Is Your AI-Built App Safe to Launch?
This free check asks ten questions about an app, internal tool or automation you made with AI, two in each of five areas: ownership and access, reliability, release confidence, recovery and support. Each area gets its own result as soon as you answer its two questions, with what to do next. There is no sign-up, and your answers stay in your browser. It sits with our other free tools.
It is for owners and managers who built something with tools such as ChatGPT, Claude, Cursor, Lovable, Bolt or Replit, or had someone build it for them that way, and now want customers or staff to depend on it. AI gets you a working app quickly. These questions are about the things around the code that decide whether it stays working once people rely on it.
What it checks, and what it doesn’t
The check never looks at your app. It doesn’t load it, scan it or ask for its address. It does not inspect code, test security or certify the app. It reflects your answers and nothing more. Use it to find where to start, then get a developer to read the code before anyone relies on it.
The five areas
Ownership and access. Whether the business owns the code and could change every password and API key today. An app that lives only inside the tool that built it, or in a freelancer’s personal account, can be lost with a cancelled subscription or a falling-out.
Reliability. Whether real people have used it for a while, and whether anyone would hear about an error before a customer complained. Something that only its builder has tried has been tested by the one person who knows how it is meant to be used.
Release confidence. Whether there is a written description of what it should do, and whether changes get checked before they go live. With AI-built software the next change is often one prompt away, and the AI can break something it wasn’t asked to touch.
Recovery. Whether the data is backed up somewhere separate, whether a backup has ever been restored, and whether you could roll back a bad change.
Support. Whether another developer could pick the app up, and who has agreed to fix it and keep it secure once people rely on it. If the answer is “nobody yet”, that is the gap to close first. Our piece on who is responsible for software built with AI goes into why.
How the result works
Every answer is fine, partly fine or a risk. “Not sure” counts as a risk, because if nobody knows where the backups are, nobody will find them in a hurry either. Each area has two questions, so it scores from 0 to 4 points of risk:
- 0 points: in good shape
- 1 or 2 points: gaps to close
- 3 or 4 points: at risk
The overall result follows the weakest area. Any area at risk means not ready to launch yet. If nothing is at risk but some areas have gaps, it is nearly ready. If every area is in good shape, nothing stands out from your answers, which is still your answers rather than a review.
A worked example
A lettings agent’s office manager built a viewing-booking tool with an AI coding assistant. The code sits in her personal GitHub account and the API keys are in a note on her laptop, so ownership and access scores 2: gaps to close. Staff have used it for a month without serious problems, and errors only show up if someone goes looking in the hosting logs, which puts reliability at 1. There is no written description and changes go straight live, so release confidence scores 4: at risk. The hosting takes nightly backups that nobody has restored, and she isn’t sure how to roll back, so recovery scores 3. Support is “probably” and “whoever built it”, 2 more.
The result is not ready to launch yet, with release confidence and recovery the two areas to fix first. The first jobs are a one-page description of what the tool must do and a test copy where changes are tried before they go live.
Send us your answers
If you want a second opinion, press “Send me my answers for review” under your result. A developer reads them and replies by email with what to do first. Nothing is sent unless you press send, and you can see exactly what we receive before you do.
When to get a developer to sign it off
If the app touches customers, payments or personal data, answers alone aren’t enough. AI Sign-Off is a developer review of software made with AI: we read the code, check it against what it is meant to do, fix what fails and sign off the exact version reviewed. It starts from £750 + VAT for a small app or automation, and a support retainer keeps someone accountable as it changes. For a fuller do-it-yourself pass first, our guide on how to get a vibe-coded app ready for real users lists the checks an owner can run.
Frequently asked questions
Does this test my app’s security?
No. It asks about how the app is owned, run and looked after. It cannot see your code, so it can’t find a leaked key or a page that opens without a login. A developer review can, and if the stakes call for someone deliberately trying to break in, that is security testing.
Is my data stored?
No. The check runs in your browser and your answers are not sent anywhere unless you choose to send them to us for review.
What does “vibe coding” mean?
Building software by describing what you want to an AI tool and accepting what it writes, often without reading the code. Our glossary has a fuller definition of vibe coding.
My app came out “nearly ready”. Can I launch?
That is your call, and it depends on what the app touches. Gaps in an internal tool used by three staff matter less than the same gaps in something that takes customer payments. Close the gaps the result lists, and if customers or money depend on it, get the code read by a developer first.
“I’ve put everything I know into how this company works — the standards, the method, the care on every project. It runs through the whole team, and I hold us all to it.”
Alexander De Sousa · Founder LinkedIn
Featured on BBC Radio Solent
Tell us what you need
A few quick questions, then a straight answer from a real person — usually within a few hours.
What happens next
- 1 We reply A real person reads your answers and gets back to you, usually within a few hours.
- 2 Questions or a short call If anything isn't clear, we ask by email or book a short call. Before any call we audit your website and make the quick fixes it finds, free.
- 3 Your fixed quote A clear, fixed price in writing, within 48 hours of any call. There is no obligation to go ahead.
Tell us what you're working on
Whether it's a new site, a platform, or a process that shouldn't be manual any more — we'll tell you honestly if we can help.