Skip to main content

AI Actions of LiveChatAI

AI Actions via Webhook

Webhooks allow your AI chatbot to communicate with any third-party system in real time. With a webhook-based AI Action, your bot can send structured data to an external endpoint and handle the response dynamically inside the chat.

🌟
This is ideal for triggering custom workflows, updating databases, sending emails, or retrieving business-specific results — all while your visitor is still in the conversation.

When to Use a Webhook

Use the Custom Webhook AI Action when:

  • You want to integrate with a system that has a RESTful API
  • You need the AI to take a dynamic action based on real-time user input
  • You want to return data back to the chat interface (e.g., delivery estimates, ticket status)
  • You don’t want to rely on a platform like Zapier or Make.com

How to Set Up a Webhook AI Action

  1. Go to AI Actions > Create New
  2. Select Custom Webhook under Create from Scratch
  • Fill in the following fields:

1. AI Action Name

Provide a clear name to describe what the action does (e.g., check-order-status or send-lead-to-crm).

2. Condition

Define a condition that triggers this action, such as: Trigger this webhook when the user asks about order delivery.

3. Endpoint URL

Enter the webhook URL where data should be sent. This should be a valid HTTPS endpoint that accepts POST requests.

4. Define Inputs (Optional)

These are variables collected from the chat conversation and sent to your endpoint. Each input includes:

  • Name (e.g., order_id, email)
  • Description (required)

Example:

{
  "order_id": "12345",
  "email": "[email protected]"
}

These inputs are passed as part of the webhook payload.

5. Fallback Message

Set a fallback message to display in case the webhook fails or does not return a usable response.

Webhook Payload Format

The chatbot sends a POST request with a JSON body structured like this:

{
  "order_id": "12345",
  "email": "[email protected]",
  "chatbot_id": "clm4ua3fj0001140f8bf7n8qx",
  "city": "abc123"
}

Response Format

Your endpoint should return a response in the following format:

{
  "response": "Your order #12345 is on its way and will arrive by Tuesday."
}

This message will be shown directly to the visitor in the chatbot.

Testing the Webhook

Once your action is set up:

  1. Click Send Test Data to send a sample payload.
  2. Check your server logs or response handler to verify everything is working.
  3. Click Update to save.

Tips for Success

  • Use clear conditions to prevent your webhook from triggering unnecessarily.
  • Secure your webhook with token validation or IP whitelisting.
  • Always return a valid response key in your response object to ensure the AI replies properly.

Good Use Cases

🌟
Checking order or booking status
Sending leads to your CRM
Fetching user-specific data from your database
Registering webinar/event signups
Generating custom pricing quotes

Need help setting up your webhook? Contact Support and we’ll be happy to assist.