How to train a team to use AI in their work

Start with a problem the team recognizes

Ask which tasks take too long, which information is hard to find and which mistakes keep recurring. It is easier to learn to evaluate a tool on a familiar problem. A fictional sales inquiry lets everyone practice with the same material and compare how the responses improve when the context becomes clearer.

Also teach how to detect a bad result

Include an incomplete answer or an outdated source in the exercise. The team needs to know when to ask for evidence, improve the instructions or handle the task another way. Do more than teach a phrase that produces a convincing answer: teach people how to verify the result.

Make the practice continue after the course

Each participant should leave with a task to work on, reusable instructions and a way to check the result. Collect examples of what worked and what failed so the team can improve together. My course uses three in-person sessions to help people move from observing AI to using it. Progress continues when that practice becomes part of their work.

Put it into practice

• Work with one task per person or function.
• Use material that the team is allowed to share.
• Practice review as well as generation.
• Save instructions and examples that can be reused.