Inside Klarna's Methodical Replacement of Its Customer Service Workforce
Klarna disclosed the number in its IPO prospectus filed with the SEC in July 2025: the Swedish payments company had reduced its total headcount from roughly 5,000 employees to approximately 3,500 over the preceding two years, a contraction it attributed in substantial part to its deployment of an in-house AI customer service assistant. That assistant, the company said, was handling the volume of work previously performed by 700 human agents. The figure was not buried in a footnote. Klarna put it in the section designed to attract investors.
The prospectus framing was careful. The word "replacement" did not appear. The company described a "leaner" operation, improved unit economics, and a customer satisfaction score it claimed was equivalent to that achieved by its human workforce. What the document did not describe in detail was what happened to the people who had staffed those roles — where they went, what severance arrangements were made, and whether the displacement tracked with what Klarna had said publicly in the years leading up to it.
The timeline is worth reconstructing. In February 2024, Klarna's CEO Sebastian Siemiatkowski announced in a post on X that the company's AI assistant — built in collaboration with OpenAI — had completed its first month handling customer inquiries at scale. He said it had managed 2.3 million conversations, representing two-thirds of Klarna's customer service chat volume at the time. He said it was doing the work of 700 full-time agents. He said customer satisfaction scores were on par with human agents. He said it was available in 23 markets and 35 languages, around the clock. The post was widely reported, including by Reuters and the BBC.
What was less widely reported was that Klarna had been simultaneously winding down contracts with outsourced customer service providers. A significant portion of Klarna's customer service workforce was not directly employed by the company but contracted through third-party firms — an arrangement common across the fintech sector. That structural detail meant the headcount reduction was partially invisible in Klarna's own employment figures, distributed instead across supplier companies in markets including the United States, Germany, and the Philippines.
"The displacement happened at the edges of the org chart," one former outsourcing industry analyst told a Swedish business publication in early 2025. "It doesn't show up cleanly in any single filing."
The scale of that outsourced component has not been precisely quantified in any public disclosure. Klarna declined to provide figures when asked by reporters following the IPO filing. What the SEC prospectus does confirm is the aggregate outcome: a workforce reduction of approximately 30 percent over roughly two years, against a backdrop of expanding revenue. Klarna reported net revenue of approximately $2.8 billion for 2024, up from $2.3 billion the year prior. The company was processing more transactions with substantially fewer people.
What the Numbers Actually Show
The Klarna case sits within a pattern documented across the financial services sector. Goldman Sachs Economics research published in 2023 estimated that roughly 300 million full-time jobs globally were exposed to automation from generative AI, with clerical and customer-facing roles among the most concentrated areas of risk. The World Economic Forum's Future of Jobs Report 2025 identified customer service and call center roles as among the positions employers expected to reduce most sharply through 2030, with AI tools cited as the primary driver.
What made Klarna notable was not simply the scale of the reduction but its speed and its transparency. Most companies that have restructured customer service operations through AI deployment have not provided explicit attribution — they have cited "efficiency improvements," "process optimization," or simply announced layoffs without naming the mechanism. Klarna, by contrast, volunteered the 700-agent figure in a press release and then repeated it in a securities filing. The motive was investor relations: demonstrating AI leverage was directly relevant to the company's valuation thesis ahead of a public offering that ultimately valued the business at approximately $15 billion when it listed on the New York Stock Exchange in July 2025.
That transparency creates an unusually clear record. It also makes Klarna a test case for whether disclosure frameworks are equipped to capture what is actually happening to workers when companies automate at this pace. The SEC prospectus describes human capital resources in terms of total headcount and benefits programs. It does not require companies to report the occupational composition of workforce reductions, the proportion attributable to automation versus business contraction, or the outcomes for displaced workers. Those questions remain outside the disclosure perimeter.
Aria Chen, who built the displacement tracking resource clankers.today after her own exit from a software engineering role in 2024, has cited Klarna repeatedly as a documentation benchmark — not because the company's actions were unusual, but because the paper trail is unusually complete. "Most of the time we're inferring," she said in an interview earlier this year. "With Klarna, they told you."
The Satisfaction Score Question
One claim in Klarna's February 2024 announcement has drawn sustained scrutiny: that its AI assistant achieved customer satisfaction scores "on par" with human agents. Siemiatkowski repeated a version of this claim in subsequent media appearances and in investor communications. The prospectus reiterated it without providing the underlying methodology.
Independent researchers have noted the difficulty of evaluating such comparisons without access to the measurement instrument, the baseline period, and the customer population sampled. Customer satisfaction in service contexts is sensitive to channel selection — customers who choose to interact via chat may differ systematically from those who prefer phone, and AI deployments are typically concentrated in text-based channels where measurement conditions differ from voice interactions historically staffed by human agents. None of these methodological details have been disclosed publicly.
The Stanford AI Index 2025 flagged the proliferation of corporate AI performance claims that lack reproducible methodology as a broader credibility concern across the industry. Klarna's claim is not unique in this respect, but given its prominence in a securities filing, it has attracted particular attention from analysts who cover both the company and the AI deployment space.
Klarna has not responded to detailed questions about its measurement approach from multiple outlets, including this one. Its investor relations materials describe the assistant as "continuously improving" and note ongoing investment in AI capabilities as a central element of its product roadmap.
What is not in dispute is the arithmetic. Klarna entered 2023 with a larger workforce and exited 2024 with a smaller one, while growing its revenue and explicitly naming an AI system as the agent of that compression. The 700-agent figure the company disclosed is not an estimate or a projection. It is the company's own accounting of what its technology replaced. The workers behind that number had jobs in customer service. Then, by Klarna's own telling, they did not. The prospectus moved on to the next section. The question of what happened to them did not make it into the filing at all.
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.