Klarna's AI Customer Service Backlash: What They Learned the Hard Way

Klarna's AI customer service rollout initially looked like a textbook success. Its autonomous agent handled 2.3 million conversations in its first month, resolved two-thirds of all tickets, and cut average resolution times from 11 minutes to just two. The buy now, pay later giant estimated the technology could add $40 million to annual profit. As AI took on more work, Klarna froze hiring and reduced its customer service workforce from around 5,000 to 3,500.

Then the warning signs appeared. A little over a year later, CEO Sebastian Siemiatkowski acknowledged that Klarna had cut too far, too quickly and lost valuable human expertise. The company began rebuilding its capacity for human support.

Klarna is not alone. Ford has also acknowledged being overly optimistic about how quickly AI could enable workforce reductions, suggesting other businesses may be making similar miscalculations.

What Do Klarna’s Customer Service Agents Do?

Klarna was an early partner of ChatGPT creator OpenAI and moved quickly to integrate natural language into its customer service. But rather than a simple chatbot, it built the offering around agentic AI. This means it doesn’t just answer questions—it can work on complex, multi-stage tasks with minimal human involvement, such as securely accessing customer data, monitoring ticket status, issuing returns, and managing payment plans. These are high-volume, repetitive tasks where AI excels.

The results were impressive: a 25% reduction in repeat requests and an estimated $40 million annual profit boost.

Human Intervention and Edge Cases

Despite high headline statistics, Klarna couldn’t ignore vocal customer dissatisfaction. For complex issues involving ambiguity, emotional responses, or edge-case technical problems, the AI agents fell short. With a 30% reduction in human staff, fewer experienced operators were available to handle these cases, leading to poorer outcomes. Satisfaction scores dropped.

Siemiatkowski admitted the company had been too aggressive with cuts and identified a strategic need to reinvest in well-trained human support workers for high-value issues.

The U-turn was widely reported as an "AI failure," but this overlooks an important fact: by most definitions, Klarna’s AI rollout was a big success. However, overlooking niche situations where AI is less capable led to sub-optimal results.

What Can Leaders and Professionals Learn?

The key lesson: AI is not a magic bullet. Even when benefits are clear, we must consider edge cases and where humans still need to be involved. By identifying its failure and acting to resolve it, Klarna learned to adapt its strategy.

Agents are built for routine, repetitive tasks with clear instructions. But real-world business is too complex, nuanced, and subjective for today’s best AIs to reliably handle. The reason agents failed in edge cases is the same reason autonomous driving isn’t yet commonplace: humans remain critical.

Businesses must chase efficiency, but when it leads to an exodus of skill and talent before there’s total confidence in the replacement, it’s likely to backfire. This means carefully retaining human skills and experience while developing future capabilities.