AI for Business · AI Case Studies
What actually happened when Klarna replaced customer service agents with AI
Klarna's AI assistant initially handled the workload of roughly 700 customer service employees and two-thirds of all queries, but by 2025 the company was rehiring human agents after customer satisfaction dropped, settling into a hybrid model where AI handles routine queries and humans handle complex or high-value cases.
Key takeaways
- Klarna's AI customer service assistant, built with OpenAI, initially handled roughly two-thirds of all customer service queries.
- The company had reduced its customer service headcount by around 700 roles as part of the initial rollout.
- By 2025, Klarna's CEO publicly acknowledged that the cost-cutting focus had led to lower service quality and began rehiring human agents.
- The company settled into a hybrid model: AI handling routine, high-volume queries, and human agents handling complex cases and escalations.
The Initial Rollout
Klarna gained widespread attention for deploying an AI customer service assistant, built in collaboration with OpenAI, that reportedly handled roughly two-thirds of all customer service queries — a workload equivalent to what around 700 human customer service employees had previously handled, with those roles reduced as part of the transition.
What Went Wrong
Over time, cracks appeared in the fully AI-driven approach — customer satisfaction metrics declined, and by 2025 Klarna’s CEO publicly acknowledged that the initiative’s heavy emphasis on cost-cutting had come at the expense of service quality, with some customer interactions suffering from the loss of human judgment and empathy in more complex situations.
The Reversal
In response, Klarna began actively rehiring human customer service agents, walking back the scale of its original AI-only approach rather than abandoning AI in customer service altogether — the reversal specifically targeted the areas where full automation had proven inadequate, not the AI usage as a whole.
Where Klarna Landed
By 2026, Klarna had settled into a hybrid model: AI continues to handle routine, high-volume queries efficiently, while human agents specifically handle complex cases, escalations, and situations genuinely requiring empathy or judgment — a structure that keeps AI’s efficiency gains while restoring human involvement where it clearly mattered most.
Bottom Line
Klarna’s case has become a widely cited example of the limits of full AI automation in customer service — the company’s shift from an aggressive all-AI approach back to a deliberate AI-plus-human hybrid model illustrates that efficiency gains from automation can come with real service-quality tradeoffs that eventually need correcting.
Estimate Your Time Savings
See how many hours and dollars using AI for a repeated task could save you with our free AI Time-Savings Calculator.
Go deeper
Frequently asked questions
Did Klarna abandon AI in customer service entirely?
No — Klarna kept AI handling routine, high-volume queries and specifically rehired human agents for complex cases and escalations, landing on a hybrid model rather than a full reversal back to entirely human-staffed support.
Related questions
- How Did Chipotle Use AI to Cut Its Hiring Time by 75%?
- How Is Morgan Stanley Using AI to Help Its Financial Advisors?
- How Does Walmart Use AI to Manage Its Supply Chain and Inventory?
- How Did Duolingo Use AI to Build 148 New Courses in a Year?
- Can AI Chatbots Fully Replace Human Customer Support?
- What role should legal counsel play before a company deploys a customer facing ai tool?
Sources
- [1]Klarna tried to replace its workforce with AI — Fast Company
- [2]Klarna backtracks AI customer service plans — EMARKETER
Written by Editorial Team
Last updated August 8, 2026
Get one well-sourced answer a week
No spam. Unsubscribe anytime.