In what tech industry analysts are calling the “most expensive ‘no sh*t, Sherlock’ moment in fintech history,” Swedish buy-now-pay-later giant Klarna has made the ground-breaking shocking discovery that humans are better at being human than artificial intelligence. After boldly replacing 700 customer service representatives with AI chatbots two years ago, the company has sheepishly announced plans to rehire actual people, citing the shocking revelation that algorithms struggle with emotional intelligence, complex problem-solving, and not making customers want to throw their devices into the ocean.
CEO Sebastian Siemiatkowski, who previously declared that “AI can already do all the jobs that we, as humans, do,” has recalibrated his position slightly to acknowledge that perhaps sentient beings with actual feelings might have some minor advantages when dealing with emotionally distraught customers who can’t pay their installments for that impulse-purchased Pizza at 1 AM.
The Golden Age of AI-Enhanced Unemployment
Klarna’s journey toward digital enlightenment began in 2022 when the company formed a partnership with OpenAI, eagerly positioning itself as “OpenAI’s favorite guinea pig,” a description that has aged about as well as milk left in a car during an Indian hot summer. The company immediately embarked on what executives called a “staffing optimization strategy” and what everyone else called “firing people via pre-recorded videos.”
By 2023, Klarna had implemented a complete hiring freeze and boasted that its AI was performing work equivalent to 700 customer service representatives. The company proudly announced $10 million in marketing cost savings as AI handled tasks such as translation, art production, and data analysis—all tasks requiring the creativity and emotional intelligence that silicon-based entities are famously good at.
Siemiatkowski celebrated by declaring that the company had reduced its workforce from 5,527 to around 3,000 employees—a 40% reduction—all while continuing to provide what executives called “adequate customer service” and what customers called “a Kafkaesque nightmare of circular logic and canned responses.”
Unforeseen Complications: AI Cannot Yet Feel Your Financial Pain
The honeymoon period of Klarna’s AI revolution came to an abrupt end when the company made a shocking discovery: customers actually prefer talking to beings capable of empathy when discussing their financial struggles. In what must have required extensive research and billions of data points to determine, Klarna’s leadership team concluded that AI chatbots—despite their impressive ability to parse language and generate responses—somehow lacked the emotional intelligence needed to properly handle a customer calling in tears because they accidentally signed up for buy-now-pay-later on groceries and can’t make the payments.
“From a brand perspective, a company perspective, I just think it’s so critical that you are clear to your customer that there will be always a human if you want,” Siemiatkowski recently told Bloomberg, apparently having experienced an epiphany that customer service might benefit from something resembling a soul.
The CEO further admitted that “cost unfortunately seems to have been a too predominant evaluation factor when organizing this, what you end up having is lower quality,” a statement that has been nominated for the 2025 “No Sh*t” Awards alongside “water is wet” and “tech CEOs sometimes overestimate technology.”
AI Achievements: A Balanced Assessment
To be fair, Klarna’s AI initiative wasn’t a complete disaster. According to the company’s numbers, their revenue per employee has skyrocketed from $575,000 to nearly $1 million, which proves definitively that firing large portions of your workforce does wonders for per-employee metrics. The company’s AI chatbots also excelled at several tasks, including:
- Confidently providing incorrect information with perfect grammar
- Misinterpreting customer emotions with remarkable consistency
- Responding to complex questions with irrelevant solutions
- Maintaining the same cheerful tone when explaining why your payment was declined as when wishing you a nice day
- Never requesting bathroom breaks, healthcare, or a living wage
In a particularly noteworthy achievement, Klarna’s AI customer service system managed to regularly convert “slightly annoyed” customers into “incandescently furious” customers in record time—a transformation that typically requires years of training for human representatives.
The Leadership Accountability Paradox
In a fascinating display of corporate physics that defies conventional laws of causality, the decision to fire 700 people and replace them with inadequate AI—described by one internal memo as “a strategic misstep of historic proportions”—has somehow resulted in zero leadership terminations. Siemiatkowski remains firmly at the helm, demonstrating the remarkable principle that in modern corporate structures, accountability flows downward but never upward!
Industry analyst Margareta Lindström explains: “It’s quite incredible. If a customer service representative fails to resolve three customer issues, they receive a performance improvement plan. If a CEO makes a decision that wastes millions of dollars, destroys customer relationships, and requires a complete strategic reversal two years later, they get to announce the new strategy as if it were their idea all along.”
This curious phenomenon, which physicists are calling “the executive accountability vacuum,” suggests that at certain levels of corporate hierarchy, the normal rules of professional consequence cease to apply entirely. Scientists are currently studying whether this effect could be harnessed as an alternative energy source.
The Rehabilitation Phase: “We’ve Always Valued Humans, Starting Now”
Rather than simply admitting error and rehiring the people they laid off, Klarna has announced a bold new “human-in-the-loop” customer service strategy that resembles an Uber-style gig model. The company plans to recruit students and people in rural areas to work remotely on an as-needed basis, a model that executives describe as “innovative” and labor experts describe as “exploitative nonsense.”
