Advanced AI Assistant Settings

Text-to-Speech (TTS) controls how your AI assistant's voice sounds to callers. AIVO supports multiple providers, each with different strengths.

Dashboard Configuration

Navigate to Settings > AI Advanced and select your preferred TTS provider from the dropdown. Configure provider-specific credentials if needed.

Speech-to-Text (STT) determines how accurately your assistant understands what callers say. Choose based on accuracy needs, language support, and latency requirements.

When enabled, expressive mode allows the AI to vary its tone, pace, and emphasis based on context. The assistant will sound more natural and engaging rather than reading in a flat, robotic tone.

How to Enable

In the dashboard, go to Settings > AI Advancedand toggle "Expressive Mode" on. Save your changes and the next call will use the new setting.

The silence_timeout_mssetting controls how long (in milliseconds) the assistant waits after the caller stops speaking before it considers the caller's turn complete and begins responding.

Controls how easily a caller can interrupt the AI while it is speaking. This affects the conversational flow and determines whether the assistant will stop mid-sentence when it detects the caller talking.

Sets the maximum length (in seconds) for any single call. When reached, the assistant will politely wrap up the conversation and end the call. This prevents runaway calls and helps control costs.

Set the max call duration in seconds from Settings > AI Advanced > Call Limits. Defaults to 1800 (30 minutes); Enterprise can request up to 7200 (2 hours).

Template variables let you inject dynamic data into your assistant's system prompt at call time. Use double curly braces to define placeholders that get replaced with real values when a call starts.

Available Variables

Example Prompt

You are the AI receptionist for {{business_name}}.
The current date is {{current_date}} and time is {{current_time}}.
Our business hours are {{business_hours}}.

If the caller asks for an appointment, check availability
and book one for them.

Passing Custom Variables via API

Custom variables are passed when triggering an outbound call through the AIVO public API. Set them under variables in the request body:

curl -X POST "https://aivo.bz/api/v1/calls/outbound" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "to": "+15551234567",
    "variables": {
      "custom.promo_code": "SUMMER25",
      "custom.vip_tier": "gold"
    }
  }'

Custom tools let your AI assistant call external functions during a conversation. When the AI determines it needs to perform an action (like booking an appointment or looking up order status), it triggers a tool call that hits your webhook endpoint.

How It Works

1. Define a tool with a name, description, and parameter schema
2. The AI decides when to call the tool based on conversation context
3. AIVO sends a webhook to your endpoint with the tool name and parameters
4. Your server processes the request and returns a result
5. The AI uses the result to continue the conversation

Dashboard Setup

Go to Toolsin the sidebar. Click "Add Tool" and fill in the tool name, description (used by the AI to decide when to call it), parameters (JSON Schema), and your webhook URL.

Tool Definition Schema

Tools follow standard JSON Schema for parameters. Define them in the dashboard or POST them to the AIVO public API:

{
  "type": "function",
  "function": {
    "name": "check_appointment_availability",
    "description": "Check if a time slot is available for booking",
    "parameters": {
      "type": "object",
      "properties": {
        "date": {
          "type": "string",
          "description": "Date in YYYY-MM-DD format"
        },
        "time": {
          "type": "string",
          "description": "Time in HH:MM format"
        }
      },
      "required": ["date", "time"]
    }
  }
}

Webhook Payload

When the tool is triggered, your webhook receives:

{
  "tool_call_id": "call_abc123",
  "function": {
    "name": "check_appointment_availability",
    "arguments": {
      "date": "2026-04-15",
      "time": "14:00"
    }
  },
  "conversation_id": "conv_xyz789",
  "assistant_id": "asst_def456"
}

Your endpoint should return a JSON response with a result field:

{
  "result": "The 2:00 PM slot on April 15 is available. Would you like to book it?"
}

Create a custom voice for your AI assistant. There are two approaches: designing a voice from a text prompt or cloning an existing voice from an audio sample.

Dashboard

Navigate to Settings > AI Advancedand find the Voice section. Click "Custom Voice" to access the voice designer or upload an audio file for cloning.

The knowledge base feeds your AI assistant with business-specific information. When callers ask questions, the assistant searches your knowledge base to provide accurate, up-to-date answers.

How Sync Works

1. You create or import articles into your knowledge base
2. Articles are chunked and converted to vector embeddings
3. Embeddings are synced to the AI assistant's retrieval index
4. During calls, relevant chunks are retrieved based on the conversation
5. The AI uses these chunks as context to answer questions accurately

Best Practices

Dashboard

Go to Knowledge Basein the sidebar. Create articles manually, import from a CSV/JSON file, or connect a Notion database to auto-sync. After creating or updating articles, click "Sync to AI" to push changes to the assistant.

API: Create Article

curl -X POST "https://aivo.bz/api/v1/knowledge" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Business Hours and Location",
    "content": "We are open Monday through Friday, 9 AM to 5 PM CST. Located at 123 Main Street, Suite 200.",
    "category": "general"
  }'

API: Trigger Sync

curl -X POST "https://aivo.bz/api/v1/knowledge/sync" \
  -H "Authorization: Bearer YOUR_API_KEY"