We teach this on your own work.
We turn the people you already employ into the people who run your AI, so there is no AI team to hire. Up to 15 seats, and every programme starts with leadership.
Every programme starts with a north-star workshop.
The first session is with your leadership. We agree what AI is for in your company, which decisions change because of it, and which ones nobody is allowed to touch.
The rule exists because a programme can only be tailored once someone has set the direction. Skip it and we teach generic tooling to people who go back to work that has not changed.
Half day, full day, two days.
-
Half day
EUR 4,000 ex VAT
One subject, one group. Enough for a north-star session with leadership, or one skill taught properly.
-
Full day
EUR 7,000 ex VAT
Two or three subjects, with hands-on work on your own cases after lunch.
-
Two days
EUR 10,000 ex VAT
A programme. Leadership on day one, teams on day two, and something built and running by the end.
All prices ex VAT, up to 15 seats. Travel within Belgium is included and you provide the venue. Accommodation sits outside the two-day package.
We pick from this list with you.
- Where your company runs its inference. Open against local, which are two questions and not one. What a machine you can order today actually runs. When buying beats renting the API, and what renting costs you beyond the invoice. Whether local makes you sovereign, which it does not on its own.
- The hub-and-spoke change model: who learns first, and how the rest of the company follows them.
- Coding agents in the terminal, for people who do not write code. The tools that let a non-developer build a working thing in an afternoon.
- Meeting notetakers. Which one fits, where the audio is processed, when a voice print becomes biometric data, and the recording policy to write before you buy licences.
- The tools radar and the model supply chain. How we pick a tool and how we drop one. Also why a model is the one piece of software you cannot take apart to see what it does.
- The AI maturity model, L0 to L5, and the jump from L2 to L3 that costs the most.
- Reverse prompting for executives: getting the model to interview you before it answers.
- Deciding when AI is the answer. Most problems wait on better people or better software. The honest advice is often to use AI to write the software that means you need less of it.
- Sorting a process into software, agent, and human with the evidence prepared.
- Building a company brain in a morning.
- EU AI Act basics for operators: risk classes, transparency, human oversight.
- Governance and evals. An eval is the written definition of what wrong means, and it belongs in place before a system gets any autonomy. It sounds technical. It is codified common sense and it usually fits on two pages.
Taught by the people who build these systems.
Every session is taught by people who build and run these systems at Machinegeist. Today that means Christophe Rosseel and Laurens Van Acker. As the firm grows it means the engineers who join them.
What does not change is the promise. A hard question in the room gets its answer in the room, rather than a follow-up mail a week later.
Tell us who is in the room and what they are stuck on.