/product-bet
Turn a promising idea into a bounded, testable investment decision.
For product managers, founders, and product engineers deciding what to build next.
npx skills add matthieu-pm/product-bet“We should add AI-generated weekly summaries.”
Who needs this? What do we know? How much should we invest?Know what you know.
Before building summaries, check whether managers need them and whether the available project data can make them useful.
Managers spend meaningful time assembling weekly status.
Tasks and comments contain enough context for a useful summary.
Current preparation time and the cost of missing an update.
Give the idea some competition.
Compare different ways to solve the problem, including doing nothing.
AI-assisted concierge summarySelected for test
Learn whether a reviewed summary helps before automating it.
Rules-based activity digest
Aggregate changes with less interpretation; may still leave too much noise.
Owner-written updates
Ask project owners for structured updates; adds recurring work for the team.
Keep the current workflow
No new investment, but no deliberate test of the idea’s value.
Make the investment finite.
Define the smallest intervention that can deliver value and test the idea.
- Owner
- 1 PM / design engineer
- Setup
- 1 week
- Participants
- 5 opted-in managers
- Duration
- 2 weekly cycles
Human review and source links for every claim. Only opted-in projects; no automatic sending.
Production integration, cross-workspace data, and automated recommendations.
Decide what each result means.
Agree on the next move before running the test. Establish the preparation baseline during screening.
Test automation feasibility.
Managers use the summaries and describe a concrete planning benefit.
Revisit the problem.
Summaries go unused or duplicate an existing workflow.
Stop the test.
Sensitive information is exposed or an error causes a consequential wrong decision.
Repair the test.
Delivery timing or missing source access prevents evaluation.
Investigate the uncertainty.
Interview participants and revise the target situation before another test.
Confidence has to be earned.
This bet stays Exploratory. Defining a test doesn’t validate the problem or its proposed solution.
Exploratory Current
The decision, target user, outcome, and assumptions are explicit.
Plausible
Needs direct evidence of the problem and support for the proposed mechanism.
Ready to test
Also requires a bounded intervention, appetite, test method, guardrails, and decision rules.
Ready to commit
Needs supporting validation, clear ownership, and implementable measurement.
npx skills add matthieu-pm/product-bet