Learn how to set up AI Voice Agents in ChatAgentLab to handle inbound phone calls automatically.
You can activate AI Voice by using:
A phone number purchased directly inside ChatAgentLab
(pricing varies by country; e.g., US numbers start around $1.75/month)
OR your own existing Twilio phone number connected through the platform
Phone numbers and AI Voice minutes are billed directly to the client.
Configure the agent by choosing:
Then set optional call settings such as background audio, call duration limits, call recording, the time zone, and the number to transfer calls to.
Your agent works best when it has access to the right information.
If your business has more details than you want to embed directly in the prompt, you can upload:
ChatAgentLab automatically trains your agent to reference this information during calls.
Extend your agent’s capabilities by connecting it to external systems such as:
This lets your AI perform real actions like creating leads, sending confirmations, or booking appointments.
After every call, you can:
Controls which AI model powers your voice agent.
Each model has its own pricing and performance profile.
Primary model (e.g., GPT-4.1, ~$0.20/min) is ideal for most business calls.
Think of this as choosing the “engine” behind your agent.
Choose which languages the agent understands and responds in.
Choose the AI voice persona callers will hear.
Each voice has a unique style, tone, and delivery.
Pick what matches your brand — friendly, professional, calm, energetic, etc.
Controls creativity in responses:
The greeting callers hear first.
Example:
“Hi, you’ve reached the virtual assistant for [Company Name]. How can I help you today?”
This sets the tone and clarifies who they’re speaking with.
This defines how your AI agent behaves.
Two parts typically exist:
Guidance for tailoring the prompt to your business.
Includes:
The logic your agent follows, such as:
You can adapt it for any client or industry.
Optional subtle audio to make the AI sound more “live.”
Examples: Coffee shop ambience, outdoor sounds, static.
Limits how long each call can last.
The agent will politely wrap up when the limit is reached.
Enable call recording for:
Ensure you follow local recording laws.
Determines whether the AI reacts to keypad input (DTMF).
Used for timestamps, scheduling, and call reports.
Enter the phone number to transfer calls to under specific conditions, such as:
Format: +1234567890 (include country code)
The Knowledge Base gives your agent deeper context about your business.
Use it for:
Your AI references the data conversationally — it doesn’t read or repeat it verbatim.
Paste any text you want the agent to understand.
Examples:
FAQ lists, service descriptions, policies, onboarding instructions.
Click Train Chat Agent to apply.
Upload documents such as PDFs, spreadsheets, or Word files.
Useful for:
The system automatically indexes these.
Add URLs from your website so your agent always has updated info.
Useful for:
Pricing pages, feature lists, product pages, blog posts.
Custom Tools allow the agent to perform actions through APIs — not just talk.
Useful for:
Define:
The AI will call the tool whenever a matching intent appears during the conversation.
The Analytics section helps you track performance and automate follow-up.
Defines how the call should be summarized.
Examples:
Create custom fields the AI should extract during calls:
These fields appear in:
Automatically send call data to external systems.
Use cases:
Receive call summaries and lead details directly in your inbox.
Useful for small teams that want immediate updates.