The Gartner Number

350 global business executives. All at companies with at least a billion dollars in annual revenue. Gartner asked them about AI adoption and what it had done to their workforce.

80% of the companies that had piloted AI or autonomous technology had carried out workforce reductions.

If the findings stopped there, this would be a predictable story. It does not stop there.

The same study found zero correlation between those layoffs and higher returns on investment. The companies cutting staff because of AI were not the ones making money from AI. The companies actually seeing high ROI were doing something different entirely: keeping their people, and using AI to make them more productive.

Gartner's VP analyst put it plainly: "Chasing value only through headcount reduction is likely to lead most organisations down a path of limited returns."

Companies are performing a human sacrifice to a god that did not ask for one, and the god is not rewarding it.

Why They're Doing It Anyway

The rational explanation for cutting staff when AI doesn't improve your returns is that cost reduction looks like ROI on a spreadsheet even when it isn't. You cut ten people, you save their salaries, you book the saving as a gain. The AI investment looks justified. Nobody has to explain why the underlying business isn't any better.

The second explanation is pressure. When a company's competitors announce AI-driven restructuring, there is boardroom pressure to do the same. The announcement produces a short-term stock bump. The long-term consequences land quietly, years later, when the institutional knowledge that walked out the door turns out to be irreplaceable.

The third explanation is that this is genuinely hard to measure in real time. The companies that kept their staff and deployed AI for augmentation will take 18 to 24 months to show the ROI difference clearly. The companies that cut and claimed efficiency show results immediately, even if those results are hollow.

The Jevons Paradox Nobody Is Talking About

In 1865, an English economist named William Stanley Jevons noticed something strange. Steam engines had just become dramatically more efficient , you needed less coal to produce the same energy. Total coal consumption went up. Massively. Because when coal became cheaper and more efficient to use, people found more uses for it. New industries became viable. New applications emerged.

Efficiency didn't reduce demand. It exploded it.

Economists are now arguing the same logic applies to AI. If AI makes certain types of work cheaper and faster, demand for that work doesn't disappear , it expands. New applications become viable that weren't viable before. New roles emerge around the technology.

The early data supports this. AI has already created more than 1.3 million new jobs globally, including roles that didn't exist in 2023. LinkedIn's fastest-growing job title in the US for 2026: AI engineer, up 143% year-over-year. Four of the top five fastest-growing job titles are AI-related.

Three years ago, prompt engineer wasn't a job title. It now has a 135.8% growth rate.

The Two Camps, and Which One Is Winning

The companies making money from AI are not the ones replacing people with AI. They are the ones making their people more capable with AI , reducing the time spent on low-value work, freeing headcount for higher-judgment tasks, and building workflows that compound over time.

The companies losing money from AI are the ones who saw a technology and saw cost reduction. They treated it as a substitution play when it is actually an augmentation play. The substitution play produces short-term optics and long-term damage.

We invented a technology powerful enough to require an entirely new profession dedicated to asking it questions correctly. That is either a breakthrough or the most expensive autocomplete in history.

Honestly, it might be both. The companies treating it as a breakthrough are winning. The ones treating it as a headcount calculator are not.