Retrospective: Activation sprint 14 (25 Aug to 5 Sep)
Outcome against goal
External collaborators can be invited and activated in staging. Event tracking slipped to next sprint.
Delivery facts
- Planned: 12. Completed: 9. Carried over: 3. Added after start: 2.
What went well
- Error states shipped two days early after design and engineering paired on day one (ACT-132).
- Review turnaround averaged under one day.
Where work waited
1. ACT-141 sat blocked for 4 days waiting on the email provider's support reply.
2. ACT-146 moved back from Review to In Progress twice; tracking events were not defined up front.
3. ACT-150 was added mid-sprint and took 2 days of QA capacity.
Patterns
- Two of three delays came from requirements that were not written on the card before work started.
Experiments for next sprint
1. No card enters In Progress without acceptance criteria.
2. Open third-party support tickets on day one for any external dependency.
3. Mid-sprint additions must name what they replace.
Questions for the team
- Would a short design-engineering pairing slot help every sprint?
- Was ACT-150 genuinely urgent?Claude prompt for project retrospective
Run retros on what actually happened: what shipped, what slipped, where work waited, and which patterns repeated.
Best for: Sprint retros, end-of-project reviews, and post-launch lessons learned.
Works with: Any assistant connected to Kanera, such as Claude, ChatGPT, GitHub Copilot, Cursor, Codex, OpenCode, or Antigravity.
You are helping me prepare a retrospective. You are connected to Kanera through MCP. This is read-only.
Board or project: [NAME]
Period: [SPRINT DATES OR PROJECT START TO END]
Original goal or scope: [GOAL]
Retro format: [START/STOP/CONTINUE, WHAT WENT WELL/WHAT DIDN'T, OR "RECOMMEND ONE"]
Before writing:
1. Read the cards completed during the period and those still open that were planned for it.
2. Read card history for due date changes, list moves backwards, and periods marked blocked.
3. Note time in progress for completed cards and flag the longest ones.
4. Read comments on the most delayed cards to understand why they waited.
5. Look for repeated causes: unclear requirements, waiting on approvals, late dependencies, scope added mid-period.
Prepare retro material with:
- Outcome against goal: what shipped, what did not, in two or three sentences.
- Delivery facts: cards planned, completed, carried over, and added after the start.
- What went well, with evidence.
- Where work waited: the three longest delays and their recorded cause.
- Patterns: causes that appeared on more than one card.
- Three improvement experiments, each specific enough to try next period.
- Questions for the team to discuss.
Describe process, not people. Keep names out of findings. Cite card keys. Label inferences clearly.Worked example output
Example result after Claude reads a two-week product sprint and its card history.
Get a more reliable result
Share the delivery facts before the meeting so the discussion starts from the same numbers.
Pick one or two experiments, not all three, and record them on next sprint's board.
Run the same prompt next time and ask Claude whether last retro's experiments changed anything.
Prompt questions
How is this different from the project wrap-up prompt?
The wrap-up prompt writes a client-facing close-out. The retrospective is internal: it looks for delivery patterns and lessons for the team.
Will it blame individuals?
No. The prompt asks Claude to describe process patterns, not people, and to keep names out of the findings.
What data does it use?
Completion history, due date changes, time in progress, blocked periods, and comments on cards from the period you choose.
Go from live Kanera work to a finished update.
Connect your AI assistant to Kanera so retrospectives start from the real delivery history instead of whoever remembers most.
Explore Kanera's AI workflow30-day Pro trial. No card required.