Organizations procure AI tools, distribute licenses, run an enablement session, and then measure disappointing usage. The conclusion drawn is usually that people need more training.
The more accurate reading is that adoption stalls where every change effort stalls: unclear permission, unclear benefit to the individual, competing priorities, and no visible model of what good looks like in this specific job.
Treating it as a change problem shifts the interventions. Leaders have to state what is sanctioned. Workflows have to be identified where the benefit is obvious to the person doing the work. Early adopters have to be visible. And the standard for output quality has to be explicit, because ambiguity here reads as risk and people respond to risk by not using the tool.
None of that is a curriculum. All of it is enablement.