AI-powered workflow isn’t triggering
You enabled AI features for standups or follow-ups, but summaries, blocker detection, or downstream actions never fire. In most cases the workflow is gated by plan, configuration, or credentials rather than a random outage.
Quick check
- Plan and feature flags — Confirm your workspace is on a plan that includes AI features and that AI is turned on in org or check-in settings.
- Check-in configuration — Open the specific check-in and verify AI options (summaries, workflows, or escalations) are enabled for that template, not only at the org level.
- API key or provider — If your org brings its own model key, confirm it is valid, not rotated out, and has the scopes the integration expects.
- Rate limits — Burst traffic or provider quotas can silently drop non-critical calls; check your provider dashboard for throttling around the times failures happen.
- Minimum data — Some AI steps need enough responses or text before they run; a brand-new check-in with zero submissions will not produce output.
Common causes and fixes
AI features not enabled on the plan
Dailybot gates advanced AI workflows to plans that include them. If you recently downgraded or are on a trial that ended, toggles may gray out or run in limited mode. An org admin should review Billing or Plan in the web app and compare to the feature list on the pricing page. After upgrading or re-enabling AI, wait for the next check-in cycle or trigger a test submission.
Check-in not configured for AI
Org-level AI can be on while an individual check-in still uses a legacy template without AI steps. Edit the check-in, open AI or automation sections, and attach the workflow you expect (summary, blocker scan, etc.). Save and have one member submit a full test response. If the workflow only runs on certain questions, ensure those questions are present and required.
API key issue
Workspaces that supply an API key must keep it current. Rotated keys, revoked keys, or keys tied to the wrong project break inference calls. Replace the key in Dailybot settings, save, and run a small test. If your security team uses IP allowlists, confirm Dailybot’s outbound IPs are permitted. Never paste keys into chat; use the secure fields in the product only.
Rate limiting
Cloud model providers return 429 or soft-fail when quotas are exceeded. Symptoms include intermittent success and “sometimes it works after lunch.” Check usage charts in the provider console. Mitigations include raising the quota, spreading check-in times, or reducing optional AI steps on very large teams. If Dailybot shows an explicit rate-limit message, wait and retry after the window it names.
Empty or malformed inputs
AI workflows may skip when responses are empty, too short, or fail validation. Ensure required fields are filled and that custom question IDs did not change after you built the workflow. After large template edits, send a fresh test response so the pipeline has clean data.
If none of this worked
Before contacting support, gather:
- Organization name and approximate team size
- Check-in name and whether AI ever worked for it
- Whether you use Dailybot-managed AI or your own API key (not the secret itself)
- Any error banners in the web app and approximate timestamps (with timezone)
- Provider quota screenshots if you suspect rate limits
- Steps you already tried from this article
Then contact Dailybot support from the Help or Contact options in the product or on the website.