Step 1 / 5 · 20 min
Make the prototype answer a product question, not demonstrate build speed.Choose one uncertainty
- Write the riskiest interaction or value assumption.
- Book test users before expanding scope.
Build · Core skill lab
48-Hour Workflow Slice
Build only enough of the riskiest interaction to watch a real user succeed, fail, or change your mind.
- Decision
- Does this workflow create enough user pull to justify production engineering?
- Timebox
- Two days + three observed sessions
- Output
- Prototype learning record
Bring to the bench
- One riskiest product assumption
- Three target users booked before polishing
- A disposable build surface
Your work saves in this browser.
Field tools
Use the instrument, not a blank page.
Copy these into your interview, agent, review, or working document. They are specific to this repetition.
script
Prototype test script
Read only the setup. Do not teach the interface.
SETUP: “Imagine [real trigger]. Use this prototype to [real outcome]. Think aloud, but I may stay quiet.” PROMPTS ONLY IF NEEDED: “What are you looking for?” “What do you expect that to do?” “What would you do next?” “What makes you trust or distrust this?” END: “What would you use instead tomorrow?” “Show me how that differs.”
checklist
Fidelity budget
Spend fidelity on the assumption, not the chrome.
Must be real: interaction under test; information needed for trust; meaningful error/recovery; timing if latency matters. Can be simulated: authentication; admin; rare integrations; notifications; data plumbing; visual polish unrelated to comprehension.
Calibrate judgment
Compare the evidence, not the polish.
Learning before architecture
A two-day cited-brief prototype is tested with three users.
Two users completed in under 20 minutes, but all three needed narration to distinguish contract citations from meeting-note citations. Reshape that trust interaction and test again; do not productionise connectors yet.
Why it works: Observed friction changes the design and protects engineering from premature commitment.
Prototype as mini-product
Built auth, dashboard, settings, integrations, and a polished landing page. Demo viewers said it looked great, so we will now harden the stack.
Why it fails: Scope hides the risky interaction, feedback is performative, and no observed user behaviour supports production investment.
Review → revise → repeat
The artifact is the beginning of the rep.
Check only standards your current artifact actually meets. Then record one consequential revision before exporting it.