Intercom Fin AI

Connect Intercom's Fin AI agent to Klero so that feature requests mentioned during support conversations are automatically captured as feedback items — no human action required.

How it works

When a user mentions a feature request in an Intercom conversation, Fin detects the intent and calls a Klero API endpoint to create the feedback item on the correct board. The item then appears in your Klero project where your team can review, triage, and let other users vote on it.

No Intercom app install required

This integration does not require installing a Klero app from the Intercom marketplace. Authentication is handled entirely by a Project API key.

Prerequisites

• A Klero project with at least one feedback board
• An Intercom workspace with Fin AI enabled

Step 1 — Create a Project API Key

In Klero, go to Settings → API Keys and create a new key. Give it a descriptive name like Intercom Fin. Copy the key — you will not be able to see it again.

Step 2 — Configure a Fin Custom Action

In Intercom, go to Settings → Fin AI Agent → Actions and create a new Custom Action.

Set the authentication to Bearer token and paste your Klero API key as the token value.

When prompted for an OpenAPI spec, paste the following:

openapi: 3.1.0
info:
  title: Klero Feedback
  version: 1.0.0
servers:
  - url: https://app.klero.ai
paths:
  /api/intercom/fin/feedback:
    post:
      operationId: createFeedbackItem
      summary: Create a feedback item from a conversation
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              required:
                - title
              properties:
                title:
                  type: string
                  description: Short title for the feedback item
                description:
                  type: string
                  description: Additional detail about the request
                board_slug:
                  type: string
                  description: Slug of the feedback board to post to. Omit to use the default board.
                customer_email:
                  type: string
                  format: email
                  description: Email of the user who made the request
                customer_name:
                  type: string
                  description: Name of the user who made the request
      responses:
        '201':
          description: Feedback item created
          content:
            application/json:
              schema:
                type: object
                properties:
                  success:
                    type: boolean
                  item:
                    type: object
                    properties:
                      ulid:
                        type: string
                      name:
                        type: string
                      url:
                        type: string

Step 3 — Write a Fin instruction

In your Fin AI instructions, tell Fin when to use the action. For example:

When a user asks for a new feature or describes something they wish the product could do, use the createFeedbackItem action to log their request. Use the user's name and email from the conversation context. Set the title to a concise summary of the request and include the user's own words in the description.

Endpoint reference

POST https://app.klero.ai/api/intercom/fin/feedback

Field	Type	Required	Description
title	string	Yes	Short title for the feedback item
description	string	No	Additional detail
board_slug	string	No	Target board slug. Defaults to the first board.
customer_email	string	No	User's email — links the item to an existing customer or creates one
customer_name	string	No	User's display name

Authentication: Authorization: Bearer YOUR_API_KEY

Response

{
  "success": true,
  "item": {
    "ulid": "01hwx...",
    "name": "Dark mode support",
    "url": "https://app.klero.ai/app/feedback/item/01hwx..."
  }
}