TICKET MANAGEMENT

Agent triages each new Intercom conversation and routes only true bugs to deduplicated Linear

An agent reads every inbound Intercom conversation, decides whether it is a bug, a feature request, or a how-to question.

CategoryTicket Management
Enginepaperclip
Difficultyadvanced
Triggerevent
Steps5
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerNew Intercom conversation createdIntercomIntercom
  • ActionAgent classifies bug vs. request vs. questionOpenAI
  • LogicRoute: bugs continue, others tag and stop
  • ActionFingerprint and dedup into LinearLinearLinear
  • OutputRecord classification on Intercom conversationIntercomIntercom

What it does

Puts an intelligent gate in front of escalation. Not every "it's broken" message is a real defect, so an agent classifies each conversation first and only genuine bugs get fingerprinted and pushed into the deduplicated Linear backlog. Feature requests and how-to questions are tagged and left in Intercom for the support team.

When to use it

Use it when raw conversation volume is too high to tag bugs by hand and you want a judgment layer that separates real defects from noise before anything reaches engineering.

How it works

  1. 1A new Intercom conversation arrives and fires the trigger.
  2. 2The agent reads the full thread and classifies it as bug, feature request, or question, extracting the symptom and affected area when it is a bug.
  3. 3A logic step routes only bug classifications onward; everything else gets an Intercom tag and stops.
  4. 4For bugs, the agent fingerprints the symptom and searches Linear, then either increments the count on the matching issue or files a new one with reproduction notes.
  5. 5It records the classification and any Linear link back on the Intercom conversation.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect IntercomConversations, contacts, articles.
  2. 2
    Connect LinearIssues, projects, cycles, triage.
  3. 3
    Connect OpenAIModels, embeddings, files.
  4. 4
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  5. 5
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  6. 6
    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.