a bold vision: **AI is not just a tool—it’s reshaping how growth teams operate**. The traditional growth team is giving way to **AI-native pods**—agile, cross-functional units powered by automation, data, and AI-driven execution. \
\
Listen to the episode here: [**20Growth: The Death of Growth Teams?**](https://podcasts.apple.com/am/podcast/20growth-the-death-of-growth-teams-how-hubspot-use-ai/id958230465?i=1000716834879) [open.spotify.com+8podcasts.apple.com+8instagram.com+8](https://podcasts.apple.com/am/podcast/20growth-the-death-of-growth-teams-how-hubspot-use-ai/id958230465?i=1000716834879\&utm_source=chatgpt.com)

While many companies are still talking about adapting to this future, **VoiceB.ai has been living it for years**.

***

### 🎯 From Teams to Agents: The Future of Growth Is Automated

Kieran predicts that traditional growth orgs will disappear or be restructured around AI agents that learn, execute, and improve autonomously.

At VoiceB.ai, that’s not a vision—it’s our product.

**Our AI sales agents**:

* Answer inbound calls
* Pitch your product
* Handle objections in real time
* Collect data, qualify leads, and even close

They’re not just "assisting" humans—they *are* the reps. And they deliver.

***

### 🔁 The Power of Memory: Our Moat

One major challenge in AI today is starting from zero each interaction. VoiceB.ai’s advantage: **agent memory**.

Every conversation:

* Teaches the agent how users respond
* Improves objection handling
* Refines tone, timing, and call flow

This leads to **compound performance improvements**—the more you use it, the better it performs. It’s retention built into the product.

> “Your VoiceB agent is a digital asset. The longer you use it, the harder it is to walk away.”

***

### 📈 Specific Tasks. Real Results.

Kieran rightly signals that **broad generalist agents aren’t reliable yet**.

VoiceB.ai stays laser-focused:

* Operating on **inbound voice workflows**
* Executing your script, pitch, and qualification
* Delivering measurable, compliant results

Our agents are **narrow by design**—which is why they actually *work*.

***

### 💡 Prompt Engineering + Memory = The New Growth Stack

We’re entering an era where **prompting equals coding**, and **memory equals data strategy**.

VoiceB.ai brings both:

* A prompt-based structure that adapts to your voice, use case, and objection patterns
* A persistent learning layer that improves with every call

This results in **a self-improving, high-performance sales asset**—active 24/7.

***

### 🚀 Built for the AI-Native Growth Leader

If you're:

* Rethinking sales strategy
* Challenging CAC assumptions
* Looking to scale without scaling teams

VoiceB.ai is **your shortcut to the AI-native future**.

We’re not here to replace your CRM or human team. We’re here to **eliminate repetitive, high-volume inbound sales work**, freeing your people to focus on what truly matters.

***

#### 🎧 Tune In

Jump into Kieran Flanagan’s full perspective here:\
👉 [**20Growth: The Death of Growth Teams? (20VC)**](https://podcasts.apple.com/am/podcast/20growth-the-death-of-growth-teams-how-hubspot-use-ai/id958230465?i=1000716834879) [podwise.ai+4player.fm+4iheart.com+4](https://player.fm/series/the-twenty-minute-vc-20vc-venture-capital-startup-funding-the-pitch/ep-20growth-the-death-of-growth-teams-how-hubspot-use-ai-to-triple-email-conversion-the-future-of-ai-seo-why-prompt-engineering-is-the-new-coding-what-every-cmo-needs-to-know-about-ai-in-2025?utm_source=chatgpt.com)

***

#### Ready to elevate your sales with AI-driven reps?

[Book a Demo](https://voiceb.ai) or [Try it Live](https://voiceb.ai) and experience the future of growth—where *smarter beats bigger*.


# How SHV Energy, Transformed Inbound Sales with AI Voice Agents

**Primagas**, a leading energy company in Spain and part of the global **SHV Energy** group, has redefined what modern customer engagement looks like — not through more agents or larger budgets, but by embracing **AI-powered voice technology** that converts, qualifies, and supports at scale.

### About Primagas & SHV Energy

<figure><img src="/files/j43g5uuIr5DbhyhcdIAv" alt=""><figcaption></figcaption></figure>

Primagas has been operating in Spain since 1996 and is a key player in the **LPG and small-scale LNG** market. As a subsidiary of **SHV Energy** — a multinational with a presence in 21 countries across 4 continents and over 13,000 employees — Primagas shares a common mission:

> To provide **decentralized, cleaner, and lower-carbon energy** solutions, especially to rural and off-grid areas.

<figure><img src="/files/5fwsLWCjZrl07KWw7mlD" alt=""><figcaption></figcaption></figure>

With over **26 million customers worldwide**, SHV Energy and its brands (including Primagas, Calor, Liquigas, and Supergasbras) are spearheading the transition away from polluting fuels such as heating oil, coal, and biomass — towards cleaner alternatives like **LPG, LNG, and bioLPG**.

<figure><img src="/files/T6J9yeP9wMSiCYKkdQRV" alt=""><figcaption></figcaption></figure>

### The Challenge: Managing Inbound Interest at Scale

For a company with a growing number of **interested consumers searching for cleaner energy solutions**, every call matters. However, inbound channels — from web forms to phone calls — were hard to scale manually. The Primagas team was seeking a solution that could:

* **Respond instantly** to every lead
* **Qualify interest** before human intervention
* **Handle objections** in real time
* **Deliver structured summaries** to help reps close faster

### The Solution: AI Voice Agents by VoiceB

Primagas partnered with **VoiceB.ai**, a next-gen conversational AI platform designed to replace and augment human pre-sales agents.

> “**With VoiceB, we’ve transformed the way we handle inbound leads**. Every call — whether from our website or landing pages — is answered in under 4 seconds, 24/7. But it's not just about speed; the AI qualifies over 40% of calls for a second conversation, pitching our value and handling objections like a pro. Our team receives a crystal-clear summary, so the human expert can focus on closing with confidence."\
> — **Jordi Alonso, CMO at Primagas - SHV Energy**

The AI voice assistant integrates seamlessly into Primagas’ inbound strategy. It **engages prospects immediately**, asks the right qualifying questions, and routes only the most promising leads to the sales team — along with a full call transcript, lead profile, and summary of the conversation.

### A Focus on Customer Experience, Not Just Efficiency

While automation often aims to cut costs, Primagas saw AI voice agents as a **customer service enhancement**, not a budget-saving tactic.

> "We didn’t implement VoiceB to cut costs — **we did it to elevate our customer service**. Now, no lead goes unattended. The AI doesn’t just filter out unqualified calls; it engages, persuades, and prepares prospects before they even speak to a sales rep. It’s like having the best pre-sales agent on every call, around the clock."\
> — **Jordi Alonso, CMO at Primagas - SHV Energy**

### The Results

* **100% of inbound calls answered** — 24/7, in under 4 seconds
* **Over 40% lead qualification rate**
* **Higher conversion rates** by delivering well-prepared, high-intent leads to human sales reps
* **Operational scalability** without compromising the customer journey

### Conclusion

Primagas’ journey illustrates how **traditional energy companies can innovate**, not just in product or infrastructure, but in **customer engagement**. By blending global expertise in clean energy with cutting-edge AI voice technology, Primagas continues to lead — not just in sustainability, but in sales performance and service excellence.


# SuperVoice Product Video Demo

In this video, I demonstrate how VoiceB's AI-powered voice assistant can effectively convert website traffic into valuable leads. \
\
Many high-intent visitors leave without engaging, but our SuperVoice can answer questions in real-time, qualify leads, and guide users toward booking demos or accessing pricing information. For B2B companies, we've seen clients increase their conversion rates from 6% to 12% and significantly boost their return on sales. \
\
Setting up SuperVoice is simple—just add one line of JavaScript to your site, and it automatically updates with your content. If you're interested in seeing VoiceB in action, I encourage you to book a demo through our website.

{% embed url="<https://www.loom.com/share/edb0c341ab644db2bbad675d1140af2a>" %}

{% hint style="info" %}
**Book a demo** [**here**](https://cal.com/team/voiceb-demo/voiceb-demo)**.**
{% endhint %}


# SuperQualification Video Demo for Insurance companies

In this video, I demonstrate how to use VoiceB's AI agent to enhance your insurance experience.&#x20;

I walk through the process of obtaining a car insurance quote on the SecureLife website, highlighting key features like comprehensive coverage and flexible payment options.&#x20;

I also emphasize the benefits of combining car and health insurance for additional discounts. Please make sure to follow the steps I outline for a smooth application process.

