Klarna’s AI customer service agent is now doing the work of 853 full-time employees. The company has simultaneously cut its workforce in half over three years and doubled revenue to $903 million a quarter.
If you’re an executive trying to make sense of where AI ROI actually shows up, this case study should be at the top of your reading list. It’s the cleanest example we have of AI doing something most companies have not yet absorbed: breaking the historical link between revenue growth and headcount growth.
Here’s what Klarna did. They partnered with OpenAI to build an AI customer service agent and integrated it directly into their primary support flow. The agent handles inquiries in 35 languages, has full context on each customer’s account, and is empowered to actually resolve cases – not just suggest answers. Today the AI handles approximately two-thirds of Klarna’s total customer inquiries. It resolves cases in under two minutes on average, an 82% improvement over prior response times. Cost per support transaction has dropped 40% in two years, from $0.32 to $0.19. Total savings on customer service alone: more than $60 million.
But the financial story doesn’t stop at customer service. Pull back the lens. Over the past three years, Klarna has reduced its overall workforce by roughly 50%. Over that same period, quarterly revenue has more than doubled, hitting $903 million in the most recent reported quarter – up from $433 million in the equivalent quarter three years earlier. Active users grew 32% year over year to 114 million.
Why This Actually Worked
Three things separate this from the long list of failed enterprise AI deployments.
First, aggressive routing beat cautious piloting. Most large companies test AI on 1-5% of traffic for nine to twelve months. Klarna pushed real traffic into the agent quickly, then expanded as performance held. AI agents get better with volume – they need real-world cases to be tuned against. The companies that pilot too conservatively starve their AI of the data it needs to improve, then conclude AI ‘doesn’t work.’
Second, full authority to act. The Klarna agent isn’t a search bar with a personality. It can refund a charge, change a payment plan, dispute a transaction, escalate to a human when needed. AI without authority is a chatbot. AI with authority is a worker.
Third – and this is the part most people miss – Klarna didn’t just save money. They broke a structural assumption. For decades, software companies assumed support cost grew roughly with user count. Klarna proved that assumption is now optional. AI uncoupled their cost-of-service curve from their growth curve. Once you decouple those two lines, the unit economics of the entire business change.
I’m Mike Partners, and I started AiExpert.org to bridge the gap between enterprise AI strategy and small business reality. Here’s your action plan.
How to Apply This to Your Business
Start by looking at your customer support workflow. You don’t need enterprise AI to make a difference here – tools like Intercom, Freshdesk, or even a well-configured chatbot on your website can handle the routine questions that eat up your team’s time. Track how many hours per week go to answering the same five questions, then automate those first. Once you see the time savings, you’ll naturally find the next process to streamline. The goal isn’t to replace your team – it’s to free them up for the conversations that actually require a human touch.
The SMB Playbook
If you run a business with anywhere from 10 to 500 employees, the principle ports directly. Here’s how to translate Klarna’s move into your scale.
1. Identify the function that currently scales 1:1 with customer count. Customer support is the obvious one. Onboarding, scheduling, FAQ, basic account management – anything where every new customer adds a unit of work to a human’s plate. That’s where your AI deployment should land first.
2. Before your next hire in that function, run the substitution test. Ask: ‘could an AI agent absorb the bottom 60% of this role?’ If yes – and for repetitive customer-facing work, the answer is almost always yes today – deploy the agent first. Hire the human to do work the AI can’t.
3. Route aggressively. Don’t run a six-month 5% pilot. Pilot for two weeks at 5%, then move to 30% if it holds, then 60%. AI agents improve with volume. Pilot starvation is the #1 killer of enterprise AI. Klarna’s instinct here was right and most companies’ instinct is wrong.
The thing to internalize from Klarna isn’t the dollar figure. It’s the decoupling. For most of business history, scaling revenue meant scaling headcount. AI changes that math at every level – from a 14-person services firm to a 14,000-person platform. The companies that build their next year of growth around this decoupling will compound. The ones that keep hiring linearly with customer count will look up in 24 months and wonder why their margins eroded.
Frequently Asked Questions
How is Klarna using AI in 2026?
Klarna has deployed AI across multiple areas of its operations, focusing on automation, cost reduction, and efficiency gains. As covered in this analysis by Mike Partners, the results include measurable improvements in both operational metrics and financial performance, demonstrating that strategic AI deployment delivers real business returns.
Can AI replace customer service agents at companies like Klarna?
AI is not fully replacing customer service agents, but it is handling a significant portion of routine inquiries. Klarna’s approach shows that AI works best when it handles high-volume, predictable requests while human agents focus on complex cases requiring empathy and judgment.
How can small businesses apply the same AI strategies as Klarna?
Small businesses can apply similar principles by starting with their most repetitive, time-consuming processes and finding affordable AI tools to automate them. Resources like AiExpert.org break down enterprise AI strategies into actionable steps sized for smaller companies, so you do not need a Fortune 500 budget to benefit from these approaches.
What is the ROI of AI automation for businesses in 2026?
ROI varies by implementation, but the pattern across major deployments is consistent: companies are seeing 20-40% cost reductions in automated processes, significant productivity improvements per employee, and faster decision-making cycles. The key driver of ROI is not the technology itself but how strategically it is deployed against the business’s highest-cost, most repetitive operations.
What AI tools should I use to automate my business like Klarna?
The right tools depend on your specific business needs. For customer-facing automation, look at chatbot platforms and AI-powered support tools. For operations, explore workflow automation platforms like Zapier or Make. For content and marketing, tools like ChatGPT, Jasper, or Claude can accelerate production. Start with one area, measure results over 30 days, and expand from there.
