Where we use AI in projects and where we don't

Artificial intelligence proves useful for specific tasks within the projects we work on, but we don't use it as a selling point or add it where it doesn't bring anything to the table.

We get asked quite a lot whether we use artificial intelligence in what we do. The short answer is yes, but not as a sales headline, rather as a tool within specific software development tasks.

Where we use it

On custom development projects, AI helps us with very specific things: generating repetitive code, spotting error patterns, or proposing initial structures that we then review and adjust ourselves. We also use it to speed up testing and internal documentation, which tends to be the most tedious part of any project.

It never replaces the work of designing an application's architecture, deciding how a system should behave, or understanding what a client actually needs. That's still our job, based on our own judgement.

Where we don't use it

We don't add "an AI chatbot" to a project just because it sounds good in a proposal. If a client needs to automate responses or classify information, we assess whether it makes technical and financial sense for their specific case. Often the most sensible solution is simpler than putting a language model in the middle of it.

We also don't blindly trust whatever an AI generates. All the code and content that comes out of these tools goes through human review before reaching production. AI proposes, but it doesn't decide.

Why we're being upfront about this

We'd rather be clear about this because there's a lot of noise around AI in the software industry. It's presented as if it could solve any problem, and that's just not true. It's one more tool, useful in certain contexts, and you need to know when to use it and when not to.

If you're weighing up a project and wondering whether AI should be part of it, that's a conversation worth having from the start, with specific cases on the table rather than general assumptions. That way you can decide, with a clear head, whether it genuinely adds something to what you need or whether it's better solved another way.