Klarna’s bold leap into AI-driven customer service has hit a human wall. Less than a year after replacing 700 staff with AI, the company is now backtracking, admitting that automation alone couldn’t deliver the quality customers expected.
In 2023, Klarna made headlines for replacing some 700 customer service employees with an AI assistant trained in collaboration with OpenAI. For a while, the numbers seemed to justify the move—the chatbot reportedly handled two-thirds of all customer queries, equivalent to the workload of its laid-off staff.
But by mid-2025, Klarna’s AI experiment has come full circle: the company is now rehiring human agents after customer satisfaction dropped and operational hiccups began surfacing.
“We focused too much on efficiency and cost,” Klarna CEO Sebastian Siemiatkowski admitted. “The result was lower quality, and that’s not sustainable.”
The decision marks a striking shift for a company once seen as a frontrunner in AI-driven operations. It also reopens a heated debate in the tech industry: just how far can automation go before it starts eating its own tail?
AI Replaces, Then Falters
The original move to AI-powered support was part of Klarna’s aggressive cost-cutting strategy following the fintech downturn of 2022–23. In a high-profile announcement, the company said its chatbot could conduct millions of conversations, citing speed, accuracy, and scale.
Yet customers weren’t convinced.
“It turns out that empathy, nuance, and complex resolution are still human strengths,” says Andriy Tatchyn, an expert in digital transformation at LaSoft. “The AI systems Klarna relied on were efficient—but not effective.”
By spring 2025, Klarna began piloting a new model: blending AI with human support. Dubbed an “Uber-style” workforce approach, the company is now hiring remote agents with flexible schedules—targeting students, parents, and even rural workers.
Why Automation Isn’t the Ultimate Answer
Klarna’s reversal highlights a broader problem: companies too often frame AI as a one-size-fits-all replacement, especially in customer service.
LaSoft previously analyzed this phenomenon in the article The Negative Impact of Artificial Intelligence on Employment, warning that cost savings from automation can be misleading when long-term brand damage and customer dissatisfaction are factored in.
“AI is a tool, not a team,” says Tatchyn. “But tech firms sometimes forget that distinction in the race to optimize every line item on a spreadsheet.”
The reality is that not all jobs are replaceable. Another LaSoft article, Jobs AI Can’t Replace, outlines roles that require emotional intelligence, ethical judgment, and adaptive problem-solving — all of which played a part in Klarna’s decision to rehire.
The New Hybrid Model
Klarna isn’t abandoning AI. Instead, it’s reshaping its vision of customer service into a human-AI partnership:
- AI handles basic inquiries and automates repetitive tasks.
- Human agents take over when issues require empathy, discretion, or escalation.
- Recruitment focuses on flexibility, diversity, and geographic inclusivity.
This model is already showing signs of improvement, and other companies may soon follow suit.
Lessons for the Tech Industry
Klarna’s backtrack isn’t a failure — it’s a recalibration. It offers a rare moment of humility in a sector often driven by hype. It also provides a valuable case study for companies racing to implement AI without fully considering its limits.
“Replacing people with machines sounds futuristic, until the machines start damaging the future of your business,” Tatchyn notes. “AI works best when it empowers people, not when it replaces them.”
In the end, the lesson is clear: technology may evolve, but trust, empathy, and judgment are still very human things—and they’re harder to scale than we thought.
