Tips & Best Practices
Practical advice for getting the most out of Pulse AI for Jira.
Scanning
Run a scan before important meetings. Pulse Score and findings only update after a scan. Trigger a manual scan via Settings → Scan Schedule → Run Now before governance reviews, sprint retrospectives, or executive reporting sessions.
Use Scan History to verify data freshness. If findings look stale, check Settings → Scan History to confirm the last scan completed successfully. A failed scan leaves the previous scan's data in place.
Automate with the daily schedule. Pulse AI runs an automated daily scan. For most teams this is sufficient. Only trigger manual scans when you need real-time data for a specific meeting or audit.
Reading Findings
Start with High findings. High findings have the most impact on your Pulse Score and carry real governance or delivery risk. Fix these before addressing Medium or Low findings.
Use Clean Signals as confirmation. Clean Signals (green) at the bottom of each section confirm that Pulse AI checked for a specific risk and found none. They are not just absence of findings - they are affirmative passing checks.
Don't over-index on Low findings. Low findings are hygiene improvements, not urgent risks. Fixing a Low finding moves your score marginally. Prioritise High and Medium findings first.
Using the Pulse AI Assistant
Always scope to the current section. The FAB assistant on each section page is scoped to that section's data. If you ask a sprint question from the Governance section, the answer may be less precise. Navigate to the relevant section first.
Ask follow-up questions. The assistant maintains context within a session. If it gives a partial answer, ask "Can you show me the specific items?" or "Give me the JQL filter for those issues."
Use the AI Prompt Playbook for ready-to-use prompts tested against each section.
Governance Cleanup
Fix Dead Owner filters before deleting user accounts. Deleting a Jira user account whose filters power dashboards or boards will break those views. Always transfer filter ownership first. See Global Config →.
Do not delete Guard objects. The Default Permission Scheme, Default Software Scheme, and system groups (jira-administrators, jira-software-users) are Guard objects. Pulse AI will not flag these for deletion. Do not attempt to delete them manually.
Deactivate users in Atlassian Admin, not Jira. User deactivation must be performed at admin.atlassian.com. The Jira REST API does not support direct deactivation on Jira Cloud.
Custom Fields
Remove a field from all screens before deleting it. A custom field marked Risky Deletion is unused but still appears on screens. Deleting it will break those screens. Remove it from each screen first, then delete the field.
Global context fields with < 2 projects are Low-risk bloat. These fields appear on every Jira issue payload even though very few projects use them. Narrowing their context to specific projects improves API performance.
Sprints
Carryover > 20% is a planning signal, not a team failure. Chronic high carryover usually means sprint scope is too large relative to team velocity. Use the Pulse AI Assistant to analyse trends across multiple sprints before drawing conclusions.
Utilisation badges are forward-looking. The At Risk and Overcommitted badges are calculated from the current state of the sprint (remaining story points vs. capacity). They reflect what is likely to happen at the end of the sprint if nothing changes.