AI blocker detection: how it works
AI blocker detection reads check-in responses for language that indicates work is stalled: waiting on another team, missing permissions, unclear decision makers, repeated “blocked” or “stuck” phrasing, or dependencies named without resolution. It runs in addition to any explicit blocker question or checkbox the template includes.
How blockers surface
When the model flags a candidate blocker, Dailybot can attach a label or section in the smart summary, list it in dashboard views, and feed follow-up automations (for example a nudge to the reporter or a task for a manager). Exact UI labels depend on your version of the product; the intent is to make implicit risk visible without forcing every team to use the same question wording.
False positives
A casual “waiting on review” might be normal process, not a blocker. Users and managers can dismiss or correct a flagged item when the product exposes that control. If detection is noisy, tighten check-in prompts so people use the official blocker field, or turn down AI blocker surfacing in settings until templates improve.
Manager notifications
Org and team admins can choose whether managers receive digests or instant alerts when new AI-detected blockers appear. Channel permissions still apply: notifications go to people configured in Dailybot, not to everyone in a public channel by default.