Setting Up Team Insights
Connect your source control provider
Team Insights needs your PR and commit history to compute anything. Connect GitHub, GitLab, or Bitbucket (self-hosted GitLab included) from Connectors if you haven't already — this is the same connection Code Review uses, so if the bot is already installed, you're covered here too.
Run the initial backfill
The first time you open Team Insights — or anytime from the Refresh Data button — Entelligence backfills your recent PR history so the Overview tab has data to show. This uses the source-control connection from step 1 directly; there's no separate token to paste in. It typically takes 3–10 minutes depending on how many PRs your org has, and tabs unlock incrementally as their data becomes ready rather than waiting for everything at once.
Connect AI tool keys (for the AI Analytics tab)
Get an API key from cursor.com/settings and paste it into the AI Analytics connect step.
After connecting, you'll map each tool's users (by email) to your Entelligence team members so per-engineer AI usage attributes correctly.
GitHub Copilot support is on the way — not connectable yet.
Connect your CI provider (for Deployment Analytics)
GitHub Actions is auto-detected once GitHub is connected — nothing to do there. For other CI providers:
A personal API token from CircleCI → User Settings → Personal API Tokens → Create New Token, plus an optional project slug (gh/org/repo) if you want to scope it.
All three are token/API based — no webhook to configure on your end. Once connected, deployment data syncs automatically (this can take a few minutes); if a sync fails, a Retry button appears.
Connect your support tool (for Support Metrics)
A 4-step wizard:
- Intro — overview of what you'll get.
- Connect Tools — supported today: Zendesk, Intercom, Pylon. Zendesk asks for your subdomain first, then authorizes; Intercom and Pylon connect via OAuth directly.
- Review Mapping — match each support agent (by email) to the corresponding Entelligence team member, one row at a time — this is what lets the dashboard attribute ticket metrics to the right person.
- Generate Insights — pulls the last 30 days of ticket data and builds the initial dashboard.
Freshdesk isn't supported yet — Zendesk, Intercom, and Pylon are the current options for Support Metrics.
Troubleshooting
| Symptom | Likely cause | Fix |
|---|---|---|
| Overview tab stays empty after connecting | Backfill still running | Give it a few minutes — check for the "Setting up your engineering context" indicator; tabs unlock as their data becomes ready |
| "No access token found" when triggering a refresh | No Git provider connected yet | Connect GitHub, GitLab, or Bitbucket from Connectors first |
| AI Analytics shows no data after connecting a key | User mapping incomplete | Revisit the connect step and confirm each tool user is mapped to an Entelligence team member |
| Deployment sync fails | Token expired or scope insufficient | Re-check the token's scope (Azure DevOps needs Build Read) and reconnect; use the Retry button on the sync screen |
| Support Metrics shows the wrong person for a ticket | Agent-to-member mapping is off | Revisit Review Mapping in Support Metrics onboarding and correct it — takes effect on the next data pull |
Next steps
- Want a recurring summary in Slack instead of checking the dashboard? Set one up in Pulse.
- Looking for sprint-level reporting instead of continuous metrics? See Sprint Assessments.