In the race to automate everything, some companies have learned the hard way that replacing humans with AI isn’t always the upgrade it promises to be. From customer service meltdowns to translation disasters, here are a few recent cases where businesses hit “undo” on their AI deployments and brought humans back into the loop.
Klarna’s Customer Service Backfire
Industry: Fintech
Klarna, the buy-now-pay-later giant, made headlines in 2024 when it laid off hundreds of employees and leaned heavily into AI. CEO Sebastian Siemiatkowski claimed that AI could do “all of the jobs that we, as humans, do.”
Once they started using AI mainly for customer support jobs, customers began complaining about robotic responses, unresolved issues, and a lack of empathy. Klarna’s brand took a reputational hit, and by mid-2025, the company quietly began rehiring humans, especially for remote support roles. Siemiatkowski later admitted that cost-cutting had overshadowed quality, saying, “What you end up having is lower quality.”
Duolingo’s Translation Trouble
Industry: EdTech
Duolingo, the language-learning app known for its quirky owl mascot, decided to replace many of its human translators and content reviewers with AI. The goal? Faster content creation and lower costs.
But language is messy, and AI struggled with nuance. Users flagged awkward phrasing, cultural insensitivity, and even grammatical errors. While Duolingo hasn’t fully reversed course, internal sources say the company is reconsidering its reliance on AI and exploring hybrid models with human oversight.
Commonwealth Bank of Australia (CBA) AI Troubles
Industry: Banking
In August 2025, the Commonwealth Bank of Australia (CBA) replaced 45 call center employees with AI-powered voice bots in an attempt to cut costs and boost efficiency, but the move quickly backfired. The bots struggled to resolve customer issues, call volumes increased, and service quality dropped, forcing team leaders back on the phones to handle the overflow. The backlash from both customers and employees was strong, and CBA ultimately reversed course, offering to rehire the laid-off staff and admitting that the decision had underestimated the value of human judgment and empathy in customer service interactions.
The AI Agent Letdown
Industry: Cross-sector
Across industries, companies have been experimenting with AI agents, autonomous bots designed to handle tasks like scheduling, customer queries, and internal workflows. But according to a recent Gartner report, these agents could only complete 24% of assigned jobs.
Executives who once planned to cut customer service roles are now backpedaling. Gartner predicts that by 2027, half of those companies will abandon their AI-first plans and reinvest in human talent. Turns out, empathy, improvisation, and context aren’t so easily coded.
The Bigger Picture
These reversals aren’t just about technical glitches, they’re mostly about trust, nuance, and the irreplaceable value of human judgment. AI can be a powerful tool, but when it’s treated as a wholesale replacement, companies risk alienating customers and damaging their brand.
As one tech analyst put it, “AI is great at doing things. Humans are great at understanding things. The best systems combine both.”
And maybe, the future isn’t about choosing between humans and machines but figuring out how they can work together without one pretending to be the other.
