CEO hiring optimism needs a gross-flow workforce plan

Two plans can coexist

KPMG’s US CEO Outlook, released on 15 October, reports that 89% of respondents expect their organisation’s headcount to rise over the next three years. In a separate question about AI-related workforce strategy over two to five years, 53% report planning reductions in some areas. These are not opposing estimates of the same quantity. KPMG’s release covers 400 US CEOs within a global study of more than 1,300 large-company leaders. Axios dates the US fieldwork to 11 August–21 September.

The study records intentions, not realised jobs, worker transitions or a causal effect of AI. The public US release does not provide respondent-level overlaps, workforce weights or a response rate. It also reports that 86% are redesigning roles and team structures for human–AI collaboration. Those plans warrant investigation; they do not establish whether affected employees will reach the new roles.

Replace the net headline with a movement ledger

For a workforce leader, the practical response is to ask which flows would make the organisation’s plan true. Maintain separate records for approved external openings, completed hires, internal transfers, voluntary departures, eliminated positions and vacancies left unfilled. Keep actual movements apart from approved plans and provisional scenarios. A rising total can otherwise conceal the disappearance of the supervised work through which a particular group learns.

This is an editorial planning recommendation, not a method tested by the survey. Start with one business unit and one planning period. Give each proposed role change a destination, a process owner and a decision date. Where a role is expected to shrink, specify whether its employees can move, what prerequisite capability is missing, and who controls release from the original team. An advertised training pathway without a receiving manager is not yet a transition.

Do not subtract 53 from 89 to estimate net growth. Both figures are shares of CEOs, not numbers of employees; the questions have different scopes and horizons. One respondent may expect company-wide expansion and local contraction simultaneously. Nor should a planner assume the two respondent groups are disjoint or distribute the percentages across a national workforce.

Make the skills budget conditional

Before commissioning a course, connect it to an approved receiving task and an observable assessment. Ask the destination manager to define acceptable work, the supervision required and a realistic practice opportunity. Record whether the employee has protected learning time and access to the tools used in that work. Measure completion separately from a successful transfer and later performance; these are different outcomes.

Review the ledger at a fixed interval with employee representatives and the managers who can authorise moves. If openings disappear or requirements change, revise the pathway rather than interpreting non-transfer as a worker’s lack of motivation. Retain uncertainty around demand and do not promise employment on the basis of a CEO survey.

For a related distinction between industry contraction and occupational demand, see technical work across company boundaries. The immediate decision is not whether AI creates or destroys jobs in general. It is whether each proposed local reduction has a credible, funded route to the work the organisation says it will add.