Connect your tools
A distribution SME had built a field application for its sales team with Lovable. The items, the customers and the sales documents stayed in Sage 100.
Read the full caseBack to the case studiesFix & stabilise
Picking up a generated application the AI could no longer fix
Every fix asked of the tool broke something else. What we took on in the code, and what the team is able to ship today.
- Sector
- Startup, online product
- Built with
- A code generation tool
- Engagement
- Code taken over, then ongoing support
- Way of working
- Delivery in batches approved by the team
The context
A startup had its product generated, then kept adding to it through the tool’s chat, without ever opening the code. The product worked and served real users.
The challenge
Bugs piled up with no obvious place to start. Every fix asked of the AI broke something else, and the back-and-forth burned credits without getting any closer to a working release. Nobody on the team could say what state the code was in, or which areas were risky to touch.
What we did
We took the code on, isolated the fragile areas and fixed what was blocking. We also made it legible: comments, documentation, written pointers for the AI tools that keep working on it. Then we supported the team over time.
The outcome
The team ships features again without breaking what works, and we now support them on an ongoing basis.
What this case shows
Picking up generated code does not mean rewriting it. It means dealing with the fragile areas, and leaving behind a project that an AI tool, like a developer, can take on without breaking everything.
The other cases
Scale up
The founder of an AI startup had built their product alone, with a generation tool. It worked, and they wanted to take it to the next stage.
Read the full caseTell us what is blocking you
Describe your situation in a few lines. We come back to you with a fixed quote within 48 working hours.
Request a quote