AI training for executives: getting the leadership team to decide

AI training for executives serves one purpose: enabling the executive committee (Comex) or management committee (Codir) to decide. Without it, leadership decides on budget, data and schedule without having seen AI work, and projects stay at the test stage. A good session starts from your own use cases and ends with a decision.

What problem does executive AI training solve?

Executive AI training solves a decision problem: leadership has to make calls on AI without having seen it work in its own company. Four symptoms come up with the IT directors we meet:

• Leadership deciding blind. Budget, data to connect, project organisation, schedule with the works council (CSE, the body that represents employees in French companies): these choices rest on impressions rather than evidence.
• Tests that never scale. A department tests a tool, the results are good, but nobody on the committee carries the next step.
• Executives at very different levels. One uses AI every day, another has never opened it. The discussion stays on principles.
• Usage already in place, with no framework. Some teams use AI with personal accounts: this is shadow AI.

The IT director then ends up alone, pushing a topic that needs decisions from executive management.

Why start with leadership?

You should start with leadership because it makes the decisions that block everything else. Once the committee has seen what AI does on its own cases, budget, sponsor and schedule follow much faster.

Executive training is step 1 of our method, awareness, meaning discovering AI through your own cases.

It also meets a regulatory obligation. Article 4 of the EU AI Act (Regulation (EU) 2024/1689) covers companies that use AI systems. It asks them to take measures to develop their staff's AI literacy. Since Regulation (EU) 2026/1744, in force since 27 July 2026, it is an obligation of means: the company must be able to show the actions it has taken, without guaranteeing a specific level for each person. A documented session for leadership is part of that.

What makes an executive session end in a decision?

An executive session ends in a decision when it meets three conditions: it starts from your use cases, it shows AI working on one of them, and it sets aside time to decide.

1. Your cases, not generic examples. Around ten repetitive tasks gathered from business departments before the session. Executives picture their own business, not another sector's.
2. Evidence on your data. An agent is an AI assistant connected to the company's tools and data. It is built in front of leadership on one of these cases. This is the moment that most often triggers a decision in our sessions.
3. A decision on the agenda. Cases selected, a named sponsor, a dated next step. Without that time planned, the session remains a lecture.

To discuss it, leadership needs a shared language. At Idoption, we distinguish four levels of AI integration: using conversational assistants (ChatGPT, Claude, Copilot), connected data, advanced agents, then a chain of agents automating a full process. Agentic AI is not always the right answer: an assistant connected to the right documents is often enough for first gains.

What happens when executives see an agent built on their own case?

Four reactions come up in the executive sessions run by Idoption:

• The potential clicks. Executives see what is feasible in the short term, on their data.
• Precise questions about risks. The conversation moves from principles to real questions: which data, which access, which approval.
• Sceptics change their minds. One or two executives who were reserved at the start become keen when they see their own case working.
• A request for a pilot. Leadership wants to test several cases with real users.

A management committee in the food industry

At a food-industry manufacturer, Idoption ran a session for the management committee. Leadership launched a pilot right after.

The pilot lasted four weeks. It covered about twice as many use cases and users as our usual pilots, and produced around twenty operational agents. At the final review, we attributed this level of adoption to leadership involvement and team motivation.

Takeaway: a leadership team that has seen AI work on its own cases carries the pilot, and teams follow.

How do you prepare your leadership team?

To prepare your leadership team, three things matter more than the session content: a sponsor, real cases and clear rules on data. Who owns what:

Which mistakes should you avoid?

• A session without real cases. Generic examples don't let leadership picture itself. Gather the cases beforehand, even imperfect ones.
• A session without a decision. Put time on the agenda to decide: cases selected, sponsor, dated next step.
• Forgetting the rules. The session raises questions about data and access. They are then handled in an AI policy and, in a company of 50 employees or more, with the CSE: see our article Works council (CSE) and AI.
• Ignoring existing usage. Better to know before the session who already uses AI, and with which tools.

When is executive training not the right starting point?

Executive training is not the right starting point when leadership already uses AI daily and has grasped what is at stake. It then needs an audit of usage and tools, or a pilot on its priority cases, rather than an awareness session.

And after the session?

After the session come governance, the audit, the use-case diagnostic, then a first pilot on your data, described in our method and illustrated by our client cases. Your teams stay in control of the agents they build.

To choose the right format for your leadership team and build it on your use cases, contact us.

Sources: Regulation (EU) 2024/1689 (AI Act) · Regulation (EU) 2026/1744 (accessed 28 October 2026).