For eighteen months the story ran one direction: company announces AI strategy, company announces layoffs, and the two press releases were suspiciously close together. In the summer of 2026 a second story started showing up underneath the first — the part where some of those same companies quietly call the people back.
The boomerang has a number now
According to Forrester, 55% of employers who laid off workers citing AI now say they regret it. That is not a rounding error or a few embarrassed outliers; it is a majority of the companies that ran the experiment concluding the experiment failed. A separate Orgvue survey lands on the same figure from a different angle: of the 39% of business leaders who made staff redundant because of an AI deployment, 55% admit the decision was wrong.
The regret is already turning into hiring. Staffing firm Robert Half found that nearly one in three U.S. hiring managers who eliminated a role primarily because of AI later rehired for the same or a comparable position. More than a third of companies that cut roles have already brought back more than half of them, most within six months. The AI didn’t quit — it just couldn’t do the whole job, and the invoice for finding that out is a second recruiting cycle.
Ford went looking for the gray beards
The cleanest example is Ford, which spent the past two years leaning on automated quality systems and was underwhelmed by the results. Chief operating officer Kumar Galhotra put it plainly: the company had been “relying more and more on automated quality systems” and the quality wasn’t there. So Ford hired 350 veteran engineers — some former employees, some pulled from suppliers — specifically to hunt for failure points before a part reaches the plant floor.
The engineers didn’t replace the AI tools; they rebuilt them and now run the troubleshooting. CEO Jim Farley says the effort has generated hundreds of millions of dollars in savings from lower warranty and recall costs, and Ford topped J.D. Power’s 2026 Initial Quality Study for the first time since 2010. The lesson Ford paid for is not that AI is useless. It is that the judgment of someone who has watched ten thousand parts fail is not a line item you can delete and re-download.
IBM and the 6% that needs a person
IBM is the other tell, and it is instructive because IBM was one of the loudest automation stories of the cycle. Its AskHR assistant resolves 94% of routine HR queries without a human — a genuinely impressive number that got quoted in a lot of layoff decks. But the company is now tripling its U.S. entry-level hiring in 2026, because the remaining 6% of cases — the ones that need ethical judgment, escalation, or just a person who can tell when the script is about to make things worse — turned out to be the 6% that mattered. Automating the easy nine-tenths of a job does not make the last tenth disappear. It concentrates it, and someone still has to be there when it lands.
Commonwealth Bank learned the same thing faster and more publicly, reversing a round of layoffs when its AI voice bots couldn’t handle real customers.
Before you frame this as good news
The rehiring is real, but the terms are worth reading before anyone celebrates. Forrester’s own Predictions 2026 report expects half of AI-attributed layoffs to be quietly reversed — and then adds the catch: many of those roles will come back offshore, or onshore at meaningfully lower salaries. The boomerang comes back, but it does not always come back to the same hand, and it is often lighter.
So the honest read for anyone watching their own field is mixed. The comforting half is that the “AI replaces you outright” narrative got tested at scale and a majority of the companies that tested it flinched. The uncomfortable half is that “we rehired the role” and “we rehired you, at your old pay, in your old country” are not the same sentence. The work that survives is the work AI can’t finish — judgment, escalation, the failure-hunting that Ford paid 350 salaries to get back. That is the safe ground. It just isn’t priced the way it used to be.
