How to run AI training for your team without wasting a day

Most AI training fails for the same reason: it is generic. People watch someone demonstrate clever prompts on invented examples, nod along, and go back to their inbox unchanged. Here is how to avoid that.

1. Keep the group small

Three to ten people is ideal. Small enough that everyone gets attention and can ask about their own work, large enough that the session creates shared habits across a team rather than one enthusiastic individual.

2. Send real material in advance

Anonymised examples of the documents, reports and spreadsheets your team actually produces are what turn a demo into training. If the exercises use your own material, the skills transfer immediately.

3. Agree your data rules before, not after

Decide what staff may and may not paste into an AI assistant, and cover it in the session so everyone hears the same thing. A simple traffic-light rule is easier to follow than a long policy document.

4. Build something reusable in the room

The highest-value hour of any workshop is the one where the group builds a Project or a set of saved prompts together - your formats, your tone, your process captured once and reused by everyone from then on.

5. Plan the follow-up

• Ask each person to name one weekly task they will do with Claude
• Share the saved prompts somewhere the whole team can reach
• Check in after a month and collect what worked

Done this way, a single three-hour on-site session changes how a team works. Done generically, a full day changes nothing.