A request for training is almost never a diagnosis. It is a symptom description delivered in the only vocabulary most organizations have for performance problems.
When people know what to do and still cannot do it consistently, the constraint is somewhere other than knowledge. It may be an unclear process, a system that makes the correct action slower than the incorrect one, conflicting priorities from leadership, or an absence of feedback that would let anyone notice the drift.
Building a course in those conditions is not neutral. It consumes budget, it consumes the time of the people you are trying to help, and — most damaging — it produces evidence that the problem was addressed. Six months later the performance gap is still there and the organization has less appetite to fund the intervention that would have worked.
The discipline is simple to state and difficult to hold: before designing anything, establish what would change if we built nothing. If the honest answer is 'nothing, because the constraint is elsewhere,' the first deliverable is a clear articulation of where the constraint actually is.
This is not a refusal to build. It is the precondition for building something that works.