> For the complete documentation index, see [llms.txt](https://docs.voiceb.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.voiceb.ai/agent-studio.md).
# Agent Studio
{% 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.
*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.
*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. |
*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.
*Ask for metrics and Agent Studio returns a structured performance overview across campaigns.*
*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").
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.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://docs.voiceb.ai/agent-studio.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.