{% embed url="<https://www.loom.com/share/dd0aa25578bb45d2a1116521a190ca54>" %}

{% hint style="info" %}
**Book a demo** [**here**](https://cal.com/team/voiceb-demo/voiceb-demo)**.**
{% endhint %}


# SuperQualification Video Demo for Online Banks

In this video, I demonstrate how banks can leverage VoiceB's AI Assistant to enhance website conversions.&#x20;

I walk through a simulated interaction where the AI Advisor assists a customer in choosing between a credit card and a personal loan for a renovation project.&#x20;

I highlight the benefits of Nova Bank's offerings compared to competitors, emphasizing transparency and ease of application.

{% embed url="<https://www.loom.com/share/e358eee6744543a8b9fd68b980626269?sid=732cff15-6925-4ec7-8840-4e4412bb12e5>" %}

{% hint style="info" %}
**Book a demo** [**here**](https://cal.com/team/voiceb-demo/voiceb-demo)**.**
{% endhint %}


# SuperQualification Video Demo for Telco companies

By Vanessa Goslow. February 19th, 2026.

One of the biggest structural problems in telecom inbound sales is **lead qualification**.

A significant percentage of calls are:

* Existing customers contacting support
* Non-qualified buyers
* Low-intent prospects
* Users comparing prices without purchase intent

The result is predictable:

* Call center costs increase
* Sales teams waste capacity
* Marketing acquisition cost rises
* Conversion volatility grows

Both Sales and Marketing suffer.

To solve this, we built a **Telco Autonomous Sales Agent** — designed specifically for telecom acquisition funnels.

Its objective is simple:\
**Close the sale autonomously whenever possible, and introduce humans only in edge cases.**

The agent executes sales in three ways:

#### 1️⃣ Digital Assisted Checkout (Web-Guided Flow)

As shown in this demo, the AI guides the customer step-by-step through the online journey, ensuring:

* Correct plan selection
* Structured data completion
* Frictionless checkout
* No data loss

The agent behaves like a real-time sales assistant embedded in the web flow.\
\
This demo shows how model 1️⃣ works in practice.

{% embed url="<https://www.loom.com/share/659df31a983040adadc28ed349c1c3b6>" %}

#### 2️⃣ Full Phone-Based Data Collection

If the customer is not in front of the website, the AI can execute the entire sales flow over the phone:

* Plan explanation
* Objection handling
* Structured data collection
* Pre-sale or sale confirmation

No dependency on the web journey.

#### 3️⃣ Qualified Human Transfer (Edge Cases Only)

If the user requests human assistance or falls outside standard flows, the agent:

* Qualifies the intent
* Collects essential data
* Transfers only sales-ready calls

Humans intervene only when needed.

***

This is not about adding another automation tool.

It’s about restructuring inbound sales economics:

* Reduce non-commercial call load
* Protect agent bandwidth
* Lower acquisition cost
* Increase autonomous revenue capture

By [Vanessa Goslow](https://www.linkedin.com/in/vanessagoslow/). GTM<br>

{% hint style="info" %}
**Book a demo** [**here**](https://cal.com/team/voiceb-demo/voiceb-demo)**.**
{% endhint %}


# SuperVoice+ Video Demo for Utilities companies

Boosting Conversions for Utilities companies with VoiceB's AI Agent

In this video, we demonstrate how utility companies are leveraging VoiceB's AI agent to enhance website conversions.&#x20;

You'll see how the AI guides me through selecting an energy plan, addressing objections, and ultimately leading to a higher conversion rate. I encourage you to consider how this technology could be applied in our projects. Let me know your thoughts!

{% embed url="<https://www.loom.com/share/7cff80a133374cfb93caf405a81f6957>" %}

{% hint style="info" %}
**Book a demo** [**here**](https://cal.com/team/voiceb-demo/voiceb-demo)**.**
{% endhint %}


# Introducing Balanced Voice Routing: Reliability, Speed, and Scale for Voice AI

How VoiceB.ai is delivering enterprise-grade uptime and performance with dual voice providers

<sup>Alex Bisbe    6/30/2025     5min read</sup>

{% hint style="warning" %}
**Good to know:** In the world of conversational AI, **latency, reliability, and uptime** are everything. One second too late, and your customer is gone. A missed call, and the opportunity vanishes. That’s why we’re excited to announce our latest infrastructure upgrade:
{% endhint %}

#### **Balanced Voice Routing**.

With Balanced Voice Routing, VoiceB.ai now routes calls between **ElevenLabs** and **Bland.ai** — two of the most advanced AI voice providers on the market — giving your AI agents best-in-class performance and reliability, automatically.

<figure><img src="/files/bOLusojCU0GUtTkr35wV" alt=""><figcaption></figcaption></figure>

### 🔁 What Is Balanced Voice Routing?

Balanced Voice Routing is a smart layer of infrastructure that dynamically:

* **Distributes live calls** between ElevenLabs and Bland.ai
* **Optimizes for lowest latency** and best audio quality
* **Provides automatic failover** if one provider is down or degraded

It’s like having two highways instead of one — if there’s traffic or a roadblock, your call is instantly routed to the better path.

### ⚙️ Why It Matters

Voice AI needs to *sound great* and *respond instantly* — every time. But cloud infrastructure can fail, APIs throttle, and outages happen. With Balanced Voice Routing, VoiceB ensures your AI agent:

* Stays online 24/7
* Speaks with minimal latency
* Delivers smooth, natural conversations
* Avoids vendor lock-in

### ✅ Advantages of Using Both ElevenLabs & Bland.ai

By balancing between **two best-in-class voice engines**, you get:

#### 🟣 **Better Latency**

ElevenLabs often performs faster in Europe, Bland.ai may be better in the US. We test and route calls where they’ll perform best.

#### 🟣 **Higher SLA**

With automatic provider fallback, your AI agent can keep talking — even if one provider is temporarily unavailable.

#### 🟣 **Resilience by Design**

No single point of failure. You get true fault tolerance and business continuity.

#### 🟣 **Voice Diversity**

You can choose or alternate voices between providers to create a more human-like and branded voice experience.

### <br>

### 🚀 Built for Enterprise Scale

This feature is especially critical for **large enterprises** and **high-volume use cases** like:

* Banking and insurance lead qualification
* Telco subscription flows
* Retail and eCommerce phone assistance
* B2B appointment setting

When uptime and quality are non-negotiable, Balanced Voice Routing ensures your AI agent meets the moment — every time.

### <br>

### 🔜 And We're Not Stopping Here

We’re working on integrating even more voice providers to give you:

* Greater regional coverage
* More languages and accents
* More voice options and personalities

Our goal is simple: **the best voice experience on the market**, powered by multiple engines, managed with zero effort on your side.

### <br>

### Ready to Try It?

Balanced Voice Routing is already live for all enterprise accounts. If you want to experience AI voice that’s **faster, smarter, and always online**, get in touch.

<br>

👉 [Talk to Sales](https://cal.com/team/voiceb-demo/voiceb-demo) and request a live demo with Balanced Routing in action.

<br>


# Voice AI Best Practices for Customer Experience

Essential strategies to boost CX and streamline support with conversational AI.

<sup>Alex Bisbe   6/30/2025  5 min read</sup>

{% hint style="info" %}
**Good to know:** voice AI isn’t one-size-fits-all. Depending on your audience and product complexity, your assistant may need a different tone, depth, or interaction pattern. Don’t be afraid to iterate and test! Sometimes a small change in tone or phrasing can double your completion rate
{% endhint %}

<figure><img src="/files/shkeqMxCykfpsmnmpTHY" alt=""><figcaption></figcaption></figure>

Voice AI is no longer a futuristic concept — it’s a competitive advantage. Whether you're in banking, insurance, telco, or SaaS, implementing **voice AI** the right way can drastically improve customer experience, reduce support load, and increase satisfaction. But success isn’t just about having the tech — it’s about how you use it.

Here are the **essential best practices** for implementing voice AI that truly enhances the customer journey.

### 🎯 1. Start with a Clear Use Case

Before jumping in, define a specific, high-impact use case. Great examples include:

* Resolving level 1 support queries (FAQs, order tracking)
* Guiding users through online applications
* Qualifying inbound leads
* Handling appointment scheduling

**Why it matters:** Focus drives clarity. By solving one pain point first, you ensure fast results and build internal momentum.

### 🧠 2. Train Your AI with Real Data

Don’t start from scratch. Train your voice AI using:

* Call transcripts
* Chat history
* Product documentation
* Help center content

With platforms like **VoiceB.ai**, you can even **automatically scrape your website daily** to keep the AI updated with your latest content — no manual updates needed.

