Klarna’s AI Pivot: What Happens When You Automate Empathy?
In early 2024, Klarna did what many tech-first companies dream of.
It replaced 700 customer service agents with AI.
Efficiency skyrocketed.
Resolution times dropped from 11 minutes to under 2.
Customer satisfaction remained steady.
And executives projected a $40 million boost to profits.
It was hailed as a masterclass in automation. A model for others to follow.
Until it wasn’t.
The Problem Wasn’t the AI
Klarna’s chatbot was fast, multilingual, always available.
It even reduced repeat inquiries by 25%—a sign that issues were being resolved more accurately the first time.
But something started to break.
Not in the backend.
In the relationship.
Customers began voicing dissatisfaction—not about the answers they received, but about how it felt to be served.
No empathy. No flexibility. No real connection.
As Klarna’s CEO admitted:
“Cost unfortunately seems to have been a too predominant evaluation factor… what you end up having is lower quality.”
They’d won the speed game.
But lost the trust game.
Substitution Thinking vs. Co-Intelligence
At The Helix Lab, we call this the trap of substitution thinking—optimizing for linear efficiency while breaking the strategy-design helix. It’s the belief that AI exists to replace humans in "repetitive" roles. But as Microsoft’s Satya Nadella proved, "Design thinking is empathy". Klarna’s lapse wasn’t an AI failure—it was a design failure to intertwine human and machine rhythms.
This brittle approach prioritises:
Tasks over relationships
Speed over trust
Accuracy over understanding
The future isn’t substitution.
It’s co-intelligence:
Machine efficiency + human empathy—amplifying each other within the Helix Blueprint.
The Return of the Human Touch
By 2025, Klarna began re-hiring agents—not to undo AI gains, but to rebalance the helix. Their new hybrid model embodies multi-loop co-intelligence:
AI loops: Handle routine requests at machine speed
Human loops: Navigate complex, emotional, and nuanced issues
Choice architecture: Customers can always request a person
Uber-like flexibility: Remote agents (students/rural workers) join dynamically
This isn’t a tactical fix—it’s a strategic reset. When service becomes relational (not transactional), humanity becomes infrastructure.
The Real Lesson for Leaders
AI is powerful—but it doesn’t build trust or read unspoken frustration. Those are human superpowers.
If you lead in any trust-driven sector, Klarna’s story is your design brief:
Don’t ask: “What can AI replace?”
Ask:
1. Where do humans create unique value?
2. How can AI amplify that value—not erase it?
3. Are we building systems that are both smart and kind?
As Klarna’s Clare Nordstrom acknowledges:
“AI provides speed, while humans offer empathy. Together, we deliver promptness when necessary—and compassion when required.”
That’s not a quote. It’s the operating rhythm for the future of work.
Suhit’s upcoming book The Helix Blueprint provides more insight into how we use co-intelligence in strategy and design. Reach out at suhitanantula.com