Enterprise AI Layoff Wave Followed by Costly Rehiring as Companies Misjudge Which Roles to Cut

A pattern is emerging where enterprises that cut workers as part of AI deployment programs are now rehiring — often in the same roles they eliminated — because they underestimated how much their AI systems depended on human judgment and institutional memory.[1] A 2025 Orgvue study found 39% of business leaders had already made AI-related redundancies, and 55% of those said they made wrong calls about which jobs to remove.[1][4] Running counter to this correction narrative, Palo Alto Networks CEO Nikesh Arora argues workers themselves are the problem: 90% of enterprise employees are behind on AI adoption, and he projects 20-25% workforce change at his own company within 12 months.[4][3]

If 55% of companies that conducted AI-related layoffs now believe they made wrong calls, the cost of misidentifying which roles AI can actually replace is measurable — not just in rehiring expense, but in operational failures like Ford's defect detection and Commonwealth Bank's call routing.[1] Whether companies revise their frameworks for AI deployment or repeat the same cycle matters for how tens of millions of workers experience the next phase of enterprise AI adoption.

A cohort of enterprises that made AI-driven workforce cuts is now reversing course. A 2025 Orgvue study found 39% of business leaders had already made AI-related redundancies, and 55% of them said they made wrong calls about which jobs to remove.[1] The failure mode appears consistent: companies eliminated workers whose value came not from routine tasks — which AI handles adequately — but from understanding exceptions, navigating escalation paths, and carrying institutional memory about how systems break. As one summary put it, AI raised productivity but broke down where work depended on judgment.[1]

Three concrete cases illustrate the pattern. Ford rehired approximately 350 veteran engineers after automated quality systems failed to catch defects that experienced humans would have flagged.[1] Commonwealth Bank reversed a plan to cut 45 service roles earmarked for an AI voice bot after call volumes and complexity proved too high for the system.[1] IBM shifted from AI-heavy HR automation toward tripling U.S. entry-level hiring.[1] Canadian employers are reportedly labeling these reversals 'AI correction hires.'[2]

Palo Alto Networks CEO Nikesh Arora offers a different diagnosis. Rather than arguing that AI-driven layoffs are ill-conceived, he frames the current period as a test workers must pass: 90% of enterprise employees are unprepared for AI, and those who do not develop skills will be replaced.[3] He projects that 20-25% of his own workforce will change within 12 months as a result of AI.[4] His argument is less fire-and-hire than 'evolve or get cut' — but he explicitly criticizes reactive layoff-then-rehire cycles as costly and avoidable.[5] The 39% Orgvue figure appears in both the rehiring-wave reporting and in Arora's commentary, suggesting it has become a shared baseline reference for the broader debate.[1][4]

What remains unsettled is whether companies now rehiring have revised their AI deployment frameworks or are simply absorbing the cost of a miscalculation before repeating it. The Arora and rehiring-wave positions agree that most enterprise AI workforce strategies have gone wrong; they disagree on whether the primary failure is corporate decision-making or worker unpreparedness.