Field note
The five-hour rule for AI training
One workshop is not a programme. The habit forms after real practice, not a demo.
The five-hour rule is a simple way to size AI training: below about five hours of hands-on practice, most people stay one-time triers, and above it they become regular users. A single lunch-and-learn almost never sticks. What changes behaviour is repeated practice on real tasks, spread over the first few weeks.
It is tempting to book one AI workshop, tick the training box, and move on. The problem is that comfort with a new tool comes from reps, not from watching a demo. People need to try AI on their own work, get it wrong, adjust, and get a result they trust, more than once, before it becomes a habit they reach for without being told.
That is why we build training as a short programme rather than an event. A first hands-on session gets people a quick win, then follow-up practice on their real tasks pushes them past the point where the habit holds. Role-specific beats generic here: the marketing team and the finance team should practise on their own jobs, not a shared toy example.
The number is not magic, and it is not a certificate. It is a planning floor. If your team has had one hour of AI training, do not expect regular use yet. Budget for the practice, and the adoption follows.
People who get more than five hours of training become regular AI users at far higher rates than those who get less.
The takeaway
Plan for at least five hours of hands-on AI practice per person over the first month, on their real work, not one all-hands session.