### 🗣️ 3. Sound Natural — Not Robotic

A robotic voice destroys trust. Choose a voice AI provider that offers:

* Natural-sounding speech
* Real-time latency under 500ms
* Dynamic tone adjustment

**Bonus tip:** For premium CX, use multiple voice vendors in parallel (e.g., ElevenLabs + Bland.ai) to balance performance and ensure uptime.

### 🔁 4. Personalize the Experience

Great voice AI should:

* Address the customer by name
* Remember previous interactions (where permitted)
* Adapt its flow based on user responses

**Pro tip:** Avoid generic “How can I help you?” intros. Instead, guide the user: “Are you calling to check your order status or speak with support?”

### 📊 5. Monitor and Improve Continuously

Use dashboards and transcripts to:

* Review failed or incomplete calls
* Detect friction points
* A/B test scripts or call flows

Platforms like VoiceB.ai provide full **real-time transcription**, **call recordings**, and **QA monitoring tools** to iterate quickly.

### 🔐 6. Stay Compliant

For voice AI in regulated industries (like finance or healthcare), make sure:

* Calls are recorded and auditable
* The AI discloses it's not human
* Sensitive data is never stored

VoiceB.ai makes this easy with built-in disclaimers, data handling rules, and customizable prompts.

### 🚀 7. Make It Easy to Escalate

Not all calls should be handled by AI. Build a fallback:

* Route complex cases to human agents
* Allow users to request a callback
* Provide a way to speak to a person at any point

This isn’t just a failsafe — it’s a trust signal.

### Final Thoughts: Voice AI That Actually Works

Done right, voice AI doesn’t just replace agents — it **empowers customers**, **scales your team**, and **delivers better CX**.

The best voice AI platforms (like VoiceB.ai) combine:

* Plug & play setup
* Automatic updates via web scraping
* Real-time scoring and analytics
* Support for both inbound and outbound calls

It’s not about replacing people — it’s about giving your business **a voice that works smarter**.

👉 Ready to deliver voice experiences your customers will love? [Talk to Sales](https://voiceb.ai/) and get started today.


# How SuperVoice Converts Website Visitors to Customers

Why 33% more sales and 70% fewer support tickets is just the beginning.

<sup>Alex Bisbe   6/30/2025  5 min read</sup>

When someone lands on your website, you have only a few seconds to engage them. Most companies try with chatbots, forms, or delayed callbacks — and lose the lead. At **VoiceB.ai**, we believe there's a better way: **SuperVoice**.

SuperVoice is our AI-powered voice assistant that lives on your website, talks to visitors in real time, answers complex questions, and drives them to convert — all without a single human agent involved.

\
![](/files/7Wn52a9jEIh5UESuF7pV)

### Real Results: More Sales, Fewer Tickets

One of our SaaS clients put it best:

> “Since switching to VoiceB's SuperVoice, sales are up 33% and 70% of support requests are resolved automatically. For complex topics like equity management, voice is the most agile channel — and VoiceB's AI handles it brilliantly. The web scraping delivers incredible answer quality. With a small team, we can now compete with the big players.”

Voice isn't just faster — it's smarter, more natural, and better suited to real decision-making conversations than chat or email. And with VoiceB’s automation, small teams can operate like enterprise sales orgs.

### Why VoiceB is Different

Most voice AI platforms require weeks of training, flow design, integrations, and ongoing maintenance. Not VoiceB.

Here’s how we make it effortless:

#### ✅ **No-Code, Plug & Play Setup**

Just paste **one line of JavaScript** into your website and go live instantly. No dev team needed. No training sessions. No onboarding friction.

#### 🔁 **Self-Updating Agent Script**

No need to write or manage scripts manually. Our platform builds your AI agent’s behavior dynamically, using your live website and help center content. As your site updates, your AI gets smarter automatically.

#### 🔍 **Daily Web Scraping**

VoiceB scrapes your website and knowledge base **every day** to stay up-to-date. That means your AI agent always knows:

* Your latest offers and pricing
* New product features
* Updated processes and policies

### What Can SuperVoice Do?

* **Answer pre-sales and support questions** in real time
* **Handle objections** and pitch your value proposition
* **Drive conversions** by booking meetings or pushing to checkout
* **Score leads** and pass them to your CRM
* **Free your human team** from repetitive conversations

### Built for Scale, Built for Results

Whether you’re a 5-person startup or a 500-agent operation, SuperVoice is designed to scale with you. No matter how much traffic you get or how complex your product is, our AI adapts, updates, and improves every day — without any work on your end.

### Ready to Turn Visitors into Revenue?

Voice is the highest-converting channel. SuperVoice makes it instant, intelligent, and automated. If you're ready to boost conversions without hiring more people, it's time to add voice to your stack.<br>

👉 [Talk to Sales](https://voiceb.ai/) and launch SuperVoice on your site today.


# Implementing SuperQualification for B2C & B2B Lead Scoring

A comprehensive guide to using SuperQualification for automated B2C & B2B lead qualification and scoring, we walk you through how to implement SuperQualification and dramatically increase your pipelin

<sup>Alex Bisbe   6/30/2025  5 min read</sup>

In B2C & B2B sales, not all leads are created equal — and wasting time on unqualified prospects is a luxury no company can afford. That’s where **SuperQualification** from VoiceB.ai comes in: a fully automated, AI-powered voice solution designed to qualify inbound leads and score them in real time.

In this guide, we walk you through how to implement SuperQualification and dramatically increase your pipeline efficiency.

<figure><img src="/files/MBGU8V4Nt03iJNboC0nc" alt=""><figcaption></figcaption></figure>

### 🎯 What Is SuperQualification?

SuperQualification is an enterprise-grade solution by VoiceB.ai that automates the process of qualifying inbound leads — via phone calls or web-based callbacks. Instead of routing all leads to your human sales team, the AI agent engages with them first, asking relevant questions, handling objections, and scoring them instantly based on your own qualification rules.

{% hint style="info" %}
**Book** **a demo** [**here**](https://cal.com/team/voiceb-demo/voiceb-demo)**.**
{% endhint %}

### ✅ Benefits for B2C & B2B Companies

1. **Automated 24/7 lead qualification**
2. **Real-time lead scoring**
3. **More efficient SDR/BDR workflows**
4. **Higher conversion rates**
5. **Reduced manual workload**
6. **Seamless CRM integration**

### 🚀 How to Implement SuperQualification in 5 Steps

#### **Step 1: Define Your Qualification Criteria**

Before launching, identify the key attributes that define a qualified lead in your funnel. These typically include:

* Company size or industry
* Budget range
* Urgency/timeline
* Decision-making authority
* Use case or intent

These rules will power the scoring model inside VoiceB.ai.

#### **Step 2: Upload Your Scripts and FAQ**

VoiceB allows you to upload qualification scripts, FAQs, and objection handling templates. You can also **scrape your website** to enrich the AI’s knowledge base — allowing it to answer common product and service questions while collecting critical lead data.

#### **Step 3: Train and Test the AI**

Train your AI agent with mock calls and real-world scenarios. You can:

* Simulate conversations across use cases
* Fine-tune language, tone, and handling logic
* Test scoring outcomes before going live

VoiceB’s console lets you monitor, tweak, and relaunch instantly.

#### **Step 4: Connect Lead Sources**

Use VoiceB’s API or VoIP integration to connect SuperQualification to:

* Web forms or demo request buttons
* Inbound call routing
* SDR email follow-ups or chat bots

As soon as a lead comes in, VoiceB’s AI kicks in to qualify and score it on the spot.

#### **Step 5: Route Qualified Leads to Sales**

Once a lead is scored as qualified:

* The AI pushes the data to your CRM
* Optionally, it books a meeting with your AE via calendar sync
* Low-scoring leads are tagged or nurtured via automation

Your sales team only spends time on high-value, ready-to-convert leads.

### 🔗 CRM & Workflow Integrations

SuperQualification integrates with:

* Salesforce
* HubSpot
* Pipedrive
* Calendly
* Webhooks & Zapier
* Custom APIs

It fits right into your existing stack — no disruption, just results.

### 🧠 Real-Time Scoring, Real-World Impact

Unlike traditional lead forms or manual qualification, VoiceB’s SuperQualification:

* Speaks directly to the lead within seconds
* Dynamically adjusts based on conversation flow
* Scores leads using your exact ruleset

It’s like having an SDR team that works 24/7, never gets tired, and never misses a follow-up.

### Ready to Deploy SuperQualification?

If you’re tired of wasting sales hours on unqualified leads and want to accelerate your B2B pipeline, it’s time to switch to VoiceB.ai’s SuperQualification.<br>

👉 [Talk to Sales](https://voiceb.ai/) to set up your campaign or see a demo.


