The Judgment Gap: Why AI Job Cuts Are Reversing
Gartner reckons half of the AI-driven workforce cuts will reverse by 2027. The reason says more about what AI can't do than what it can.
That prediction, and the Gartner research behind it, opened Episode 67.
Jenni and I spent a good chunk of the episode on this one, because the pattern underneath it is so tidy it's almost embarrassing. Forrester found that 55% of leaders who restructured around AI now regret the decision. The companies unwinding those cuts had replaced roles that needed human judgment and got information retrieval instead. Those are not the same thing, and the gap between them is where the money went.
Klarna is the case everyone points to, and rightly. In 2024 the company said its AI assistant did the work of 700 customer service agents. By 2025 it was hiring people back, not for the easy, repeatable queries the bot handles fine, but for the fraud disputes, the complex billing, the emotionally charged calls. A different, more skilled profile than the outsourced support it had cut. The automation worked. The assumption that it could cover everything did not.
The line from the episode I keep coming back to is that AI has read everything but it's lived nothing. There's a real difference between reading about surgery a thousand times and having done surgery a thousand times. One produces knowledge. The other produces judgment. Plenty of organisations spent the last two years confusing the two, and the reversal is the bill arriving.
My take is blunter than the research: if the work needed judgment, you shouldn't have let those people go in the first place. The smart version was always to deploy the bot on the volume and keep your people on the hard stuff. Instead a lot of leaders decided they'd be cleverer than everyone else, made the cuts, and are now quietly rehiring, often at higher pay, while framing it as strategy rather than a correction.
That's the part worth sitting with. Almost nobody is going to stand up internally and say "we got this wrong." So the honest question for anyone in comms is what you do when leadership wants a reversal dressed up as a plan. Does 2027 even feel slow to you? Because from where we sit, it's already happening.
Also in this episode:
The office amenity that actually improves teamwork. A Harvard Business Review study of more than 6,000 knowledge workers found the strongest teams aren't defined by perks or work arrangements, but by their ability to protect focused work through meeting-free periods and fewer interruptions. Jenni and I got into why "busy" became a badge of honour, why focus blocks die within a week when nobody protects them, and why the real skill organisations need to teach is saying no. It sounds heavenly on paper. Making it stick needs backing from the top, not another calendar tip.
Internal comms is useful, but it's not important. Thanks to Jason Anthoine for sharing Cameron Wood's piece, one of very few LinkedIn articles I've read all the way through. Two lines landed. First, that opening isn't reading, reading isn't understanding, understanding isn't believing, and believing isn't changing behaviour, so blending the first outcome with the last is how comms teams have misled themselves for years. Second, that most leaders think the function is useful rather than important, and there's a difference. Jenni brought her own research, including VMA Group data showing the value CEOs place on internal comms actually dropped after the pandemic. Useful in the moment, then put down. Our take: start from the behaviour you want to change and work backwards, and stop treating outputs as proof of outcomes.
Germany tightens its sick leave rules. The coalition government now requires a medical certificate from the first day of illness and has scrapped the pandemic-era option of phoning in sick. The economic argument is real, with absence at record levels, but the signal it sends is "prove it" rather than "let's understand why people are off." Jenni and I weighed whether it just pushes unwell people into work, or into a doctor's waiting room they can't get into, and whether it treats a symptom while ignoring the root cause. My expectation is that it fails, and that other countries watch closely before copying it.
Written by Chuck Gose, founder of ICology.
The Frequency Podcast
Real talk about comms, culture, and employee experience.
Chuck Gose and Jenni Field skip the buzzwords and get straight to what matters. New episodes every week.
Listen & Subscribe