Ford, IBM and CBA reverse AI-driven layoffs as automation stalls

The interesting story in AI headcount right now is not a fresh wave of cuts. It is the number of companies quietly walking back the ones they already made. CNBC reports that, of the 39% of business leaders who made employees redundant citing AI deployment, 55% now say those decisions were wrong, according to a survey by Orgvue. Data from staffing firm Robert Half lands in the same place: about 32% of U.S. hiring managers said they eliminated a role primarily due to AI and later rehired for the same or a similar position.

Three named examples give the pattern texture. Ford, per CNBC, rehired, newly hired, or promoted 350 experienced engineers after automated quality-control systems could not catch design flaws the way seasoned humans could, and the company then topped JD Power's 2026 Initial Quality Study for the first time since 2010. IBM's AI system deployed to take over human resources work managed to address roughly 94% of incoming requests, but the roughly 6% it could not resolve, including situations involving ethical judgment, exposed the limits of full automation; IBM now plans to triple its U.S. entry-level hiring across all business units in 2026. Commonwealth Bank of Australia laid off more than 40 customer service staff and replaced them with an AI voice bot, but the bot could not cope with rising call volumes and CBA reversed the cuts.

The throughline the reporting draws is that firms budgeted on tech to replace humans without investing in training or upskilling, which left teams unprepared to leverage AI. Where outputs came back inconsistent or wrong, humans had to be reintroduced to provide oversight, often at premium cost, duplicating work the automation was supposed to eliminate.

The honest caveat is that this is one survey, one staffing dataset, and a handful of high-profile anecdotes, not a settled economic picture. The reporting does not tell you whether rehired workers are returning at prior pay, which sectors beyond auto, banking and tech are seeing the sharpest reversals, or how many quiet walkbacks are not showing up in press releases. Take the specifics as reported, not as a definitive turn.

The direction worth watching is the framing shift. Firms treating AI as augmentation and pairing it with training and human oversight are the ones showing up in these reversals with entry-level hiring plans intact, while firms that booked the automation savings first are the ones eating rehire costs and lost institutional knowledge. If your organisation is still pitching AI internally as a headcount line item, the 55% figure is the one to put in the deck.