# Terms & Conditions

### **General term of use**

1 ) This document shows the General Terms of Use of the Website VoiceB.ai (the “Site”) property of Sekondary S.L. (“VoiceB“), located at Avda de Can Picanyol, nº 30, PB, 1<sup>st</sup> door,  08173, Sant Cugat del Vallès,  Barcelona, Spain, Registration Code Number B72794613  and entered at the Commercial Registry of Barcelona, Tome 48567, Folio 71, Page B 588425, registration 1.

&#x20;

Certain services of the Site may be subject, together with the General Terms of Use, to its own terms of use (from now on Particular Terms of Use).

VoiceB may modify the General Terms of Use and all the Particular Terms of Use (called together \`Terms of Use), at any time without prior notice, in order to adapt them to the current law applicable at each time, jurisprudential changes and usual market practice. Access and use of the Site by the User implies their express acceptance of all the Terms of Use on the published version on the Site when accessing it.

2 ) Access and Use of the Site

1. User Status

Access to the Site attributes the status of User. To access and use certain services, it may be necessary to be a registered user. This registration must be carried out only in the specified way. The User understands and accepts that the use of the contents and/or services offered on this Site shall take place at his own risk and/or liability.

2. Proper Use of the Site

The User undertakes to use the Site and all its contents and services in accordance with law, morality, civil order, and Terms of Use. Similarly, the User undertakes to make proper use of the services and/or contents of the Site, which shall not be used under any circumstances for illegal purposes or those which may constitute a crime, threaten third parties’ rights, and/or constitute an infringement of intellectual and industrial property rights protected by law or any other regulations of the applicable legal code.

&#x20;

3\.    User Obligations

The User undertakes not to transmit, insert, distribute, or put at third parties’ disposal any kind of material and information (data, messages, pictures, sound and image files, photographs, software, etc.) contrary to law, morality, civil order, and Terms of Use.

&#x20;

The User agrees NOT to:

* insert or distribute any information, false, ambiguous, or inaccurate contents that lead the information recipients into error.
* Impersonate other Users of the Site.
* Carry out any action, nor distribute, transmit, or put at third parties’ disposal any kind of information, item, or contents that constitute an infringement of the intellectual and industrial rights protected by law, patents, marks, or copyright property of the owner of the Site or of third parties..
* Carry out any action or omission that implies or may imply an infringement of telecommunications secrecy and Personal Data Protection Law.
* Carry out any action or omission that promotes, incites, and/or constitutes xenophobic conduct, an apology for terrorism, or that threatens human rights..
* insert or distribute into the Internet any data program (virus or harmful software) liable to cause damage to the computer systems of the access provider, its providers, or third-party users of the Internet.
* Transmit advertising that has not been requested or authorized, advertising material, spam, chain letters, pyramid schemes, or any other way of request, except in those areas (such as advertising spaces) that have been exclusively designed for that purpose..
* Distribute information or content that threatens fundamental rights and public freedoms recognized by the Constitution and international treaties.
* Share, transmit, or make available to third parties any information that constitutes illegal or unfair advertising

&#x20;

3\)  The User understands that all the contents of the Site, information, and material contained on it, structure, selection, arrangement, and display of its contents and computer programs, software, databases relating to it are protected under VoiceB or third parties´ intellectual and industrial rights. As a result, the User may not reproduce, transform, modify, disassemble, carry out reverse engineering, distribute, hire, render, put at disposal or permit public access by means of a public communication method to any item and/or contents of the Site. The User shall refrain from suppressing identification symbols of the intellectual and industrial rights or any other belonging to VoiceB or third parties are shown on the Site, and from the different services offered through it. Similarly, the User shall refrain from evading or manipulating any technical device provided by VoiceB or third parties for the protection of their rights.

&#x20;

4\)  VoiceB does not guarantee availability, access, and continuity of the operation of the Site. In accordance with the limits established in the current legal code, VoiceB shall not be liable for the damages caused to the User as a result of the lack of availability, access errors and lack of continuity of the Site.

&#x20;

5\) Contents and services provided by third parties may be accessed through the Site. VoiceB shall only be liable for the Services provided by itself and the contents directly created by VoiceB. The aforementioned liability will be excluded in the event of unavoidable circumstances or if the User´s system configuration is not appropriate to permit the right use of the Site services.

&#x20;

6\) Through the Site, the User shall access, use and contract services and/or contents liability and property of third parties. VoiceB neither previously controls, approves nor is liable for the content and services contracted and/or provided by these third parties through the Site. In the same way, VoiceB does not guarantee legality, reliability, usefulness, veracity, accuracy, exhaustiveness and update to the content and services provided by third parties through the Site. The User understands and accepts that contracting processes of services initiated through the Site and any contractual or extra-contractual relationship formalized with the third parties contacted through the Site are considered to have been solely carried out between the User and the third party. Therefore the User accepts that VoiceB is not liable for damage of any nature caused on the occasion of their contracting process, business, talks and/or contractual or extra-contractual relationships with the third parties contacted through the Site, and whose products are the subject of marketing through the Site. With the indicated restrictions by applicable current rules as regards Information Society Services, VoiceB shall not be liable, neither indirectly nor subsidiarily, for damages of any nature arising from the use and contracting of the contents and services provided by third parties through the Site, as well as the lack of legality, reliability, usefulness, veracity, accuracy, exhaustiveness and update to them. For your information, under no circumstances limiting, VoiceB shall not be liable for the damages of any nature arisen from: a) defective compliance and / or failure to comply with the contractual and / or extra-contractual agreements assumed by third parties; b) acts of illegal advertising or unfair competition; c) services and contents of third parties which are inappropriate and fail to meet expectations; d) all kind of vices and defects of the Services and contents provided through the Site by third parties; e) the use and processing carried out by third parties relating to the personal data provided through the Site by the User.

&#x20;

7\) The Site may link technical devices and search tools at the User´s disposal that permit access to websites property of other entities (links). The User understands and accepts that the use and access to the linked websites shall take place at their own risk and responsibility and VoiceB is exonerated from any liability relating to technical availability of the linked websites, quality, reliability, accuracy and/or veracity of the services, information, items and/or contents the User may access on them and in the search directories included on the Site. With the indicated restrictions by applicable current rules as regards Information Society Services, VoiceB shall be liable, neither indirectly nor subsidiarily, for the damages of any nature arising from: a) operation, lack of availability, inaccessibility and lack of continuity of the linked website and / or the available search directories; b) the lack of maintenance and updates to the content and services contained on the linked websites; c) the lack of quality, accuracy, legality and usefulness of the content and services of the linked websites.

&#x20;

8\) Use of Personal Data

A – Personal data for third parties.

• When the User accesses content or initiates the contracting process of a Service and/or product advertised and/or marketed through the Site by a third party, the User shall provide their personal data for the abovementioned entity which is liable and owner of the service and contents accessed to or whose contracting has been initiated.

• The company which owns the Service and /or content shall be in charge of the personal data processing. In each of the registration forms that must be formalized by the User to access the various services and / or products provided by the third party, the entity liable for the processing shall be identified, as well as the various points established under the current law applicable to personal data provided for VoiceB as owner and manager of the Site.

B – Personal data for VoiceB

• As regards the personal data of the User obtained by VoiceB as owner of the Site and as a result of the access and use of it by the User (e.g. IP addresses, e-mail addresses sent to VoiceB, etc.), as well as other data voluntarily provided by you for VoiceB as manager of the Site, please read the VoiceB Privacy Policy.

&#x20;

9\) Use of Cookies. Please note that the website uses both our cookies and third-party cookies in order to improve our services and to display advertisements related to your preferences by analyzing your browsing habits.

Cookies are tools used by web servers to store and retrieve information on their visitors. They are small text files which we send to your computer in order to maintain a record of your preferences and remember them upon your return.

These cookies can store information in order to show contents and offers of special interest in accordance with your profile and to register the number of times a message is shown to a User (general novelties in the website).

Users may freely decide about whether cookies used are implemented in their hard drives or to receive a warning on their screens when a cookie is received and to decide whether or not to implement it in the hard drive. In order to do so, we suggest using the help section in your web browser to find out how to change the configuration currently being used.

For more information on the use of Cookies please visit our Use of Cookies Policy.

&#x20;

10\)  VoiceB reserves the right to: a) Carry out the changes considered suitable, and may modify, delete and unilaterally include without prior notice new contents and/or Services, as well as the way in which those shall be displayed and located. b) Deny or withdraw access to the Site, without needing prior notice, at its own request or at third parties´ request, to those Users that may be breaking these Terms of Use.

&#x20;

11\. All the questions relating to the Site are subject to Spanish Law and submitted to the jurisdiction of the Courts and Tribunals of Spain.

