Topic insights and what to automate next
Topic insights and what to automate next
What your customers ask about, how well each topic is handled, and the next best improvement.
Tickets are grouped into topics for the last 7, 30 or 90 days. For each topic you see volume and trend, the share answered automatically, review and triage rates, knowledge coverage, paid-search use, reopen rate and an estimated satisfaction score.
| Recommendation | Means |
|---|---|
| Write a help article | Low knowledge coverage and mostly triaged |
| Review drafts / adjust threshold | Good coverage but mostly held for review |
| Answers aren’t working | High reopen rate — check the articles used |
| Needs a backend action | Customers ask TR-AI to do something (refund, cancel, order status) — add a procedure |
| Working well | Already mostly answered automatically |
Tips & tricks
- Work through “Automate next” from the top — it is ranked by volume × how much is still manual × how feasible it is.
- Click “Name topics with AI” once for friendlier topic names (one short AI call per topic, cached).
- Estimated satisfaction is a heuristic; real CSAT replaces it wherever you collect it.