Companies replacing workers with AI: what the record shows, company by company (2026)

Every week brings a new list of companies replacing workers with AI. This week it is buyouts, displacement estimates and budgets moving from payroll to compute.

The lists repeat the same names. Almost none of them check what each company actually said, in its own filing, about what the AI did.

TIN has read seven of those companies against the public record. The most quoted number, Klarna’s “700 full-time agents”, was a stated workload equivalence, not 700 layoffs [source].

What “companies replacing workers with AI” means

Companies replacing workers with AI are firms that cut roles and name AI as the work that replaces them, as distinct from firms that cut roles for other reasons in the same year.

That distinction is the whole post. A layoff and an AI rollout in the same quarter is not a replacement. A replacement needs the company to say the AI is doing the work the people did.

What companies are replacing workers with AI?

The named cases sort into five groups, and only one of them is AI doing the job instead of people. The others are equivalence claims, mix shifts, AI failures and AI disruption the company did not blame on AI.

Workload equivalences. Klarna said its assistant handled 2.3 million conversations, two-thirds of its customer-service chats, in its first month [source]. The 700-agent line was its estimate of what that volume equalled.

Octopus Energy’s CEO made the same move: its drafting tool did “the work of 250 people”, and he said it was “unlikely to lead to job cuts” [source]. In 2024 Octopus announced 4,000 new jobs [source].

A mix shift. Zalando said in 2018 it would replace “200 to 250 positions previously carried out by people with algorithms and artificial intelligence” [source]. The same report carried a plan to hire an additional 2,000 people [source].

TIN has written up each of these three on its own: the Klarna reversal, the Octopus Energy case and the Zalando announcement. This page puts them side by side with the cases lists leave out.

Are AI layoffs real?

Yes, and the clearest ones on file were caused by AI failing, not by AI succeeding. Two companies cut a fifth or more of their staff after an automated pricing model bought homes it could not sell at a profit.

Zillow announced on 2 November 2021 that it would wind down Zillow Offers, its iBuying business. CEO Rich Barton said “the unpredictability in forecasting home prices far exceeds what we anticipated” [source].

CNN reported the shutdown meant “a 25% reduction in its staff” and that “the company said it will be cutting some 2,000 jobs” [source].

A year later Opendoor reported “an inventory valuation adjustment of $573 million” in one quarter [source]. Weeks after, it let go “about 550 people, or 18% of the company” [source].

Those are real job losses with an algorithm at the centre. They are the opposite of the replacement story: the people went because the model’s prices were wrong.

Is AI causing layoffs where the company does not say so?

Sometimes, and Stack Overflow is the documented case. A peer-reviewed study measured the AI effect on its business, and the company’s own layoff statement did not mention AI.

The study found that within six months of ChatGPT’s release, activity on Stack Overflow “decreased by 25%” relative to counterpart forums [source].

In October 2023 Stack Overflow laid off 28% of its workforce, and its CEO cited “macroeconomic pressures impacting the entire tech industry” [source]. TIN’s case file keeps the measured decline and the stated reason apart, and so does this post.

Why are companies replacing workers with AI?

Companies say cost, and the one firm that ran the experiment to its end said cost was the mistake. Klarna’s CEO told the press that cost “seems to have been a too predominant evaluation factor” and “what you end up having is lower quality” [source].

Klarna’s own headcount did fall, from 5,527 at the end of 2022 to 3,422 at the end of 2024 per its IPO prospectus [source]. Its CEO attributed much of that to attrition under a hiring freeze, “not solely due to AI” ([source]).

The other reason is the one nobody puts in a press release: the AI story is worth more than the AI. Presto Automation told investors its drive-thru voice AI reached automation rates of 95% to 99% [source].

The SEC found that on the advanced 2023 pilot, humans intervened on approximately 70% of orders, and on the original version “in all instances” [source]. The replacement was off-site human agents entering orders behind a voice.

How many workers has AI replaced, by the measured record?

None that a national register can see yet, in the one study that linked AI use to official pay and hours. That is a narrower claim than “none”, and it is the one the evidence supports.

Humlum and Vestergaard linked chatbot surveys of about 25,000 Danish workers in 11 exposed occupations to administrative records. They found “precise null effects on earnings and recorded hours”, ruling out effects larger than 2% two years after ChatGPT [source].

It is a working paper, not peer-reviewed, and it covers one country and two years. It does not say AI never replaces anyone. It says the replacement wave in the headlines has not yet shown up where wages are recorded.

The proof

TIN’s case files hold each company’s claim next to what independent sources and filings show:

• Klarna’s assistant: the 2024 numbers and the 2025 walk-back, with headcount figures that do not reconcile printed side by side.
• Octopus Energy’s Magic Ink, where the satisfaction figure moves between statements and is shown, not merged.
• Zalando’s 2018 announcement, with the hiring plan the retelling drops.
• Zillow Offers’ shutdown and Opendoor’s 2022 quarter, both from SEC filings.
• Stack Overflow’s measured decline, Presto’s SEC order and the Danish register study.

Read the limits. Seven companies are not a census, and this post claims no total for AI job losses. It shows what kind of claim each famous name actually is.

Companies replacing workers with AI, by what the record shows

The bottom line

The list everyone shares assumes each name is a company where AI took over the work. Read the filings and almost none of them are. The equivalences were capacity estimates, the biggest AI-linked layoffs followed AI getting the job wrong, and the one firm that cut on cost said quality fell.

When a company says AI replaced its people, ask which column its claim belongs in before you repeat it: that question works on any automation headline, not only this one.