Should we cut headcount now that AI can do some of the work?

Action before thinking is a precarious position for an executive, and there is now a public record of what happens when this particular action comes first.

Ford has reportedly rehired experienced engineers to address quality issues that automated systems could not handle. Their vice president of vehicle hardware engineering put it plainly: artificial intelligence is a fantastic tool, but it is only as good as the information you use to train it.

Commonwealth Bank of Australia replaced more than 40 customer service staff with an AI voice system, found the system could not cope, saw call volume increase, and reversed the cuts. The bank acknowledged it had not adequately considered all relevant business factors and should have been more thorough in assessing which roles were required.

IBM moved a large share of routine HR requests to AI and handled the great majority of them successfully. The remainder, which included ethical judgment calls, it could not. The company subsequently announced plans to expand entry-level hiring, with its chief human resources officer noting that without continued investment in entry-level hires, the pipeline simply dries up in three to five years.

Survey work by the workforce analytics firm Orgvue and separate data from Robert Half both point the same direction: a substantial share of leaders who made AI-driven redundancies later concluded the decision was wrong, and a meaningful number rehired for the same or a similar role. The pattern behind those findings is consistent with what the named reversals show.

Where AI outputs are inconsistent or hard to apply, companies end up reintroducing human oversight, which duplicates effort and erases the productivity gain. Hollowing out the workforce, particularly at entry level, is a dangerous long-term strategy. AI augments the human. It does not eliminate the need for one.