TICKET MANAGEMENT

Intercom Duplicate-Conversation Agent Resolver

An agent investigates incoming Intercom conversations against the open queue, decides whether each is a duplicate of an active thread.

CategoryTicket Management
Enginepaperclip
Difficultyadvanced
Triggerevent
Steps5
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerNew Intercom conversationIntercomIntercom
  • ActionAgent searches open conversation queueIntercomIntercom
  • LogicAgent reasons + applies confidence gateOpenAI
  • ActionMerge context into canonical conversationIntercomIntercom
  • OutputTag duplicate + post rationale noteIntercomIntercom

What it does

When a new Intercom conversation arrives, an agent reads it, searches the open conversation queue for siblings about the same problem, and reasons about whether to fold it into an existing canonical thread. If it merges, it carries the customer's new detail into the canonical conversation and tags the duplicate so agents see the full picture in one place.

When to use it

For support inboxes where the same outage or billing issue generates a flood of separate chats. Use it when judgment matters more than a fixed score — the agent weighs phrasing, customer, and timing rather than a single similarity number.

How it works

  1. 1A new Intercom conversation triggers the agent.
  2. 2The agent searches open Intercom conversations and pulls candidate siblings.
  3. 3It reasons over the candidates, applying a confidence gate before declaring any duplicate.
  4. 4For a confident match, the agent appends the new context to the canonical conversation and tags the duplicate.
  5. 5The agent posts a short rationale note and, when no confident match exists, leaves the conversation untouched and logged as unique.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect IntercomConversations, contacts, articles.
  2. 2
    Connect OpenAIModels, embeddings, files.
  3. 3
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  4. 4
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  5. 5
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.