&#x20;

Last Updated: 24th, March 2025

&#x20;


# Privacy Policy

Updated on October 26th, 2025

<p align="center"></p>

<p align="center"></p>

The company SEKONDARY S.L. DBA VOICEB ("We (Us)" “VoiceB” or “Sekondary"), is committed to protecting your privacy. Through this Privacy Policy we inform you on how VoiceB process the personal data you provide us by any means and / or channel made available to you, including this Website.

&#x20;

Who is the data controller?

&#x20;

SEKONDARY, S.L., is the data controller for the processing of Personal Data, which implies that we are responsible for protecting the privacy of your Personal Data. We will use your personal data in a fair, legal and transparent manner and, in accordance with General Data Protection Regulation (EU) 2016/679 of April 27, 2016 ("GDPR") and the current regulations applicable to the protection of personal data.

&#x20;

Please find below our contact information:

&#x20;

Address: Avda de Can Picanyol, nº 30, PB, 1st door,  08173, Sant Cugat del Vallès,  Barcelona, Spain.

&#x20;

Contact of Data Controller: Legal Department, email: <hello@voiceb.ai>

&#x20;

&#x20;

For any queries related to the processing of Personal Data, you can contact the Data Controller at the email address above.

&#x20;

Why do we process your data?

&#x20;

In compliance with the provisions of the GDPR and the current regulations applicable to the protection of personal data, we inform you that your personal data shall be processed by VoiceB for the purposes of:

&#x20;

(i)              allowing access to and use of the website and, as well as the management, development and control of your registration and use of the website

(ii)             managing your registration as a client of VoiceB 's services and the requests, of any nature, that you send us.

(iii)            administering and developing the pre-contractual or contractual relationship formalized with you

(iv)           improving our services and products and send service updates and keep you informed of all our latest developments, offers and products which are like those you contracted and, that may be of interest to you and.

(v)            Delivering commercial communications, by any means regarding products information and / or services provided by VoiceB and of any third-party products and services when the user of the website has expressly accepted the privacy policy when provide the personal data that will be used for the sending of this information.

&#x20;

What is our legitimacy to process your Personal Data?

&#x20;

The legal basis that justifies the above processing referred to in sections (i) (ii) (iii) and (iv) is the performance of the existing pre-contractual and contractual relationship and the legitimate interest of VoiceB to improve its relationship with its clients and promote your products and services. The legal basis that justifies the processing indicated in section (v) above is your consent.

&#x20;

In relation to the processing referred to in section (i), please read VoiceB's Cookies Policy (link).

&#x20;

International transfer of Personal Data

&#x20;

Your data is stored and processed within the EU / EEA or in countries that are considered to offer an adequate level of protection by the European Commission, or within the U.S  processed by Data Processors who have subscribed to agreements that fully comply with the legality of transfers from other countries  or provide adequate safeguards (Standard Contractual Clauses)

&#x20;

&#x20;

How long do we retain your Personal Data?

&#x20;

Your personal data will be retained for the duration of the relationship between VoiceB and you and / or use of this website and, once this relationship has been finalised, data will be kept for as long as liabilities may arise or for the period required by applicable law.

&#x20;

Personal data will be securely deleted or anonymized when no longer necessary for the purposes for which it was collected, in accordance with VoiceB’s internal **data retention and deletion procedures**. Marketing or informational communications will continue until you inform us otherwise.

&#x20;

&#x20;

Your rights and Data Deletion Requests

&#x20;

We inform you that you can exercise your rights of access, rectification, cancellation, portability, limitation and / or opposition to the processing of your personal data in accordance with applicable data protection laws.

&#x20;

These rights may be exercised free of charge by sending a communication to the following email address:  <hello@voiceb.ai>.

&#x20;

To submit a data deletion request, please contact us by email at <hello@voiceb.ai>  with the subject line “Data Deletion Request.”

&#x20;

Upon receiving your request, VoiceB will:

* Confirm receipt of your request within the legally required timeframe;
* Verify your identity to protect your personal data; and
* Delete your personal data from our systems and records, unless retention is required for compliance with legal, contractual, or security obligations.

&#x20;

The categories of personal data that may be deleted include:

&#x20;

* Account and registration details
* Contact information (such as name, email, or phone number)
* Usage and communication data associated with your account
* Support and inquiry records

&#x20;

Please note that VoiceB may retain certain data for a limited period as required by law (for example, for tax, billing, or audit purposes)

&#x20;

VoiceB maintains an internal process and record of all data deletion requests to ensure compliance with the **GDPR**,

&#x20;

In any case, if you consider your data protection rights are breached and / or that Voice B does not comply with its privacy obligations, you may file a complaint with the competent data protection authority (agpd.es).

&#x20;

Contact VOICEB

&#x20;

You may withdraw the consent referenced to in this Privacy Policy at any time by sending an e-mail to the email address on <hello@voiceb.ai>

&#x20;

&#x20;

For any queries, or any question related to the processing of Personal Data, you may contact the Data Controller at the following email address: <hello@voiceb.ai>

&#x20;

If you believe your data protection rights have been violated or that VoiceB has not fulfilled its privacy obligations, you may file a complaint with the **Spanish Data Protection Agency (AEPD)** at [www.agpd.es](http://www.agpd.es)

&#x20;

&#x20;

&#x20;


# Data Processing Agreement

Last updated: January 2026

This Data Processing Agreement (“DPA”) forms part of and is incorporated by reference into the Master Subscription Agreement (“Agreement”) between the Client and Sekondary S.L., doing business as VoiceB (“VoiceB” or the “Data Processor”).

The purpose of this DPA is to establish the terms and conditions under which VoiceB will process personal data on behalf of the Client in connection with the provision of the Service, in accordance with Article 28 of the GDPR and other applicable data protection laws.

By subscribing to and/or using the Service, the Client (the “Data Controller”):

* Acknowledges that this DPA applies automatically without the need for a physical signature.
* Confirms that it has read and understood this DPA.
* Agrees that VoiceB will process personal data as described herein, exclusively for the provision of the Service.

This DPA shall remain in force for the duration of the Agreement and is an integral part thereof.

&#x20;

<p align="center"><strong>STIPULATIONS</strong></p>

**1. Definitions**

**"Data Protection Legislation"** shall be interpreted as the Regulation, Directive 2002/58/EC, as well as any other rules and / or regulation implemented or created with the aforementioned Directives, or that modifies, replaces, recreates or consolidates any of them, as well as any other laws applicable to the processing of personal data that may exist in any jurisdiction, including, where appropriate, the guides and codes of practice published by the control authorities.

**"Data controller",** "Data processor", "data subjects", "processing" and "appropriate technical and organizational measures" shall be interpreted in accordance with the Data Protection Legislation applicable in the relevant jurisdiction.

**"Personal data breach"** means any breach of personal data that causes the destruction, loss or accidental or unlawful alteration of personal data to which the data subject have had access and / or have been treated by the data processor during the provision of the Service.

Any other terms defined and used in this DPA shall be interpreted in accordance with the provisions of the same.

&#x20;

**2. Purpose**

The purpose of this DPA is to:

2.1 Define the terms and conditions according to which the Data Processor will carry out the processing of the personal to which it has access in the framework of the Service.

2.2 Establish the applicable confidentiality conditions that the Data Processor, and all its personnel, collaborators and / or agents must comply with at all times during the provision of the Service and once it has finished.

&#x20;

**3. Obligations of the Data Controller.**&#x20;

3.1 The Data Controller recognises and accepts that knows and complies with the Data Protection Legislation and assumes all the responsibility in relation to the fulfilment of the obligations that it holds as Data Controller. The Data Controller shall defend and hold the Data Processor harmless against all claims, losses, damages, costs and expenses that may arise or relate, directly and indirectly with claims, procedure and / or fines of any nature with respect to the breach of the Data Controller of the guarantees and obligations established in this section.

3.2 Specifically, as Data Controller, it is obliged, and will be solely responsible for the following obligations and guarantees: (i) That the personal data is collected and processed in accordance with and in compliance with the requirements of the Data Protection Legislation; (ii) it complies with all obligations, including those related to obtaining the consent and information that are necessary in accordance with the Data Protection Legislation for the processing of personal data that is made as a consequence and / or in relation to the Services Contract; (iii) to implement and update the notices necessary for personal data protection and the privacy policies on the Websites that are owned by them. In the event that the Service is provided on the website held by the Data Processor , the Data Controller shall provide the Data Processor , prior to the start of the provision of the Service, the notifications of data protection and privacy regulations that must be inserted by the Data Processor  on the aforementioned website and / or the support used to collect the personal data.

3.3 The delivery of the legal notifications mentioned in the previous section 3.2 on behalf of the Data Controller to the Data Processor is a necessary requirement for the commencement of the Services.

