Sprint Retrospective AI Prompt

A structured prompt for using AI to prepare a retrospective for an AI-assisted development team. Generates data-informed conversation starters, format recommendations, and action item candidates without replacing the human facilitation.

Syntax

scrum-master-ai

Sprint [number] Retrospective Preparation

Context:
- Velocity this sprint: [X points]
- Escaped defects: [n]
- Stories with significant AI contribution: [n of total]
- Previous retrospective top action items: [list]
- Team size and remote/in-person: [details]

Please generate:
1. Three data-backed conversation starters based on the above metrics
2. Two questions about AI tool quality patterns we should discuss
3. A retrospective format recommendation for this team context
4. Five candidate action items based on patterns you can infer
5. One "elephant in the room" question we are probably avoiding

Example

scrum-master-ai

// Sample output for a team with 3 escaped defects and 60% AI-contributed code:

conversationStarters: [
  "We shipped 3 defects that passed our DoD — what did our review process miss?",
  "60% of our stories had significant AI contribution. How confident do we feel about what we shipped?",
  "Our previous retro action was to improve PR review quality. What changed?"
]

aiQualityQuestions: [
  "When AI wrote the code and the tests, how did we verify the tests were actually testing meaningful behavior?",
  "Did anyone approve AI-generated code this sprint that they didn't fully understand? What would it take for that to be safe to say?"
]

recommendedFormat: "Start/Stop/Continue — straightforward, time-efficient, surfaces specific behaviors"

elephantInTheRoom: "Is our Definition of Done actually being applied, or are we calling things done based on test passage?"