Organisations are answering the wrong question. The right question is not "what skills do our people need?" — it is "what does the work unit require of humans once AI handles the rest?" You cannot out-train a moving target.
Redesigning work before reskilling produces higher readiness
- McKinsey's 2022 global reskilling survey: 87% of executives anticipated significant skills gaps within five years, yet fewer than one in three had a structured plan. Three years on, AI has not waited — and the planning deficit has compounded.
- Deloitte's 2024 Global Human Capital Trends: organisations that redesigned work units before building competency frameworks reported significantly higher workforce readiness scores than those that ran learning programmes alone.
Closing the gap is a design decision, not a training budget
Audit your competency framework against your AI deployment roadmap — they should be the same document, built from the same analysis of what each work unit requires of humans once AI is embedded. They almost certainly are not. That misalignment is where the competency gap lives, and closing it is a design decision, not a training budget decision. D5 frames this as a workspace design problem: the work unit is the intervention point, and the training catalogue is downstream of it.