3.4 The Data Controller states that the processing of personal data by the Data Processor does not entail a high risk for the rights and freedoms of natural persons.

&#x20;

**4. Purpose and object of the processing of personal data. Obligations.**

4.1 The Parties acknowledge that the Client is the Data Controller and VoiceB., is the Data Processor in relation to the personal data that will be processed during the provision of the Service.

4.2 The object of the processing is the provision of the Service and this DPA will have the same duration as the Service Contract of which it is an inseparable part. Appendix A of the DPA establishes the nature and purpose of the processing, the type of personal data processed by the Data Processor , as well as the categories of the data subjects whose personal data will be processed in the framework of the provision of the Service.

4.3 The processing of personal data that the Data Processor fulfils will be limited to the necessary processing's to render the Service to the Data Controller and, in any case, it will be subject to the documented instructions of the Data Controller.

4.4 The Data Processor agrees not to carry out any other processing on the personal data or to apply or use the personal data for any purpose other than the provision of the Service.

4.5 If the Data Processor is obliged to process the data for any other purpose under Union or Member State law, it shall inform the Data Controller of this legal requirement prior to the processing, unless such Law prohibits it for reasons important to the public interest.

&#x20;

**5. Security of personal data**

5.1 The Data Processor declares that it is obliged to implement and maintain the appropriate technical and organizational measures to protect personal data against unauthorized or illegal processing and against accidental or unlawful loss, destruction, alteration, damage or accidental or unlawful theft. These measures shall be appropriate considering: (i) the risks presented by data processing, in particular as a consequence of the destruction, loss or accidental or unlawful alteration of transmitted personal data, and (ii) the nature of the data of personal nature object of protection. At a minimum, these measures shall comply with the requirements established under Appendix B of this DPA.

&#x20;

**6. Prohibition of communication of personal data**

6.1 Except for the cases of subcontracting existing in accordance with the below Seventh stipulation, the Data Processor  undertakes to keep under his / her control and custody all the data of the Data Controller that it accesses due to the provision of the Service and not to disclose them, transfer them, or in any way communicate them, not even for their conservation, to other people.

&#x20;

**7. Subprocessors**

7.1 The Data Processor is authorized to engage sub processors if it is necessary for the fulfilment of the Service. In the event that the services subject to subcontracting imply the access and processing by the sub processor of the personal data of the Data Controller, the Data Processor will subscribe, prior to the commencement of the subcontracting, a Data Processing Agreement with the Sub processor in no less demanding terms than those contained in this DPA..

7.2 The Data Controller acknowledges and accepts that, as of the effective date of this DPA, the Data Processor uses the sub processors listed in Appendix A.

7.3 The Data Processor  may update the list of sub processor from time to time. Any such updates will be communicated to the Data Controller through the Website or by email. The Data Controller shall have a period of 15 calendar days from the notification to raise any legitimate and duly reasoned objection.

7.4 For all purposes, the sub processor will have the status of data processor. In the event of non-compliance and / or defective compliance by the subcontractor of the formalized processing assignment or the Data Protection Legislation, the Data Processor  will remain fully responsible to the Data Controller regarding the compliance of the obligations of the sub processor.

7.5 In the event that (i) the European Commission establishes or any supervisory authority adopts standard contractual clauses for the matters referred to in Articles 28.3 and 28.4 of the Regulations, in accordance with the provisions of Articles 28.7 and 28.8 of the Regulations (as applicable), and (ii) the Data Controller notifies the Data Processor  of its desire to incorporate elements of such standard contractual clauses to this DPA, the Data Processor  will accept the changes required by Data Controller to achieve this objective.

&#x20;

**8. Obligation to return the data**

8.1 Upon termination of the Agreement, VoiceB will delete or return Personal Data in accordance with its standard retention policies, unless otherwise required by law.

8.2 Once the obligations referred to in the previous paragraph have been fulfilled, the Data Processor , through an authorized person, must if required by Data Controller,  issue a formal communication stating that it has been carried out faithfully.

8.3 In the event that the Data Processor, its staff, collaborators and / or agents are legally bound to preserve an endorsement of its corporate information during a legally stipulated period, the material of the Data Controller has to be destroyed or returned once the legally stipulated period has expired.

&#x20;

**9. Backup copies**

9.1 The Data Processor undertakes not to copy or reproduce the information provided by the Data Controller, except when necessary for its processing or to implement the security measures to which it is legally bound as the data processor.

9.2 In the latter case, each of the copies or reproductions will be subject to the same commitments and obligations set forth in the clauses of this DPA, and must be destroyed or returned, as indicated in the previous Stipulation.

&#x20;

**10. Responsibilities**

10.1 The Data Processor  undertakes to comply with the obligations established in this DPA and in the Data Protection Legislation. Pursuant to Article 82 of the Regulation, and subject to the Master Service Agreement, the Data Processor  is liable to the Data Controller for damages caused to the data subject or third parties, including administrative sanctions, arising from judicial or extrajudicial claims or sanctioning procedures of the Supervisory Authority, which are a consequence of the direct non-observance of the instructions assumed in this DPA and / or of the non-compliance or defective fulfilment of the Regulation.

&#x20;

**11. Other obligations of the Parties.**

11.1. The Data Processor undertakes to:

(i)    Process Personal Data in accordance with the documented instructions of the Data Controller (which may be specific or general) or as periodically notified by the Data Controller directly, and solely and exclusively for the purpose of providing the agreed Services. Likewise, the Data Processor  commits not to apply or use personal data for any purpose other than that of this contract, nor to disclose or transfer them, not even for storage, to other persons.

(ii)   Immediately inform the Data Controller if it considers that any of the Data Controller's instructions violate Data Protection Regulations or any other provision on data protection of the Union or Member States.

(iii)  Maintain the duty of confidentiality regarding personal data accessed under this assignment, even after its termination.

(iv)  Ensure that its authorized personnel or collaborators handling Personal Data have expressly and in writing committed to confidentiality or are subject to a statutory confidentiality obligation, and that they comply with the corresponding security measures, which must be duly communicated to them.

(v)   Make available to the Data Controller all necessary information to demonstrate compliance with its obligations if requested.

(vi)  Keep a written record of all processing activities carried out on behalf of the Data Controller, which must include:

a. Name and contact details of the Data Processor .

b. The categories of processing carried out on behalf of the Data Controller.

(vii)Cooperate with the Data Controller in handling requests for the rights of data subjects in accordance with Data Protection Regulations. In any case, the Data Processor  must immediately forward to the Data Controller any request for the exercise of the rights of access, rectification, erasure, objection, restriction of processing, data portability, and the right not to be subject to automated individual decision-making (including profiling) made by an affected party whose data has been processed by the Data Processor  in compliance with this contract. This ensures that the Data Controller addresses the request within the time limits set by the applicable regulations. The transfer of the request must be made without undue delay as quickly as possible, without undue delay.&#x20;

11.2. The Data Controller acknowledges and accepts that it is their duty to provide information to data subjects whose personal data is collected for processing, at the time of collection.

**12. Personal data Breach**

12.1. In case of suspicion of a personal data breach, the Data Processor will take immediate action, with the purpose of investigating the possible personal data breach and to identify, prevent, mitigate and, if necessary, remedy the effects of the possible personal data breach.

12.2 Immediately notify the Data Controller as soon as there are reasonable grounds to believe that a personal data breach has occurred, and, to the extent possible, provide the Data Controller with a detailed description of the personal data breach, which will include:

(a) The possible impact of the personal data breach.

(b) The categories and the approximate number of affected data subjects, as well as their country of residence and the approximate number of personal data records affected.

(c) The risk that has caused the personal data breach to the data subjects.

(d) The measures adopted or proposed by the Data Processor to remedy the breach of data breach and to mitigate the possible negative effects.

(e) Provide periodic updates and any other information requested by the Data Controller in relation to the personal data breach.

(f) Will not disclose or publish any communication, notice, press release or report related to the breach of personal data without the prior written authorization of the Data Controller (except in the case that it is obliged to do so in accordance with the provision of Law).

If it is not possible to provide the information simultaneously, and to the extent that it is not, the information will be provided gradually without undue delay.

The Data Controller acknowledges that it is their duty to report data security breaches to the Spanish Data Protection Agency and to notify the affected individuals as soon as possible when the breach is likely to pose a high risk to the rights and freedoms of natural persons.

&#x20;

**13. International Transfer of personal data**

13.1 The Data Processor will not process the data outside the European Economic Area or in any country for which the European Commission has not guaranteed an adequate level of protection without the written authorization of the Data Controller or without complying with the requirements established by the GDPR for carrying out international transfers.

