Customizing Your AI Assistant

Learn how to tailor your AI assistant to best serve your needs.

Basic Customization

  1. Navigate to your assistant's settings page
  2. Adjust the assistant's name, description, and basic parameters
  3. Configure the initial message your agent will send to users
  4. Add preconfigured responses for quick user interaction
  5. Save your changes

Knowledge Base

  1. In the assistant dashboard, locate the knowledge base section
  2. Upload relevant documents or data to expand your assistant's knowledge
  3. Organize and manage the uploaded information as needed
  4. Your assistant will automatically reference this information when answering questions

Email Collection

You can configure your assistant to collect user email addresses:

  1. Enable the email collection feature in your agent's settings
  2. Choose when to collect emails:
    • At the beginning of the conversation
    • At the end of the conversation
  3. The assistant will naturally ask for the user's email at the specified time

Custom API Functions

Your assistant can now make calls to external APIs:

  1. In the Custom API Functions section, click "Add API Function"
  2. Configure your API endpoint:
    • Function Name: A unique identifier for the API (e.g., "get_weather")
    • Description: Explain what the API does
    • Endpoint URL: The API endpoint (e.g., "https://api.weather.com/current/{city}")
    • Parameters: Define the parameters your API accepts in JSON Schema format

Example parameters for a weather API:

{
  "city": {
    "type": "string",
    "description": "The city to get weather for",
    "required": true
  }
}

The assistant will automatically use these APIs when relevant to the conversation.

Behavior Configuration

  1. Define the assistant's personality and tone
  2. Specify how it should handle different types of queries
  3. Configure response formats and styles
  4. Set up any restrictions or limitations
  5. Test the changes to ensure they meet your requirements

Tools and Capabilities

Your assistant can use several built-in tools:

  1. Code Interpreter: For handling programming-related queries
  2. File Search: To search through uploaded knowledge base
  3. Email Collection: For gathering user email addresses
  4. Custom API Functions: For making external API calls

Enable or disable these tools based on your needs.

Integration Options

You can integrate your AI assistant into your website using our embeddable chat widget:

  1. In your agent's settings, find the "Show Embed Code" button
  2. Click the button to reveal the embed code
  3. Copy the provided JavaScript snippet
  4. Paste the snippet into your website's HTML, just before the closing </body> tag

The embed code will look similar to this:

<script src="https://app.agnt.one/embed.js"></script>
<script>
  window.onload = function () {
    window.chatWidgetConfig = {
      url: "https://your-site-subdomain.agnt.one",
      id: "your-agent-id",
    };
  };
</script>

Testing and Refinement

  1. Use the Preview feature to test your assistant's behavior
  2. Monitor conversations to ensure proper functioning
  3. Adjust settings based on actual usage patterns
  4. Regularly update the knowledge base and API configurations as needed

Best Practices

  1. Keep API endpoints public and authentication-free for best compatibility
  2. Use clear, descriptive names for custom API functions
  3. Provide detailed descriptions for API parameters
  4. Test API integrations thoroughly before deployment
  5. Regularly review and update custom API configurations

For more detailed information or assistance, please contact our support team.