The Prototype Brief

The prototype carries the solution. It is also the least valuable of the three artifacts, and that is by design. Its job is to align the team, then to be rebuilt properly. Do not fall in love with the prototype. Fall in love with the problem.

Paste the block below into Claude Code, Cursor, v0, Bolt, or whatever you build in. Fill the brackets. The one-page brief is your context; paste it in whole.

Build a working prototype that lets a user [core action].

Context: [paste the entire one-page brief here]

Constraints:
- Real interactions, not screenshots. It must be clickable.
- Cover the happy path end to end, plus the two most likely failure paths.
- Do NOT build settings, auth, or edge cases yet. Align on the core flow first.
- Mock the backend and fake the data. We are testing the flow, not the plumbing.

Success looks like: a teammate can use it without me narrating over their
shoulder, and can tell me where the flow breaks.

How to brief a prototype that aligns instead of just demos

• Describe the experience, not the interface. Not "build a dashboard with charts." Instead: "A user opens their account list, clicks an account, and instantly sees a red / yellow / green risk score." Tell it the moment of value.
• Name real scenarios. Real names, real numbers, real edge cases from an actual customer. Lorem ipsum hides the problems you are trying to surface.
• Say what you will NOT build. Scope is a gift to the model and to yourself.
• Happy path first. Get the core flow clickable before anything else exists.

Rules for this artifact

1. The prototype is disposable. The moment it has done its aligning job, its remaining value is close to zero. Keep the problem understanding. Throw the prototype away.
2. Pair one hour building with one hour testing. A prototype nobody tries is just a fancier mockup.
3. Two weeks, maximum. If you are still iterating at week three, build the real thing or kill the idea. Go back to customer conversations, not more prototype tweaks.
4. When it aligns the team, extract the commitments into evals (see the eval template) and hand the real build to engineering.