Short answer
AI training that makes a difference is not about which tools exist. It is about the participants’ own work — and at least half the time should be spent with hands on the keyboard, not listening.
The measure of whether it succeeded is simple: how many people in the room have used it in their real work three weeks later? If the answer is zero, it does not matter how good the talk was.
Why most training achieves nothing
The most common format is an hour’s overview of what AI is, a few impressive examples, and a closing remark that the technology changes everything. Everyone thought it was interesting. Nobody does anything differently on Monday.
Three causes recur:
The content was about the technology, not the work. Participants learned what a language model is. They were not told what to do with their own Tuesday.
Nobody wrote anything themselves. Watching someone else use a tool gives you as much skill as watching someone else swim.
Nobody decided what would happen afterwards. Without a task to take back, everything reverts to the familiar within a week.
What a half-day should contain
This is how I lay out the four hours, and the proportions matter more than the order.
1. Thirty minutes: what the tool actually is. Short and demystified. That it produces what is probable rather than what is true, why it invents sources, and why it sounds just as confident when it is wrong. This section exists to calibrate trust — both in the person who thinks it is magic and in the one who thinks it is worthless.
2. Thirty minutes: what applies with us. Which tools are approved, what must never be entered, who is responsible for what gets published. If you have no policy yet, this is often where one begins.
3. Ninety minutes: your own tasks. The core. Each participant brings three real tasks — an email to be written, a summary to be made, a document to be condensed — and we work on them on site. Not examples, not invented cases. What is in their inbox.
4. Forty-five minutes: what gets worse. Where it should not be used in your particular setting, how to recognise an answer that looks right and is wrong, and which tasks should remain human. This section is what separates a training course from a sales pitch.
5. Fifteen minutes: one thing to do on Monday. Each participant writes down one concrete task they will use it for within a week, and who they will tell that they have done it. It sounds trivial, and it is the single most important quarter of an hour.
Who should be there
Mix roles. It looks messy on paper and works better in practice, because what is obvious to an administrator is often unknown to a manager — and the other way round.
Include the sceptic. That person asks the questions that keep the training honest, and often becomes the one who uses the tools most thoughtfully afterwards.
And make sure someone with authority is in the room. Without that, most ideas stay in “we should raise that some time”.
What happens afterwards
The effect is decided by the two weeks after, not by the day itself. Three things usually need to be in place:
Someone who answers questions. A person inside the organisation, not a consultant. It is enough for them to be half a step ahead of the others.
A place to share what has worked. A channel, a document, an item at the weekly meeting. What one person solves is usually applicable to five.
A follow-up booked before everyone goes home. An hour about a month later. Without it, the training becomes a pleasant afternoon.
What it costs
A half-day training costs SEK 8,000 and is adapted to what you actually do during the day. I always ask for three typical tasks in advance, and build the session around them.
If the broader work needs doing first, there is AI strategy from SEK 15,000, which answers what is worth doing and in what order, and policy, GDPR and the EU AI Act for SEK 12,000. What the different parts involve is set out in the article on what an AI consultant costs.
The most important thing
If you only take one thing away: measure the training by what is done differently three weeks later.
A training where everyone thought it was interesting but nobody changed anything is not a successful training. It is a pleasant afternoon with an invoice.
Best wishes, Mikael GlännströmWellbeing coach, author and father of two