Skip to content

AI for companies of 50 to 250 people.

Rules first · who can see what · fixed terms

More users, more rules, more people who need to be kept in the loop. I take care of it, in steps your IT team can check.

Book a call

Feel free to bring someone from IT.

  • 01

    Rules first

    Data processing agreement (DPA), risk class and working rules come before the first line of code.

  • 02

    Multiple users, each with their own role

    Who sees what is settled up front. And who did what is logged, so you can trace it later.

  • 03

    Fixed terms

    A fixed price per step. A standing weekly check-in during the build. Documentation with every handover.

What changes compared with a small company

  • One point of contact: me. Your IT team and management sit in where needed.

  • A pilot (a first working version for one process) in one department. Only then does it go wider, built on what the pilot taught.

  • Permissions and roles per user, set up on your own login.

  • Data processing agreement (DPA) before the build, risk class in the handover.

  • Everything is transferable: the code, documentation and logins are yours.

Properly arranged, in writing too.

For anyone who first wants to know how privacy and the rules work. Four agreements, in plain language.

  • EU AI Act

    For each system, I determine the risk class. Most SME work falls into the lowest. If a chatbot talks to customers, it says so.

    Which risk class are you in?

  • Data processing agreement

    If the process involves personal data, you sign a data processing agreement (DPA) before the build. One with me, and one with the AI vendor. It sets out who may do what, and for how long.

    Part of the build plan

  • Where your data sits

    EU servers are the starting point. If a tool cannot offer that, I tell you up front and you decide. And in every case, an agreement that your data does not train anyone else's AI.

    How I handle it for the chatbot

From first call to working system

Clear steps. You always know where you stand.

Step 1 · you pick a time · AI Scan

Call

I walk through your week with you. After that you can stop. No invoice.

How you pay

No invoice for the call.

What you do

You pick a time and describe how your week looks.

What you get

  • Whether AI pays off for you now, and if not, why not
  • What I would not do
  • Which process comes first

Questions from larger teams

You work alone. Does that work for a company our size?

Yes. I build in steps, at a fixed price, and your IT team checks each one. If the project grows beyond what I can handle, I tell you in advance.

What about our existing systems?

I build on what you already have. Microsoft 365, Google Workspace or your industry software: the integration comes first.

Who is responsible for the data?

You remain the data controller. While I build or maintain the system, I am the data processor. That goes in writing, including what the AI is allowed to see.

What if the system makes a mistake?

Where a mistake is costly, I build in a check: a person approves before anything goes out. What can run without a check is set out in the working rules. If you spot a mistake, I adjust the system.

A call first. Then a first version in one department, for a fixed fee.

Feel free to bring someone from IT.

Book a call