Back to the Playbook
OperateLearningOP-06

Run product retrospective

Convert delivery experience into a few owned improvements, with the learnings linked to epics and code and fed back into Discover.

Product statusAssisted workflow

Context in4 required inputsCompleted work · Previous retro actions · Delivery signals such as cycle time and rework
TransformationProduct Manager + Product AgentAtlassian · Local Files
OutcomeRetro summary with owned improvement actions.Artifact: Improvement actions

Owner

Product Manager

Human-led

Participants · Agents

EngineerDesignerProduct Agent

Trigger

ScheduledEnd of a sprint, release or initiative, or a recurring team cadence.

Required context

JiraCompleted workJiraPrevious retro actionsJiraDelivery signals such as cycle time and reworkJiraIncidents and blockers

Optional context

CodeQuality metricsFilesTeam feedbackProduct BrainOutcome reviews

Skills

AtlassianLocal Files

Activity steps

  1. 1Review the previous actions.
  2. 2Gather observations and cluster the themes.
  3. 3Identify root causes.
  4. 4Choose one to three improvements with owner and date.
  5. 5Feed durable learnings into the Product Brain.

Output

Retro summary with owned improvement actions.

Artifacts

Creates: Improvement actions

Updates: Product Brain

What good looks like

  • Blameless and evidence-aware.
  • Few actions, each with owner and follow-up.
  • Learnings are linked to the epic and the code.
  • Repeated findings are flagged for the Playbook.

Quality gate

The team agrees the actions.

Destination

JiraProduct Brain

Usually next

People and agents work from the same Product Brain. The owner stays accountable. Assigned agents prepare and check. A named person approves at the gate.

Deep dives from the AI PM Lab

Articles that explain the thinking behind this Activity.

Show us exactly how product work happens today.

Thirty minutes with a founder. We compare your process with the Playbook and pick the first Activity to run on your Jira and your repo.