“We’re not going back to the old ways,” explained Siemiatkowski in a recent interview. “We’re moving forward with a hybrid model that combines the best of AI with the best of human capability, while maintaining the worst of gig economy employment practices.”
The new system will allow customers to speak with actual humans when they require assistance beyond the capabilities of AI, such as understanding tone, context, or basic human empathy. Meanwhile, the AI will continue handling simpler tasks, primarily directing customers to the human representatives who can actually help them.
The Silent Industry-Wide Recalibration
Klarna isn’t alone in its AI humbling. According to a January 2024 survey of 1,400 executives, widespread dissatisfaction with AI integration is common, with many citing underwhelming results. In the UK, a survey revealed that 55% of business leaders who had replaced humans with AI regretted the decision, though most would rather walk barefoot on LEGO bricks than publicly admit it.
Tech industry analyst Henrik Johannsson notes that many companies are quietly recalibrating their AI strategies: “What we’re seeing across the board is a silent retreat from the ‘AI can replace everyone’ position. The new narrative is ‘AI enhances human capability’ rather than replaces it. It’s the corporate equivalent of saying ‘I meant to do that’ after tripping in public.”
This strategic pivot is reflected in job postings across the tech sector, where roles once proudly advertised as “AI-replaceable” are now being rebranded as “AI-enhanced,” “AI-collaborative,” or “human-essential.” The industry has smoothly transitioned from “AI will replace all humans” to “we always meant AI would be a tool for humans” without acknowledging the contradiction.
The Science of Customer Service: Emotions Required
The fundamental issue underlying Klarna’s AI misadventure is one that computer scientists have understood for decades: emotional intelligence cannot be synthesized through algorithms alone. A comprehensive analysis of AI customer service failures at Klarna, obtained exclusively by TechOnion, revealed the following issues:
- Rigid and unhelpful answers that didn’t fully address customer queries. One customer reported asking for help with a payment issue and receiving instructions on how to download the Klarna app—which they were already using to make the complaint.
- Lack of conversational flow, forcing users to rephrase questions multiple times. In one documented case, a customer had to rephrase the same question about a refund in seven different ways before the AI understood, at which point it referred them to a nonexistent department.
- Misinterpretation of complex inquiries, causing unnecessary escalations. The system frequently misunderstood emotional cues, once interpreting a customer’s sarcastic “Thanks for nothing” as sincere gratitude and responding with “You’re welcome! Is there anything else I can help you with today?”
While AI excels at handling repetitive tasks, research consistently shows that human customer service representatives outperform AI in empathy, problem-solving, and building trust. As one customer service expert put it: “Turns out the ‘service’ part of ‘customer service’ benefits from understanding what it means to be a human who is frustrated, confused, or on the verge of throwing their phone across the room.”
The Great Rehiring: Gig Economy Edition
Klarna’s solution to their AI customer service disaster represents a masterclass in admitting failure while refusing to actually fix the problem. Rather than returning to a stable workforce of full-time customer service representatives, the company is implementing what it calls a “flexible human assistance model” that industry critics are calling “Uber but for helping people who are angry at AI chatbots.”
Under this new system, workers will log in when they want and take customer service calls on demand—a model that conveniently shifts scheduling risk from the company to the worker while maintaining the fiction of “flexibility.” The company estimates this will save them approximately 30% on benefits, paid time off, and other inconvenient aspects of traditional employment, while providing workers with the “freedom” to work whenever they are desperately in need of income.
“It’s really a win-win,” explained a Klarna spokesperson who definitely exists. “We get human intelligence without human employment costs, and workers get to experience the thrill of never knowing if they’ll make enough money to pay their bills this month.”
As the digital dust settles on Klarna’s great AI experiment, the company finds itself exactly where critics predicted it would be two years ago: acknowledging that customer service requires actual humans with actual empathy. The only differences are the millions of dollars wasted, the 700 careers disrupted, and the irreparable damage to customer relationships.
The company does, however, have one genuine innovation to show for its efforts: it has conclusively demonstrated that when a corporation’s leadership makes catastrophic decisions based on overinflated tech promises, the consequences flow exclusively to the workers, customers, and shareholders—never to the decision-makers themselves.
What do you think? Have you had your own nightmare experiences with AI customer service? Are you one of the lucky 700 who might get to return to Klarna as a gig worker? Or are you an executive who’s currently planning to replace your workforce with AI despite all evidence suggesting it’s a terrible idea? Share your thoughts in the comments below, where a sophisticated AI will pretend to read them before forwarding the interesting ones to an underpaid human moderator.
Support TechOnion’s Human-Written Content
If this article has made you appreciate the value of actual humans writing actual content, consider supporting TechOnion with a donation of any huge size. Unlike our AI competitors, we require food, shelter, and occasionally therapy after interviewing tech CEOs. Your contribution helps us continue employing carbon-based writers who can experience emotions like skepticism, amusement, and existential dread—all crucial ingredients for quality tech journalism. Rest assured that 100% of your donation will go toward keeping actual humans employed, unless our CFO decides to blow it all on an AI financial planning system that recommends investing in blockchain-powered fidget spinners.
