How do I teach AI when some employees are beginners and others already use it?
Give everyone the same task and the same definition of a useful result, then change how hard the practice is to fit where each person starts. Everyone gets held to one standard, at different speeds.
Teresa Carlson described the different starting points inside an organization during the summit conversation. Some employees pick up the tools quickly, and others need more time and support. That's a reason to redesign the learning, and nobody has to be labeled resistant or ready.
One task can serve both groups
A shared task for the session could be preparing a meeting brief from an approved set of documents. Agree on what a useful result looks like: accurate facts and visible sources, with open questions clearly marked.
From there, vary the practice. A beginner might work through the assignment with a guided example and a short checklist, while someone with more experience tests it against conflicting information or an incomplete source. Both are learning to judge the output, at different levels of difficulty.
Ask people to show what they checked and what they changed. The fastest draft doesn't always come from the person who's learned to use the tool responsibly.
For your next session, prepare one core exercise and an optional harder case, and leave time for questions and revision. You can hold everyone to the same standard without expecting everyone to reach it the same way.