AI and Automation

I bring AI where it actually helps: cutting the tedious work, surfacing useful insights, making your software smarter. Without the hype.

Not every problem needs AI. Often a plain rule or a clean connection between two systems does the job, and then I build that instead, because it costs less and is easier to trust. Where AI earns its place, I start with a single task, and you see how it does on your own cases before it goes further.

Your team stops doing work that a computer should be doing

Good candidates are tasks that come up often, follow clear rules and are still done by hand: moving data between systems, sorting documents, preparing replies. The more often a task recurs, the more it is worth automating. One-off exceptions are better left to people.

What goes is the work nobody enjoys that eats time anyway. Decisions and contact with people stay with your team. Whoever used to retype data gets to do something more useful.

AI assistants and smart workflows built for your specific context

A general model does not know your prices, your processes or your customers. It becomes useful once it can reach your own data and knows what it is allowed to do with it. I build that connection, not the model itself.

Manuals, price lists and earlier proposals are made searchable, and the model answers from them rather than from memory. Every answer can be traced back to its source, so you can check where it came from.

Connects only to what it needs to. Your data stays in your control.

Your input is not used for training. I use access tiers where the provider does not train on it, and I put in writing what goes where. Where it makes sense, the model runs on your own infrastructure and nothing leaves your building.

AI will sometimes get things wrong, so I do not build systems that decide on their own. Suggestions go to a person, and critical steps need approval. What gets automated is the preparation. The responsibility stays with your people.