This is analysis, not investment advice.

For a year, the story was that AI would let companies do more with fewer people. Now a quieter story is unfolding: some of the firms that cut jobs and blamed AI are hiring those people back, CNBC reports — having learned the hard way that today's AI couldn't actually replace them.

The reversal

Surveys of employers point to a notable share of AI-driven layoffs being walked back within months, with many managers admitting regret — and analysts at Gartner have projected that a large chunk of companies that swapped customer-service or operations staff for AI will end up restaffing those roles (often under new titles). Exact figures vary by survey and should be read with caution, but the direction is consistent: some cuts went too deep, too fast.

The pattern is clearest where the real-world failures were visible.

What went wrong

The common thread: AI handled part of the job, not all of it — and the missing part needed humans.

  • Klarna, the Swedish fintech, leaned heavily on an AI assistant to handle customer service and cut back its support staff — then publicly changed course, moving to rehire human agents after service quality and customer satisfaction slipped.
  • Air Canada was held legally liable when its chatbot invented a refund policy that didn't exist.
  • McDonald's scrapped an AI drive-thru ordering test after a run of errors.

Independent research points the same way: customer-service is one of the areas where AI stumbles most, and surveys find most consumers are frustrated by AI-only support, CNBC reported. Many AI projects underdeliver not because the idea is wrong but because the data and integration work was never done — companies cut first and figured out the AI later. In some cases, "AI" appears to have been a convenient cover for ordinary cost-cutting.

The honest, unsettled bigger picture

Here's the part that resists a tidy conclusion. The rehiring wave does not mean AI won't reshape work — and it sits against real weakness elsewhere. Boursel has tracked a "low-hire" job market (the JOLTS hiring rate near multi-year lows), soft entry-level and young-worker hiring in AI-exposed fields (per research from Anthropic and others), and continued layoffs even at giants like Microsoft, which keeps trimming staff while pouring billions into AI.

So both things are true at once: AI is genuinely changing work and pressuring some jobs — and the claim that it can wholesale replace workers right now has been overstated. The reversal is a story about the gap between AI hype and current capability, not proof that the technology is a dud.

Why it matters

For workers, it's a small but real corrective to fear-driven headlines: today's AI is more often a tool that augments people than a drop-in replacement. For companies, it's a costly lesson — cutting staff on the promise of an AI that isn't ready can backfire, forcing expensive rehiring and lost quality; the smarter path is to prove the AI works before removing the humans. And for investors and executives, it's a reason to treat sweeping "AI will let us do more with less" claims with the same skepticism as any other projection. Boursel makes no forecast on AI and employment; the takeaway is that the future of work with AI is arriving slower and messier than last year's headlines promised — and, for now, the humans are still needed.