AI training for staff: how to choose what each department needs

The problem with one course for everyone

An accountant, a salesperson and a warehouse manager use AI for very different tasks and so make different mistakes. A general course covers a bit of everything, so one group is bored and the other doesn't learn what it needs. Article 4 of the AI Act also asks that measures take account of staff knowledge, experience and context.

Start with a short measurement

First have staff complete a few work tasks from their own field. The result per department shows where the real gaps are. Perhaps sales gives AI good instructions but doesn't check promises in quotes. Perhaps accounting checks figures well, but nobody knows which data may go into which tool.

Choose short, targeted lessons

Link one or two lessons per department to what the measurement showed. Short lessons on a concrete skill fit a working week better than half a day of theory, and staff see the point sooner. Everyone also gets the basic rules for safe use.

Show that it worked

Measure again after a few weeks, with different tasks at the same level. The difference shows management whether the training paid off, and it belongs in your Article 4 record.

Also check whether your training costs qualify for regional support, such as the SME portfolio (kmo-portefeuille) in Flanders or training vouchers (chèques-formation) in Wallonia. Conditions differ per scheme and provider.