**14. Audit**

14.1 The Data Processor shall allow the Data Controller and its respective auditors or authorized agents to verify compliance with the obligations set out in this DPA. For such purposes, the Data Processor shall make available all information reasonably necessary to demonstrate compliance with the obligations established herein, including where applicable and when available, third-party certifications and audit reports (such as ISO, SOC 2, or equivalent), as well as any additional documentation reasonably requested by the Data Controller.

14.2 All information obtained and derived from the audits will be treated as Confidential Information of the Data Processor.

&#x20;

**15. Entry into force**

15.1 This DPA will enter into force upon Client´s acceptance of the Master Service Agreement and shall remain in force for as long as VoiceB processes Personal Data on behalf of the Client under the Agreement. The obligations under this DPA that, by their nature, extend beyond the termination of the Agreement (including confidentiality and data deletion obligations) shall remain in effect after termination.

<p align="center"> </p>

<p align="center"><strong>APPENDIX A</strong></p>

<p align="center"> </p>

<p align="center"> </p>

**Special categories of data:**

No categories of Sensitive Personal Data as defined by Applicable Privacy Law shall be processed for the purposes of this Processing Appendix.

**Personal Data Processing Activities**

**Nature and Purpose of Processing**

Collection, Access, Storage, Consultation, Comparison, Analysis, Anonymization.

The purpose of the processing is the provision of the Service under the Contract.

**Categories of Data Subjects**

The categories of data subjects whose personal data will be processed by the Data Processor  within the scope of service provision are:

* Internet users who access the service from the landing page where the service is presented (data subjects).
* Client.

**Categories of Personal Data**

The categories of personal data accessed by the Data Processor  are as follows:

a)   **Identifying Data:** Name, surname, ID card number, residential address, email address, IP address, mobile phone number, and bank account details.

b)  **User Content:**  Voice recordings, transcriptions, messages, contact information, and any other data provided during phone communications.

c)    **Usage Data**: Interactions with the Service, such as call duration, frequency, metadata.

d)    **Device Data:** Including IP address, browser type, operating system, device identifiers, and mobile network information.

e)    **Website Navigation Data:** Cookies and tracking technologies.

f)      **Data required for DDI procurement:** Information requested by the telecommunications provider (Twilio) for the acquisition of direct dial-in (DDI) numbers, which may include identifying details of the contracting company or individual, legal documentation, billing address, and any other country-specific requirements. The specific information requirements will be subject to the prevailing regulations of the respective country and can be consulted at [Twilio Regulatory Guidelines](https://www.twilio.com/en-us/guidelines/regulatory).

&#x20;

&#x20;

**Special Categories of Personal Data (if applicable)**

No special categories of personal data are processed.

&#x20;

**Retention Period of Personal Data**

For the duration of the contract or for the time specified by the client.

<p align="center"> </p>

**List of sub processors**

The up-to-date list of sub processors engaged by the Data Processor for the provision of the Service is available at the following URL:

<https://docs.voiceb.ai/legal-and-compliance/data-sub-processors>

&#x20;

&#x20;

<p align="center"> </p>

&#x20;

<p align="center"> </p>

<p align="center"> </p>

&#x20;


# Trust Center

Trust Center page

{% embed url="<https://trust.voiceb.ai/>" %}


# Data Sub-processors

Last Updated, July 23rd 2025&#x20;

### 📘 Legal & Compliance: Data Sub-processors

At[ VoiceB.ai](https://voiceb.ai/), we are committed to ensuring transparency, security, and full compliance with data protection regulations, including the General Data Protection Regulation (GDPR) and other applicable privacy laws.&#x20;

To support the delivery of our services, we engage certain reputable third-party service providers (sub-processors) who process customer personal data on our behalf and under our instructions. We carefully assess each provider to ensure they meet high standards of privacy, security, and legal compliance.

Below is a comprehensive overview of our current data processors.

### ✅ What Is a Sub-processor?

A sub-processor is a third party engaged by VoiceB to process personal data on behalf of our customers (the data controllers), while VoiceB acts as a data processor. These providers assist us in delivering functionality such as hosting, AI processing, authentication, and payment services.

Each sub-processor is bound by a Data Processing Agreement (DPA) that includes obligations regarding confidentiality, data protection, security measures, and breach notification.

### 🔐 List of Authorized Sub-processors

| Sub-processor | Description                                           | Jurisdiction  | Data Location             | Purpose                                             | Privacy Policy                                                | Compliance & Security Links                                                                                                                                                                                                                                   |
| ------------- | ----------------------------------------------------- | ------------- | ------------------------- | --------------------------------------------------- | ------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Vercel        | Hosting platform for the VoiceB website and dashboard | United States | EU (configurable), US     | Hosting, Static content & app delivery              | [Privacy Policy](https://vercel.com/legal/privacy)            | [DPF](https://vercel.com/legal/privacy) · [DPA + SCCs](https://vercel.com/legal/dpa) · [Security Measures](https://vercel.com/security) · [Certifications: ISO 27001, SOC 2, HIPAA](https://vercel.com/security)                                              |
| Clerk.dev     | Authentication & user management                      | United States | US (default), EU optional | Login & account handling                            | [Privacy Policy](https://clerk.dev/legal/privacy)             | [DPF](https://clerk.com/changelog/2024-02-29) . [DPA + SCCs](https://clerk.dev/legal/dpa) · [Security Overview](https://clerk.com/docs/security/overview) · Certifications: SOC 2 Type II (in progress)                                                       |
| ElevenLabs    | Voice synthesis                                       | Poland        | EU (configurable), US     | AI voice generation for agent calls                 | [Privacy Policy](https://www.elevenlabs.io/es/privacy-policy) | [DPA + SCCs](https://www.elevenlabs.io/legal/data-processing-agreement) · [Certifications: SOC 2 Type II](https://compliance.elevenlabs.io/)                                                                                                                  |
| OpenAI        | LLM provider (optional, depending on integration)     | United States | US                        | Language processing for AI agent behavior           | [Privacy Policy](https://openai.com/policies/privacy-policy)  | [DPA+SCCs](https://openai.com/policies/data-processing-addendum/) · [Security Practices](https://openai.com/security) · [Certifications: SOC 2 Type II](https://openai.com/security)                                                                          |
| HubSpot       | CRM & Marketing Automation                            | United States | US & EU                   | Sales, support, marketing workflows                 | [Privacy Policy](https://legal.hubspot.com/privacy-policy)    | [DPA + SCCs](https://legal.hubspot.com/dpa) · [Security](https://legal.hubspot.com/security) · [Certifications: SOC 2, ISO 27001, GDPR](https://trust.hubspot.com/?hsCtaTracking=98c3bf73-03eb-4096-9ec7-dfb3078d4b63%7C3ecab3c5-8bc7-4fa6-91fe-9f0d1731cf4d) |
| Stripe        | Payment processing infrastructure                     | United States | EU (configurable), US     | Subscription billing & invoicing                    | [Privacy Center](https://stripe.com/privacy)                  | [DPF](https://stripe.com/privacy) · [DPA + SCCs](https://stripe.com/legal/dpa) · [Certifications: PCI DSS Level 1, SOC 1 Type II, SOC 2 Type II, EMVCo, ISO 27001 aligned](https://stripe.com/docs/security/stripe)                                           |
| Twilio        | Telephony infrastructure & DDI routing                | United States | EU (configurable), US     | Voice call handling and phone number provisioning   | [Privacy Policy](https://www.twilio.com/legal/privacy)        | [DPF](https://www.twilio.com/legal/privacy) · [BCR](https://www.twilio.com/legal/bcr) · [Security Measures](https://www.twilio.com/security) · [Certifications: ISO 27001, ISO 27017, ISO 27018](https://www.twilio.com/security)                             |
| Bland AI      | Voice call infrastructure provider                    | United States | US                        | VoIP infrastructure, call routing                   | [Privacy Policy](https://www.bland.ai/privacy)                | [DPA + SCCs](https://www.bland.ai/legal/dpa) · [Certifications: SOC 2 Type I, SOC 2 Type II](https://www.bland.ai/security)                                                                                                                                   |
| Nango         | Unified API integrations platform                     | United States | EU (configurable), US     | Connecting VoiceB to third-party CRMs via OAuth/API | [Privacy Policy](https://www.nango.dev/legal/privacy)         | [DPF](https://www.nango.dev/legal/privacy) · [DPA + SCCs](https://www.nango.dev/terms#dpa) · [Certifications: SOC 2 Type I](https://www.nango.dev/product-updates/soc-2-type-1-compliant)                                                                     |

### 📄 Data Processing Agreements

VoiceB.ai has a Data Processing Agreement (DPA) in place with each of its sub-processors. These agreements ensure that any processing of customer personal data is conducted in accordance with the GDPR and other applicable data protection laws.<br>

