Sprint 3 · Due · 10 points
Review is the phase most students skip. It's uncomfortable — you just built something and now you have to honestly assess it. But a review that names what broke is worth more than a demo that hides what doesn't work.
— Prof. Sathya
Three parts. 200 words max (not counting the AI critique paste).
Try your build yourself as if you were your user. Describe what happens. Where does it work? Where does it break or feel wrong?
Copy-paste this prompt into Claude or ChatGPT:
I built [describe what you built] for [your specific user] during [the specific moment you're designing for]. Here's what it does: [brief description]. Here's what I think could go wrong: [your failure mode from the Build Plan]. Before critiquing, research online from credible sources to understand the scenario, users, and processes I am designing for. Use what you find to check my design against reality, not just general design principles. Then challenge this design: What assumptions am I making about my user that might be wrong? What would happen if the constraint was worse than I expected? Where is this most likely to fail in real use? Be specific and direct — ground your critique in what you found.
Paste the AI's critique, then write: which critique actually landed — the one that made you think "that's a real problem" — and what you're going to do about it in v2.
Based on your self-test and the AI challenge, name one specific thing you're changing in v2 and why.
Put yourself in the physical situation. If your user is standing in a hallway, stand up. If they have 3 seconds, time yourself. If they're wearing gloves, try using your phone with gloves on.
Don't just agree with everything the AI says. Some critiques will be generic ("what if the user doesn't have internet?"). Pick the one that's specific to YOUR design and YOUR user.
Your self-test notes + the AI critique (copy-pasted) + your response identifying which critique landed + your one change for v2. Can be submitted as text in Canvas or as a document.