Spreading training thinly across everyone gives a good completion rate but rarely moves anyone past guided use, while depth costs visibility; here is how to decide where the hours go.
We wrote last time about the seven capability levels and why Level 3, independent competent use, is the floor below which nothing durable has happened yet. The obvious next question, in any budget conversation, is where the limited training hours actually go: spread thin across everyone, or concentrated on fewer people until they cross that floor. Most training plans quietly avoid answering this question and default to breadth, because breadth is the answer nobody has to defend in a room.
This piece is about naming that default for what it is, and about the cost that comes with each side of the choice, because both sides have one.
Training everyone a little produces a completion percentage that looks good in a board pack, treats every department as equally deserving, and avoids the genuinely uncomfortable conversation about which team gets the deeper, more expensive investment first. None of this is dishonest. It is simply easier to announce "eighty per cent of staff completed AI awareness training" than to explain why finance got three days of hands-on coaching and the rest of the organisation got a ninety-minute session.
A wide, shallow rollout genuinely does something: it lowers fear across the organisation, builds a shared vocabulary so people can at least discuss the tools without embarrassment, and reduces the resistance the next, deeper phase of training will meet. What it does not do is move anyone past guided use. Breadth training, honestly assessed, produces Level 2 at best, and Level 2 is the level that evaporates the moment the trainer leaves the room.
Concentrating the same hours on a smaller group, until they reach Level 3 and beyond, produces something breadth cannot: people who can do the actual work correctly, alone, and who can then unblock a colleague without escalating to the champion. A department with three genuinely competent people spreads capability sideways in a way that thirty people at guided use never do, because the three can teach, and the thirty at guided use have nothing durable to pass on.
Breadth trains everyone to nod along. Depth trains someone who can actually do the work when the trainer has left the room.
Depth training excludes people, visibly, in a way that invites a fair question from everyone not chosen: why not me. Breadth training excludes no one and answers that question before it is asked, at the cost of producing nothing that survives past the training session for anyone. The honest answer is that both costs are real, and an organisation that pretends otherwise, by calling breadth training "capability building" in a board report, is spending real money on a number that will not hold up when someone tests it in six months.
Spend most of the training budget on depth for the department that will use the capability daily, and be explicit that broader awareness training across the rest of the organisation is building vocabulary and reducing fear, not building capability; do not let a board report conflate the two.
No external statistic cited; this article extends the organisation's own capability-level model rather than presenting third-party evidence.
If this is the question on your desk, a thirty-minute call tells you whether the service fits, or that you do not need us yet.