An LLM inside the corporate network
Where does the data go when an employee pastes a document into ChatGPT? What actually protects you, what is only security theater. Real options: private Azure, on-prem deployment, Claude Enterprise.
A lab for executives and senior managers of industrial mid-caps. It does not produce AI technicians — it produces leaders who ask sharper questions and recognize opportunities that were invisible before.
Every session is built around the principles of the EU AI Act. Not as a regulatory appendix — as the structure of the reasoning itself.
The people who choose vendors, validate proposals, sign contracts — and need to know what to ask, what to demand, what to refuse. Not a technical course: a shift in vantage point.
Foundations, recognizable examples, the difference between "AI" as a word and AI as a technology. No mathematics.
The three architectures everything rests on. What they do, what they do not, when to choose each. No code.
Where AI truly changes the way work gets done. Where it creates new risk. How to tell them apart at a glance.
When to trust the system, when not to. How to build control without paralyzing the work.
From the written rule to the real one: the Monday email to your team that sets what is allowed and what is not.
The core AI Act requirement made operational. Where a human must be present, how you prove they were, why it counts.
What changes over the next 24 months. What it means for your company. What your customers will start asking.
How to pass on what you learned to the rest of the function. AI literacy as a legal duty — and a competitive edge.
Three deep dives into the points where the real difference is decided: between a serious AI program and a cosmetic one.
Where does the data go when an employee pastes a document into ChatGPT? What actually protects you, what is only security theater. Real options: private Azure, on-prem deployment, Claude Enterprise.
Where does your data end up with a cloud model? When is it worth keeping everything in-house? What does an on-prem model really cost? The technical questions behind a strategic choice.
Not the legal counsel version. The operating version: what you do on Monday, what you put in the internal policy, what you demand from the vendor. What you risk if you do not.
Classes of 6 to 12 participants — the discussion only works at that size. The agenda is built around the real processes of your company, starting from a preliminary call with Francesco or Michele.
A 30-minute call with Francesco or Michele. Not a pitch — a read of your context, to decide together whether a program makes sense.