An AI plan and staff training are not the same thing

Staff training is often bought as if it were the whole answer to AI adoption. It is not, and research from four separate consulting firms keeps finding the same pattern: executive and board-level understanding of AI, not frontline staff competency, is what predicts whether a deployment produces a measurable result. Eighty-eight percent of companies use AI somewhere in the business. Six percent get a significant financial benefit from it. Only one percent of US business leaders describe their AI rollout as fully ready.

The reason is distributional. Roughly seventy percent of the value AI produces comes from changing how a company works, its workflows, roles and decision rights, a leadership decision no amount of staff tool training can substitute for. The remaining value sits closer to the technology and the people using it day to day. Skip the leadership-level redesign and the best-trained staff in the world are still operating inside a process nobody redesigned around them.

That does not make staff training optional. Seventy-two percent of employees report their skills requirements have changed, and only thirty-six percent say they received adequate training for it. MIT researcher Paul McDonagh-Smith frames the requirement as a three-way alignment: the work itself, the people doing it, and the structures surrounding them all have to move together, or results stall regardless of how good any single piece is.

The board's own knowledge gap compounds both problems. Sixty-six percent of boards report minimal hands-on AI experience, and fifty-five percent of board members think a fellow board member should be replaced for not knowing enough about AI, the highest figure ever recorded on that question. Strategy and capacity building are not sequential. An organisation that funds one and defers the other has funded half an answer, and it shows up in the same six percent that funded neither.