Klarna reintroduces humans after AI support both sucks, and blows

Tombstone icon

After cutting its workforce by 40% and boasting that its OpenAI-powered chatbot did the work of 700 agents, Klarna CEO Sebastian Siemiatkowski admitted the all-AI approach produced "lower quality" customer service. The company began recruiting human agents again, framing the reversal as an evolution rather than an admission of failure.

Incident Details

Severity:Facepalm
Company:Klarna
Perpetrator:Executive
Incident Date:
Blast Radius:Service quality/customer experience issues; operational/personnel cost; reputational damage.

Klarna is a Swedish buy-now-pay-later fintech serving approximately 150 million customers. In 2023, CEO Sebastian Siemiatkowski struck a partnership with OpenAI and told Sam Altman he wanted Klarna to become OpenAI's "favorite guinea pig." Over the next two years, Siemiatkowski made Klarna the most visible corporate test case for replacing human workers with AI at scale.

The 700-Agent Announcement

In February 2024, Klarna published a press release announcing that its OpenAI-powered AI assistant had handled 2.3 million customer service conversations in its first month of global operation. The chatbot was handling two-thirds of all customer service chats. Klarna claimed customer satisfaction ratings were at the same level as human agents, and stated the assistant was "doing the equivalent work of 700 full-time agents."

The number 700 was not an abstract projection. Between 2022 and 2024, Klarna eliminated approximately 700 positions, primarily in customer service and support. The company's total headcount fell from 5,527 full-time employees at the end of 2022 to 3,422 by the end of 2024 - a roughly 40% reduction, driven by a combination of AI investment and natural attrition through a companywide hiring freeze.

Siemiatkowski was not subtle about the strategy. He told interviewers the company would continue shrinking through attrition as AI absorbed more work. In December 2024, Klarna revealed it had hired no human workers for the entire preceding year, with the CEO boasting that "people internally at Klarna are just rallying to deploy as much efficiency AI as they can."

The company also cut ties with Salesforce and Workday, replacing both platforms with AI-powered internal systems. Siemiatkowski publicly framed the move as part of his vision for an AI-run company, one where enterprise software vendors were unnecessary overhead.

The Quality Problem

The narrative held through early 2025, powered by Klarna's pre-IPO publicity cycle. Then Siemiatkowski told Bloomberg something different. In a May 2025 interview, he acknowledged that Klarna's full pivot to AI-based customer support had resulted in "lower quality" work. The company would begin recruiting human customer service agents to work "in an Uber type of setup" - freelance-style arrangements rather than full-time hires.

He told CNBC the same month that AI had helped the company shrink its workforce by about 40%, framing it as an efficiency gain. But the Bloomberg admission was harder to spin: the AI-first approach to customer service had produced worse outcomes, and the company was reversing course.

A Klarna spokesperson told Fortune the company was maintaining its policy of not replacing employees who leave, apart from hiring freelance customer service agents through its outsourcing division, adding "we're very much still AI-first." The statement was technically accurate and practically meaningless - Klarna was hiring humans to do customer service work because the AI couldn't handle it, regardless of what label the company put on the arrangement.

The Rehiring

By September 2025, the reversal was more concrete. Business Insider reported that Klarna was reassigning internal workers to customer support roles after quality concerns. CX Dive reported the company was "again recruiting humans for customer service," over a year after claiming the AI chatbot had made those roles redundant.

Klarna's head of communications, David Nordstrom, framed the shift carefully: "While Klarna pioneered the use of AI in customer service with groundbreaking results, this strategy will now evolve to the next level." The language positioned the rehiring as a strategic evolution rather than a correction.

External analysts were more direct. Julie Geller, principal research director at Info-Tech Research Group, said the takeaway was straightforward: "AI should augment human agents - not replace them."

The IPO Complication

Klarna's AI story was central to its IPO narrative. The company went public in July 2025, with shares surging 30% on their debut and the company reaching a post-IPO valuation of $19.65 billion. The AI-driven efficiency gains - fewer employees, lower costs, consistent customer handling - were a selling point for investors.

The timeline creates an awkward sequence. Klarna pitched itself to public market investors partly on the strength of its AI-first customer service model. Within months of going public, the CEO was admitting that model produced lower-quality outcomes and the company was hiring humans back. Investors who bought into the AI efficiency story in July were learning by September that the company was quietly undoing the strategy.

Siemiatkowski acknowledged that Klarna had "overestimated AI's capabilities and underappreciated the human aspects of service delivery." This is a fair assessment, but it's also a remarkable admission from a CEO who spent two years as the most vocal corporate champion of replacing human workers with AI. The 700-agent press release in February 2024, the hiring freeze, the Salesforce and Workday departures, the "no human hires for a year" boast - these were not cautious experiments. They were public declarations that AI had made human customer service workers obsolete.

The Broader Pattern

Klarna's reversal landed during a period of increasing skepticism about AI customer service. An IBM survey referenced by Fortune found that just one in four AI projects delivers the return it promises, and even fewer are scaled up successfully. Gartner had separately predicted that by 2027, half of companies that cut customer service staff because of AI would need to rehire.

The Klarna case is particularly instructive because the company went further and faster than most. This was not a pilot that quietly fizzled. It was a flagship corporate AI strategy, championed by the CEO at every opportunity, executed across the entire customer service operation, and linked to the company's public market debut. And it produced customer service that was, by the CEO's own admission, of lower quality than what the human employees it replaced had delivered.

The freelance "Uber type of setup" Siemiatkowski described for the rehired human agents is worth noting. The workers who were eliminated were full-time employees. The workers being brought back are freelancers. Klarna cut 700 permanent positions, found out AI couldn't replace them, and is now refilling the gap with contract workers who presumably lack the benefits, job security, and institutional knowledge that full-time employees had.

Whether that qualifies as "evolving to the next level" depends on your definition of the word "evolution."

Discussion