What goes wrong
What goes wrong in AI-built apps.
AI coding tools are good at making things work. They are less good at making things safe, and they don't tell you which is which. These are the nine problems I check in every review. Each page says what the problem is, what it looks like from your side, what it costs, and how to check for it yourself in about 30 seconds.
None of this means your app has the problem. It means you can find out.
- Secrets in the browserYour private keys are inside the app, where anyone can copy them.
- Users who can see each other’s dataOne user can see or change another user’s data.
- Payments taken on trustYour app believes anything that says it’s from Stripe.
- Forms that accept anythingYour app accepts whatever it’s sent, including things that break it.
- Silent failuresErrors are swallowed, so you find out from a customer.
- Slows down as you growFine with ten users. Timeouts with a thousand.
- Building blocks nobody checkedCode libraries the AI made up, or ones nobody maintains.
- No testsNothing tells you something broke until someone clicks on it.
- AI with no guardrailsEvery session guesses your rules again, and the code drifts apart.
Find out before your users do.
The Production Readiness Review is $250, takes 2–3 business days, and ends with a clear verdict: fix, restructure, or rebuild.
Book the $250 review