Sprint Retrospective Prompt Pack

AI prompts for retro facilitators — What Went Well, deltas, actions

Ad placeholder (leaderboard)

A sprint retrospective is the agile ritual where a team inspects how it worked and commits to improvements. Done well it compounds — small fixes accumulate into a faster, happier team. Done poorly it becomes a complaint session with no follow-through. This prompt pack gives facilitators a structured way to run the session and, crucially, to turn the raw feedback into prioritised, owned action items afterwards.

How it works

You set the team size and sprint length so the prompts match your group dynamics and cadence, then choose a format: classic (What Went Well / Improve / Actions), starfish (Keep / More / Less / Start / Stop), or 4Ls (Liked / Learned / Lacked / Longed For). The pack produces two prompts. The first is a facilitation guide with timeboxed sections, opening questions, and prompts to draw out quieter voices. The second is a synthesis prompt: you paste in the raw feedback and the model clusters it into themes, surfaces the highest-leverage issues, and drafts SMART action items with suggested owners. Everything is generated in your browser.

Tips and examples

Run the facilitation prompt before the meeting to set your timeboxes and opening question, and keep the conversation human — the AI never reads the room for you. Capture feedback verbatim during the session, then drop it into the synthesis prompt while it is fresh; clustering is far more accurate with real quotes than with your paraphrase. Cap your action items at two or three — a retro that produces ten “improvements” produces zero, because none get done. Assign every action an owner and a definition of done, and open the next retro by reviewing whether the last batch shipped, which is the single biggest driver of teams actually improving.

Ad placeholder (rectangle)