Contact me

AI in 2026: The Companies Learning the Expensive Way

Jon Jensen

by Jon Jensen, Product Designer

Last year I wrote about how AI doesn't make your expertise obsolete. It makes expertise more important, because someone has to catch what the machine gets confidently wrong. That was an argument about individuals. This is what happened when companies bet the other way.

The headline of 2024 and 2025 was the layoffs. Cut the team, hand the work to AI, watch the margins climb. The headline of 2026 is quieter and far more revealing: a remarkable number of those companies are walking it back.

The reversal nobody bragged about

A February 2026 Careerminds survey of 600 HR leaders who had made AI-driven cuts found that roughly two in three were already rehiring for roles they had eliminated. They weren't waiting long, either. More than half rehired within six months, and some within three.
A separate Forrester survey found that 55% of employers regret replacing workers with AI. And Gartner now projects that half of all companies that cut jobs for AI-related reasons will rehire for similar roles by 2027.

That is not a rounding error. That is a large share of an industry discovering, in under a year, that the demo didn't survive contact with production.

Same technology, opposite outcomes

The post-mortems all rhyme. AI turns out to be a genuine accelerator and a catastrophic full replacement, and the difference between those two results is almost entirely a staffing decision.

Klarna became the cautionary headline after replacing hundreds of customer service reps with a chatbot, hitting a quality wall, and quietly reversing course. The instructive contrast is IKEA, which handed roughly 47% of customer queries to an AI bot named Billie but kept its people, reskilling 8,500 call-center workers into remote interior design advisers. That human-plus-AI channel generated €1.3 billion, about 3.3% of total revenue, in a single fiscal year.

Same technology. Opposite philosophy about what to do with the humans. Opposite result. The companies that treated AI as a reason to delete expertise got worse; the ones that treated it as a tool to amplify expertise got better.

Not everything is an AI story, but the rehiring is

It's worth being honest that a meaningful share of recent tech cuts weren't really about AI at all. A lot of it was a correction to pandemic-era over-hiring, dressed up in AI language because "we're more efficient now" is a more flattering board-meeting line than "we hired too many people in 2021." Industry analysts have been clear that AI is a contributing factor to the layoff wave, not the main driver for most of it.

But the rehiring wave cuts straight through the spin. You can rationalize a layoff a dozen ways. It's much harder to explain why you're posting the exact same job six months after eliminating it. When a company removes a function, tries to run it on AI alone, and is back hiring for it within two quarters, that's not a narrative. It's a controlled experiment with a published result.

The competitive read

Here's the part that should matter to anyone weighing their own position, whether you're an IC wondering if you're safe or a leader deciding where to cut.

The firms pulling ahead in 2026 are not the ones that cut deepest. The real risk, as one analysis put it, is over-correcting: cutting too far and opening skills gaps that quietly erode quality until they show up in the product. A competitor that kept its experienced engineers and designers and handed them AI to amplify their output now operates with both speed and judgment. The company that fired those people and kept only the AI has speed and no one left who can tell when the output is confidently wrong.
And the output still is wrong, often. Even the best models in 2026 hallucinate well into the double digits on genuinely hard, open-ended tasks. Reasoning-tuned models can actually be more prone to fabricating with confidence, not less. So the choice isn't really "humans or AI." It's "AI with experienced people steering it" versus "AI with nobody qualified watching." In a year where the technology is this capable and this unreliable at the same time, that's not a close race.

The takeaway

A year ago the question was personal: will AI replace my job? The answer was that it makes your expertise more valuable, not less.

The 2026 version is organizational, and the market has started answering it for us. AI doesn't make expertise obsolete. It makes the absence of expertise visible, fast, and expensive, and a growing list of companies have the rehiring budgets to prove it.

If you're a builder, that's reassuring. If you're deciding who to cut, it's a warning. The teams that win the next few years won't be the ones that replaced their people with AI. They'll be the ones that gave their best people the best tools and got out of the way.

More articles

"Just sprinkle some UX on it..."

A common misconception in the tech industry is that UX can be added at the end of a project like a finishing touch. This article explores why this approach is fundamentally flawed and how to integrate UX thinking from the very beginning.

Read more

AI in 2025: You're Still the Expert

I've heard many designers and developers share fears about AI replacing them in the future. I outline reasons this is only true for designers that don't adopt AI into their daily routine.

Read more

Tell me about your project

Say Hey