Blog · Security
Built your app with AI? Ten things to check before real customers arrive
AI coding tools can take an idea to a working app in a weekend. That's genuinely useful. But "it works" and "it's safe for customer data" are different things, and AI tools tend to optimise for the first.
Before real customers sign up, run through these ten checks.
Access
1. Every page and API checks who's asking. Try opening another user's data by changing an ID in the address bar. If it works, so can anyone else.
2. Admin features are actually locked. Hidden buttons aren't security. The server has to refuse the request.
3. Your database isn't open to the internet. Many hosted databases have row-level security settings that are off by default. Check yours.
Secrets
4. No keys in the code. Search your repository for API keys and passwords, including old commits.
5. Keys in the browser are public. Anything shipped to the front end can be read by visitors. Only publishable keys belong there.
Data
6. Inputs are checked on the server. Forms validated only in the browser can be bypassed in seconds.
7. You know where personal data lives. Make a short list: what you collect, where it's stored, and who can see it.
8. Backups exist and you've restored one. A backup you've never tested is a hope, not a plan.
Running it
9. Spend limits on every paid service. Hosting, AI and email providers can all run up bills if something loops or gets abused.
10. You'll know when it breaks. Set up basic uptime alerts and error logging, so you hear about problems before customers do.
Fixing what you find
Most of these take minutes once you know about them. Write each issue as a clear, specific instruction and your AI coding tool can usually apply the fix. Then test it yourself rather than taking its word for it.
Want a second pair of eyes on how your firm uses AI? Book a free 20-minute call.