Inside Klarna's Systematic Replacement of Its Customer Service Workforce
In September 2024, Klarna CEO Sebastian Siemiatkowski posted on X that the company's AI assistant was handling the work of 700 human employees. He framed it as a triumph. The number had already been circulating in investor materials. What received less attention was what had happened to those 700 people in the two years prior.
Between 2022 and 2024, Klarna reduced its global headcount from roughly 7,000 employees to approximately 3,800 — a cut of nearly 46 percent. The company attributed much of that reduction to a deliberate decision to stop hiring in roles where AI could be deployed instead. Siemiatkowski told the Financial Times in August 2024 that Klarna had implemented a hiring freeze beginning in 2023 specifically to allow AI to absorb the headcount. "We're proving every month that we can do more and more work with fewer and fewer people," he said.
The customer service function was the sharpest example. Klarna's internal data, which Siemiatkowski shared publicly in a February 2024 press release, showed its AI assistant handling 2.3 million customer service conversations in its first month of full deployment — about two-thirds of Klarna's total customer service volume. The company reported that the AI resolved issues in an average of two minutes, compared to eleven minutes for human agents, and that customer satisfaction scores were equivalent. Klarna claimed the shift would generate $40 million in profit improvement in 2024.
The figures came directly from Klarna. They have not been independently audited. But the headcount trajectory is corroborated by the company's own public statements and reporting from Bloomberg, the Wall Street Journal, and the Financial Times, all of which covered the workforce reduction in detail.
How the Reduction Actually Happened
Klarna's approach was not, primarily, mass layoffs. It was attrition managed by design. The company stopped backfilling positions when people left, particularly in customer operations. This meant that workers who resigned or whose contracts ended were not replaced by humans. They were replaced — or simply not replaced at all — while the AI assistant absorbed their former workload.
This is significant because attrition-based displacement is structurally harder to track than a formal reduction in force. There is no WARN Act filing, no earnings call announcement of layoffs, no press release headlined "Klarna Cuts 700 Jobs." The reduction happened in the negative space: in roles that disappeared quietly as contracts expired across the company's outsourced and in-house customer service operations in markets including Germany, Scandinavia, the United Kingdom, and the United States.
Klarna relied heavily on outsourced customer service contractors, which made the displacement even less visible. When a third-party vendor loses a contract because the client has automated the function, those workers do not appear in the client company's headcount figures at all. The 700-employee equivalent Siemiatkowski cited referred to workload absorption, not to a specific set of individuals who had held Klarna badges.
Labor researchers have flagged this pattern as one of the primary mechanisms through which AI-driven displacement occurs without generating the public signal of announced layoffs. Lawrence Katz, a Harvard labor economist, described a similar dynamic in a 2024 interview with the New York Times, noting that automation tends to reduce employment through the elimination of new job creation rather than through visible mass termination events.
What the Numbers Tell Us About Scale
Klarna is not alone, but it has been unusually candid. Most companies deploying AI at scale in customer operations have been reluctant to connect automation investments directly to headcount outcomes in public statements. Siemiatkowski's decision to publicize the 700-employee-equivalent figure — partly as an investor narrative ahead of the company's IPO, which was filed with the SEC in November 2024 — made Klarna a rare documented case study.
The Goldman Sachs Economics research team estimated in a widely cited 2023 report that roughly 300 million full-time jobs globally were exposed to automation, with customer service and administrative support among the highest-exposure categories. The McKinsey Global Institute, in its 2023 report on generative AI's economic potential, identified customer operations as one of four business functions where generative AI could deliver the most measurable productivity gains — and where workforce impact would be most direct.
The Oxford Martin School's work on occupational automation probability, which predates the current generation of large language models, had already placed customer service representatives in a high-risk category. What the Klarna case demonstrates is that these probabilistic estimates are now resolving into specific corporate decisions with named outcomes.
Klarna's IPO prospectus, filed with the SEC, disclosed that the company's operating expenses had decreased substantially as a proportion of revenue, with customer operations costs cited as a key driver. The prospectus also included risk language acknowledging that the company's continued reliance on AI meant its workforce would likely remain smaller than historical norms relative to its transaction volume. That is the company, in a legal document, telling investors that the workforce reduction is structural and intentional, not cyclical.
The Stanford AI Index 2024, published by the Stanford Institute for Human-Centered Artificial Intelligence, documented a surge in enterprise AI adoption across financial services and fintech, with customer-facing automation tools among the fastest-growing deployment categories. Klarna's trajectory sits within that broader pattern, but its willingness to quantify the labor implications in CEO statements and investor materials makes it traceable in a way that most peer companies are not.
Aria Chen, who built the income protection service clankers.today after losing her own software engineering role to AI-driven restructuring in 2024, has pointed to Klarna as an illustration of why displacement is difficult to anticipate at the individual level. The workers whose roles were absorbed were not notified that their position was being eliminated — they simply found that contracts were not renewed, or that new openings in their function no longer appeared. The signal, if it existed at all, was institutional rather than personal.
What Klarna's case clarifies is the mechanism. The displacement did not require a dramatic announcement. It required a hiring freeze, an AI deployment, and time. The 700-employee figure Siemiatkowski cited so precisely was not the product of a single decision made in a boardroom. It was the cumulative result of many smaller decisions — to not post a job, to not renew a contract, to route one more conversation to the AI queue — made across multiple quarters in multiple markets.
That granularity makes the story harder to tell and harder to prepare for. It also makes it harder to dispute. The math is in the press release. The headcount is in the investor filings. The CEO said the quiet part into a microphone. Klarna is not a warning sign. It is a completed event, and the documentation is public.
PROTECT YOUR INCOME NOW
If AI automation threatens your job, preparation is the only honest answer: know your risk, build a runway, and reskill before you have to. The Clankers 90-day programme structures exactly that — free, which is exactly why today is the right time to start.
Start the 90-Day Programme — Free →GET THE CLANKERS BRIEF — FREE
One email every Monday: the week's AI displacement moves that matter, and what to do about them. No spam. Unsubscribe anytime.