Services › AI agents in schools
Which work would you give to an agent?
We help your school or network choose which work an AI agent can take over, and build that agent with your people in the platform you already have.
Since February 2025 the EU AI Act requires organisations that use AI to ensure sufficient AI literacy among the people who work with it.
Start with your own work.
We work in Microsoft 365 Copilot, Google Gemini, ChatGPT or Dify, whichever your school already uses. Your people build along, so the agents stay when we leave. LeX works with professionals who know the field: developers and trainers. Guido van Dijk, director of LeX, holds final responsibility and is your point of contact.
This study coach is one of six example agents from our public guides. You can read them before you call us, and they stay with the school. Workshops in English use the Dify series. See the guides
Our director also works for Stichting VO-content, a Dutch foundation. VO-content develops open learning materials and AI for the whole sector. LeX works at the level of a single school or school board, and does not speak on behalf of VO-content. How we keep roles apart
Limits give room.
Once you know what an agent does not do, you see how much room is left. That is why we always start here.
- There is a human in the loop. The agent makes a proposal, a member of staff checks it and decides.
- Arranged first, then student data. The GDPR applies to a prompt as well. If students work with an agent themselves, it starts with a DPIA: the data protection impact assessment your data protection officer knows.
- Assessment stays with people. AI that evaluates learning outcomes or helps decide on admission is high risk under the AI Act. We build agents that support, not agents that award grades.
- No supplier, no licences, no hosting. You work in your own environment and manage it yourself. That keeps you in charge.
How to work with us
- Talk Twenty minutes for a team, a board or a network event: what an agent is, and where to start.
- Half-day Participants build an agent themselves, in Copilot or Dify, and take the session plan and design card home.
- Multi-day programme Each participant builds an agent for their own work and tests it at their school.
- Hackathon Teams build an agent for a real problem in a single day. We provide the technical lead, as at two hackathons for teachers in 2025.
- Engagement The first agents in your own environment, with a working group. If the question is about how education is organised, the work continues as education transition.
Is this for you?
- You want something that lasts. A talk or half-day is a start, with the question which work you want to change.
- You want to learn from work that already runs. A general inspiration session on AI we leave to others.
Which work would you hand over first?
Tell us. We will think with you about a good first step.
Prefer email? info@l-e-x.nl