All of our sub-processors:<br>

1. Are subject to a GDPR-compliant DPA with VoiceB.ai
2. Are contractually obligated to implement appropriate technical and organizational security measures
3. May only process data based on our documented instructions
4. Must promptly notify VoiceB.ai in the event of a personal data breach or security incident
5. Are required to support data subjects’ rights to access, rectify, erase, or restrict their personal data

***

### 🌍 International Data Transfers

Some sub-processors are located outside the European Economic Area (EEA). When customer personal data is transferred internationally, we ensure that such transfers are conducted in full compliance with the GDPR, using one or more of the following legal mechanisms:

* Standard Contractual Clauses (SCCs) approved by the European Commission
* Adequacy decisions, including participation in the EU-U.S. Data Privacy Framework, where applicable
* Binding Corporate Rules (BCRs), where adopted by the provider
* Other lawful mechanisms permitted under Chapter V of the GDPR

We thoroughly assess the legal and privacy posture of all sub-processors prior to engagement, and require appropriate safeguards for any international transfer.<br>

***

### 📥 Contact

For any questions or to request a signed copy of our **Data Processing Agreement**, please reach out to:

📧 **<hello@voiceb.ai>**


# Data Journey, Storage & Security

End-to-end data flow, GDPR compliance, and enterprise-grade security. By Alex Bisbe. January 14th, 2026.

This article explains **how data travels through VoiceB**, **where it is processed**, **where it is stored**, and **how it is protected**.\
It is written for **IT, Security, Legal, Procurement, and Enterprise stakeholders** who need a clear, auditable view of the platform.

VoiceB is designed **EU-first**, **privacy-by-design**, and **production-ready for regulated industries**.

***

### High-Level Architecture Principles

Before diving into the diagrams, a few non-negotiables:

* **EU data residency by default**
* **Stateless AI reasoning** (no LLM data storage or training)
* **Clear separation** between real-time processing and persistent storage
* **Client remains Data Controller** at all times
* **VoiceB acts strictly as Data Processor**
* [**VoiceB Data Sub-processors list**](https://docs.voiceb.ai/legal-and-compliance/data-sub-processors)

***

### Process 1 — Inbound Calls (DID / DDI)

#### Data Flow Diagram

```
Caller
  |
  v
Inbound Phone Number (DID / DDI)
  |
  v
Twilio (EU)
(Call routing, SIP, optional recording)
  |
  v
Voice Engine
- ElevenLabs (EU – default)
- Bland (US – backup, GDPR-compliant)
  |
  v
LLM Reasoning Layer
- Gemini 2.5 / Gemini 3 (Google – EU processing)
(NO DATA PERSISTENCE)
  |
  v
Voice Response (TTS)
(ElevenLabs)
  |
  v
VoiceB Control Plane
- Vercel (EU)
- PostgreSQL (EU)
  |
  v
Client CRM / Systems
```

#### What Happens Here

* The call is received via **Twilio (EU)**
* Voice is synthesized and understood by **ElevenLabs (EU)**
* The AI reasoning is executed by **Google Gemini**, processed in the EU
* Results (transcripts, outcomes, structured data) are stored in **Vercel (EU)** backed by PostgreSQL
* Business data is pushed to the **Client CRM**, which becomes the system of record

#### Storage Summary

* Call recordings: **Twilio (EU)** if enabled
* Transcripts & metadata: **PostgreSQL (EU)**
* No data stored in the LLM layer

***

### Process 2 — Click-to-Call (API-Triggered Calls)

#### Data Flow Diagram

```
Website / App User
  |
  v
Click-to-Call API
(Vercel – EU)
  |
  v
Twilio (EU)
(Call initiation)
  |
  v
Voice Engine
- ElevenLabs (EU – default)
- Bland (US – backup)
  |
  v
LLM Reasoning Layer
- Gemini 2.5 / Gemini 3 (EU)
  |
  v
Voice Response
  |
  v
VoiceB Control Plane
- Vercel (EU)
- PostgreSQL (EU)
  |
  v
Client CRM / Lead Systems
```

#### What Changes vs Inbound

* The trigger is an **API event**, not a phone number
* The rest of the architecture is **identical**
* No browser audio or frontend PII is stored

#### Storage Summary

* Same persistence rules as inbound calls
* No additional data captured beyond operational needs

***

### Process 3 — Web VOIP (Widget-Based Calls)

#### Data Flow Diagram

```
Website Visitor
  |
  v
VoiceB Web VOIP Widget
(Vercel – EU)
  |
  v
Twilio (EU)
(WebRTC / SIP bridge)
  |
  v
Voice Engine
- ElevenLabs (EU – default)
- Bland (US – backup)
  |
  v
LLM Reasoning Layer
- Gemini 2.5 / Gemini 3 (EU)
  |
  v
Voice Response
  |
  v
VoiceB Control Plane
- Vercel (EU)
- PostgreSQL (EU)
  |
  v
Client CRM / BI Systems
```

#### Additional Notes

* Web widget configuration is stored in **Vercel (EU)**
* Call recordings (if enabled) are stored **only in Twilio (EU)**
* Fully compatible with WebRTC security standards

***

### Where Data Is Stored (Clear & Simple)

#### Stored in the EU

* Call metadata (timestamps, duration, outcomes)
* Transcripts (optional, configurable)
* Structured fields collected during calls
* Prompts, configurations, and audit logs

#### Not Stored Anywhere

* LLM prompts beyond runtime execution
* Any data inside Gemini
* Any data reused for training
* Any cross-customer data

***

### Data Processing Agreement (DPA)

VoiceB provides a **standard GDPR-compliant DPA**, aligned with **Articles 28–32**.

#### Roles

* **Client** → Data Controller
* **VoiceB** → [Data Processor](https://docs.voiceb.ai/legal-and-compliance/data-sub-processors)

#### Covered Topics

* Scope and purpose of processing
* Categories of data subjects and data types
* Sub-processors and hosting regions
* Technical and organizational measures
* Incident notification SLAs
* Data subject rights and deletion workflows

#### [Sub-Processors](https://docs.voiceb.ai/legal-and-compliance/data-sub-processors) (Default)

* **Twilio** — EU
* **ElevenLabs** — EU
* **Google (Gemini)** — EU
* **Vercel** — EU

US-based providers (e.g. **Bland**) are **failover-only**, GDPR-compliant, and covered by SCCs.

***

### Security Architecture

#### Infrastructure Security

* EU-based cloud infrastructure
* Logical tenant isolation
* Hardened environments
* Continuous monitoring

#### Data in Transit

* TLS encryption end-to-end
* Secure SIP and WebRTC channels
* Signed and authenticated APIs

#### Data at Rest

* Encrypted PostgreSQL databases (EU)
* Encrypted call recordings (Twilio EU)
* Encrypted backups with restricted access

#### Access Control

* Role-based access (RBAC)
* Least-privilege principle
* Full audit logs
* No internal access to content without authorization

***

### Retention & Deletion

* Retention policies are **client-configurable**
* Automatic deletion supported
* Manual purge on request
* Full GDPR Right-to-Erasure compliance
* Call recordings can be disabled entirely

***

### What VoiceB Explicitly Does NOT Do

* ❌ No training on customer data
* ❌ No data resale or reuse
* ❌ No hidden storage in AI models
* ❌ No uncontrolled sub-processors

***

### Enterprise Readiness

VoiceB is built to pass:

* Enterprise security reviews
* GDPR and privacy assessments
* Telecom, energy, and banking compliance checks
* Procurement and legal due diligence

Available on request:

* Signed DPA
* Sub-processor list
* Visual data flow diagrams (PDF / slides)
* Security annex

***

### Bottom Line

VoiceB’s data journey is:

* **Transparent**
* **EU-first**
* **Contractually protected**
* **Technically locked down**

Data flows in, intelligence happens in real time, outcomes flow to your systems—**and nothing unnecessary persists**.

That’s the architecture. No shortcuts. No ambiguity


# Cookies

### **Use of Cookies.**  Please note that the website uses both our cookies and third-party cookies in order to improve our services and to display advertisements related to your preferences by analyzing your browsing habits.

Cookies are tools used by web servers to store and retrieve information on their visitors. They are small text files which we send to your computer in order to maintain a record of your preferences and remember them upon your return.

These cookies can store information in order to show contents and offers of special interest in accordance with your profile and to register the number of times a message is shown to a User (general novelties in the website).

Users may freely decide about whether cookies used are implemented in their hard drives or to receive a warning on their screens when a cookie is received and to decide whether or not to implement it in the hard drive. In order to do so, we suggest using the help section in your web browser to find out how to change the configuration currently being used.

For more information on the use of Cookies please visit our Use of Cookies Policy.


