Connect to IntercomDetect Knowledge Gaps in Intercom Fin AI
Fin AI answers questions based on the content you feed it. When customers ask questions that fall outside that content, Fin escalates. When Fin answers incorrectly, customers rate low. Both signals point to knowledge gaps — the question is whether you see them quickly enough to fix them.
The reactive approach
- 1Review escalation transcripts after enough of them accumulate.
- 2Identify the questions Fin failed to answer.
- 3Create or update knowledge base articles to address the gap.
The problem: Accumulation-based review means knowledge gaps persist for weeks before being identified. Every day the gap exists, Fin is deflecting customers to human agents unnecessarily.
With Supportman
Dissatisfied emoji ratings on Fin-handled conversations post to Slack within seconds. Your team can review the transcript immediately and tag the gap reason in the thread, building a prioritized backlog in real time.
- Route 🙁 😠 😐 ratings to a dedicated DSAT channel in the Supportman App Home (Pro and above).
- Team member reviews the Fin conversation from the Slack link and tags the reason (e.g. "pricing question", "feature X not covered").
- Content team has a real-time backlog of gaps to address, not a monthly batch.
Knowledge gaps show up in customer ratings first — Supportman posts dissatisfied Fin CSAT to Slack in real time so your team can tag and fix gaps while context is fresh.
Can Supportman automatically categorize escalation reasons?
Not currently — categorization is done by a human in the Slack thread. Automated categorization is on the roadmap.
How do I know when a knowledge gap has been fixed?
After updating the knowledge source in Intercom, monitor dissatisfied ratings for that topic in your DSAT channel. A successful fix shows up as fewer 🙁 😠 ratings on similar Fin conversations over the following weeks.