Product critic / Patchwork
Don't add an AI dashboard.
Finish the release handoff.
The decision: what should one developer build this week so a small team can prepare a release without maintaining a second spreadsheet?
Evidence boundary
- The tracker stores release, owner, status and severity.
- The founder currently copies open blockers into a spreadsheet before release meetings.
- There is one development week available and no approved ad budget.
- Whether other teams share this workaround.
- Whether an export is worth paying for.
- Whether the real bottleneck is missing data, rather than viewing it.
A release-filtered view with a copyable blocker list.
Reuse existing data. Show unassigned issues explicitly; copying a list must not imply that missing ownership is resolved. Defer forecasting, integrations and generative scoring.
One next test
Use the prototype in the founder's next real release preparation. Observe whether they can build the meeting agenda from it without recreating the spreadsheet. Ask permission before observing another team; no outreach is authorized by this recommendation.
Reconsideration rule
If the founder still needs the sheet because status or ownership is stale, stop expanding export features and examine how updates enter the tracker. If another consenting team tries it, treat that as qualitative evidence, not market validation.
What would change the decision?
Repeated examples of decisions blocked by issue volume rather than stale information would strengthen the case for better triage assistance. It would still not establish that a model-generated score is trustworthy.