Every director goes home with their own AI agent

Yesterday, on behalf of Education Oracle, I ran a workshop for directors and managers with one clear approach: don't treat AI as hype or a standalone tool, but as something you deliberately stay in control of.

We started the afternoon with a practical workshop. First we looked at the difference between prompts and AI agents, the opportunities and the risks, and the ethical frameworks you want to work within as a leader. Not as a legal story, but as a concrete design framework: what may AI support, where do you need to set limits, and how do you always remain ultimately responsible yourself?

Then we got to work. In groups by level, from learning to prompt more effectively to building and refining Copilot agents, participants worked on applications that tie in directly with their day-to-day practice. Think of support for decision-making, policy analysis, weekly overviews or structuring complex files.

The best moment came at the end of the workshop. Every participant went home with their own AI agent. Not a generic chatbot, but an agent with a clear role, clear boundaries and a concrete task.

After the workshop, Nadia, together with Marieke and the team, took over with an explanation of the plans for MondriAI-Edu, the AI code of conduct and the further roll-out within the organisation. That linked individual practical experience to an organisation-wide vision.

What stays with me most: the calm that arose once it became clear that AI doesn't have to be a threat, as long as you design it well. AI that thinks along, organises and asks questions. And people who keep making the decisions.