AI assessment for companies

The assessment ends in concrete material you can decide on, whether you build with me, with your own team or not at all.

• A map of the processes that take the most time, with estimated hours per week and who does the work today.
• A ranked list of where AI and automation pay off, with the estimated time saved and the cost to build for each item.
• A view of the risks: which data is involved, security, GDPR and where the EU AI Act applies.
• A proposed first step that can be built in a few weeks and measured.

The best opportunities are rarely the most talked about. They are usually work that repeats every day: email being sorted, data being moved between systems, documents being summarised, and reports being assembled from numbers in several places.

I judge each step by three questions: how much time it takes, how predictable it is and what happens if it goes wrong. Steps that are time-consuming, predictable and easy to check suit an agent. Steps where a mistake is expensive keep a person, sometimes with AI as support.

The result is not a report that ends up in a drawer. Every item on the list has an owner, an estimated effect and a way to measure that it works once it is built.

Many AI projects start with a solution and then look for a problem. That gets expensive, and it is why so many pilots never reach production. An assessment turns the order around: first where the time goes, then what is worth building.

Because I also build and run these solutions, the judgements stay practical. I know what an integration with a business system tends to cost, how long an agent takes to become stable and what it takes for it to still work in a year, after the models and tools have changed.