If you are the manager tasked with getting your team to actually use the new AI tools, you have probably noticed the gap. People sit through the rollout webinar. They nod at the policy slide. They favorite the link to the approved tool. And then, on Monday morning, almost nobody uses it.

The problem is rarely the tool. The problem is the training.

Most AI rollouts get treated like an standard operating procedure (SOP) read-through. Here is what it does. Here is what you cannot put into it. Here is the policy. Any questions? People leave with a vague sense of permission and zero muscle memory. They have not done anything. They have only been told about something.

A learning experience is different from a teaching experience, and the difference is where adoption actually lives.

The shift: design for the experience, not the slide deck

The instructional design field has known this for a long time. Adults learn when the task is relevant to work they actually do (andragogy). They retain procedural knowledge when they perform it, not when they hear about it (Merrill's First Principles of Instruction). And they build judgment fastest when they can compare options side by side rather than evaluate one thing in isolation (contrastive examples).

None of that has to show up on the slide. It just has to show up in the activity.

Two examples of what this looks like in practice

The Rewrite My Boring Email activity. Staff take a bland internal email and ask the approved AI tool to rewrite it three ways: friendly, concise, and executive. They line the versions up next to each other and pick the one they would actually send. Nobody lectures them on tone as a prompt instruction. They feel it. The contrast does the teaching, which is exactly what the research on contrastive examples predicts. Twenty-five minutes, real task, immediate transfer.

The AI Self-Portrait Prompt activity. Not that long ago, trending LinkedIn posts featured people appearing as superheroes. This training activity capitalizes on that by turning it into a fun yet instructive exercise. Staff use an approved AI image tool to make a caricature of their work personality. The first attempt is always slightly off. They iterate. By the third try they have internalized three core prompting moves (be specific, layer details, refine through follow-ups) without anyone ever calling it a lesson. The playful frame is not a gimmick. Humor and surprise are documented anchors for procedural memory, and a low-stakes task gives staff permission to fail in public, which is the single biggest blocker to AI experimentation in most organizations.

Both activities take less than 35 minutes. Both produce something the participant can show a colleague that afternoon. Neither requires the facilitator to be the smartest person in the room about AI.

Benefits for the facilitator

If you are running the session, here is what changes for you:

You stop being the expert and start being the host. The activity does the teaching. You frame it, set guardrails, run the timer, and lead the debrief. That is a much easier job than "stand up and explain large language models to skeptical colleagues."

You get visible outputs. Three rewritten emails. A wall of cartoon characters. Something to point at when leadership asks how the training went. Compare that to an attendance sheet from a webinar.

You spend less time worrying whether anyone learned anything. People who have just done a task in front of you have demonstrably done it. The evidence is on the screen.

Benefits for the attendee

If you are sitting in the session, here is what changes for you:

You leave with a thing you can do, not a thing you have heard about. You have already used the approved tool once for a real task. The next time is just the second time, which is a much smaller psychological step than the first.

You stop worrying about doing AI wrong. A cartoon of yourself cannot be wrong. An email rewrite that you do not like can simply be discarded. The low-stakes frame is doing real cognitive work: it dismantles the perfectionism that keeps staff from trying the tool at all.

You build judgment, not just knowledge. You have picked between three versions of an email. You have iterated on a prompt until it produced what you wanted. Those are the skills that actually transfer to next week's member communication, board memo, or campaign brief.

The takeaway for managers

If you want adoption, stop running AI training as compliance theater. Run it as twenty-five minutes of real work, with a real task, in front of other humans. The tools are not the hard part. The behavior change is. And behavior change is built by doing, comparing, laughing, and choosing, not by being told.

Pick one short activity. Run it well. Then run another one next week. That is how the rollout starts to stick.

Want more activities? Check out our AI Practice Sets toolkit!

Getting stuck? No time to facilitate the training?

Lada Consulting can provide a turnkey training plan for your team. Our training materials focus on the practical side of artificial intelligence. Book a call today to discuss how we can help!