Part 1: Klarna’s AI Support Chatbot – A U-Turn in Progress
Swedish fintech Klarna had been vocal about the success of its AI chatbot, which reportedly handled 2.3 million conversations in its first month and claimed to match or surpass human agent performance in many metrics. However, less than a year later, the company is backpedalling—rehiring human agents and rebalancing its support workforce.
Why the reversal?
- Customer complaints: Reports emerged of increased frustration, unresolved queries, and conversations cut short by the chatbot.
- Regulatory scrutiny: European regulators, especially in Sweden and Germany, began questioning whether AI systems were compliant with consumer protection standards.
- Brand risk: Klarna’s customer base values trust and clarity, and the automated experience reportedly eroded both.
The takeaway: Customer-facing automation must meet a higher bar, not just in task completion, but in empathy, nuance, and escalation.
Root-Cause Analysis: When Chatbots Disappoint
- Limited context memory — Bots often failed to retain multi-turn intent.
- No escalation signal — Many interactions didn’t flag for human intervention when needed.
- Script bias — Users sensed canned replies, reducing satisfaction and trust.
- Tone mismatch — Even helpful responses felt “robotic,” especially for emotional or urgent cases.
In short: Customer support isn’t just about accuracy—it’s about experience.
Part 2: ABB’s PixelPaint – Robotic Automation at Its Best
On the factory floor, a very different story is playing out. ABB’s PixelPaint technology, now live in Mercedes-Benz’s S-Class manufacturing line, uses AI-guided robots to spray two-tone finishes with micron-level accuracy—delivering a 15% reduction in paint waste and eliminating the need for masking tape.
Key performance gains:
- Precision: 100µm spray resolution, far beyond human capability.
- Efficiency: Reduces setup and clean-up time dramatically.
- Sustainability: Minimises overspray and paint consumption.
Unlike customer service, paint application is a closed, repeatable process with defined tolerances and low emotional variability—making it a perfect candidate for full automation.
Use this framework to evaluate whether your automation roadmap is aligned with business goals—or creating more risk than reward.
Final Thought
Klarna and ABB offer a timely reminder that automation is not a one-size-fits-all solution. Where Klarna stumbled, ABB soared—not because one used AI and the other didn’t, but because each applied it in contexts that either magnified or constrained its effectiveness.
As companies explore AI agents, robotic workflows, and digital co-pilots, leaders must learn not just how to build and deploy them—but when to hold back.
Anjin Digital helps teams make smart, context-aware decisions about when and how to deploy AI agents—balancing automation with human oversight for maximum impact and trust. Learn more at anjin.digital.
