How do you choose between generic AI training and training on real tasks?

Choose by the outcome you need. Generic AI training builds awareness: what the tools are, what they can do, what the risks look like. Task-based training changes behavior: a specific team does specific recurring work differently after the session. If your people have never touched AI, a generic introduction has a place. If your company already pays for Copilot, Gemini or ChatGPT and daily work still starts the old way, another generic session will not move anything.

What each option actually delivers

Generic training delivers a shared vocabulary, tool orientation and inspiration. It is cheap per head, easy to schedule for large groups, and its effect is honest but shallow: people know more and mostly work the same.

Task-based training starts from the team's own reports, emails, research, meeting notes and procedures. Participants rebuild their real tasks with AI support during the session, under data rules, with a verification step. It costs more to prepare, works best in smaller groups, and leaves behind reusable examples, shared prompts and practices a manager can observe a month later.

The decision checklist

Answer five questions before buying:

1. Do people already have tool access? If yes, awareness is not the gap; practice is.
2. Can you name the tasks that should change? If nobody can list five recurring tasks per team, run a short diagnostic first, not a bigger training.
3. Who will attend? Mixed audiences of 80 push you toward generic content; intact teams of 10 to 15 make task-based work possible.
4. What must exist afterwards? If the answer includes documentation, rules or evidence for clients and EU AI Act Article 4 purposes, generic training cannot produce it.
5. How will you know it worked? Attendance and satisfaction scores measure the event. Changed tasks measure adoption.

The sequencing that works

The two formats are not enemies; the order matters. A short generic session can level vocabulary across a company. But budget the serious money for intact teams working on their own tasks, and measure before and after. Companies that spend the whole budget on a large inspirational event usually return a year later with the same problem and less internal credibility.

For a deeper comparison of prompt-focused courses versus adoption programs, see The Unlearning School vs prompt training.

Frequently asked questions

Is generic AI training ever the right choice?

Yes: for genuine beginners, for very large audiences that need common vocabulary fast, or as a mandated awareness layer. Just be honest about what it produces: awareness, not adoption.

Why is task-based training more expensive?

Preparation. Someone has to collect the team's real tasks, select the ones worth rebuilding, prepare examples with the right data boundaries and design exercises around them. That work happens before the session and is most of the value.

How many people can attend a task-based session?

It works best with an intact team of roughly 8 to 16 people who share the same work. Beyond that, exercises stop being about anyone's real tasks and the session drifts back toward a demo.

What should we do before booking anything?

Get a baseline. The free 7-minute AI Adoption Score shows where the gap between paid tools and daily work is largest, which tells you which team and which tasks to start with.

Next step

Compare the published formats on the workshops page, or start with the AI Adoption Score to see which team needs task-based work first.