Just 5% of firms profit from AI — the rest are quietly rehiring workers

Companies that rushed to replace human workers with artificial intelligence are now quietly bringing them back, new data shows — a reversal that underlines the limits of automation hype.

A recent analysis by people analytics firm Visier found that 2025 saw the highest rate of employee rehires since 2018, with more companies reinstating staff they had previously terminated for AI-driven roles. Typically, about 5.3% of laid-off workers are re-employed by their former organisations, but that figure has climbed sharply this year, Axios reported.

The trend follows an MIT study revealing that only 5% of companies that fully embraced AI realised any measurable profit from their investment. Many firms appear to have overestimated the immediate cost benefits of replacing humans with algorithms — only to discover operational gaps, training costs, and inefficiencies that offset the expected gains.

Andrea Derler, Visier’s head of research and value, said the spike in rehires points to “AI-driven hiring and firing practices” that were poorly planned. “The idea that now AI is coming and replacing absolutely every job is still really not proven,” she told Axios.

Derler added that many executives underestimated the complexity of integrating AI into workflows and failed to weigh the true costs of implementation. “Leaders should ask: Which roles can be replaced by AI? What’s the cost, what’s the risk, and what’s the plan for the people and skills left behind?” she said.

Visier’s 2025 report, Embracing the AI-Driven Workforce, found that firms often lacked a strategic framework for reskilling or redeploying human talent after automation. “Many very senior executives haven’t had time to really understand what AI can — and crucially, can’t — do,” Derler said, noting that hype-driven decisions have led to disjointed workforce transitions and productivity shortfalls.

While AI systems promised efficiency and savings, companies have since encountered unforeseen challenges: long model-training cycles, inconsistent task accuracy, and the human effort required to supervise and refine machine output. “Lay-offs are never free,” Derler noted. “There might be immediate salary cost savings, but training AI for complex tasks can be costly and frustrating — and you can’t replace the versatility of a human in some roles.”

The trend suggests that firms are entering a period of recalibration, with many shifting towards a “human-plus-AI” approach rather than outright substitution. Analysts say the rehire wave signals a maturing view of artificial intelligence — one that recognises its potential as an augmentation tool, not a replacement for human capability.

As companies reassess their workforce strategies for 2026, the focus appears to be moving from automation for efficiency’s sake to integration for sustainability — a shift that could define the next phase of the AI workplace era.