AI layoffs backfire: 32% of companies bring back jobs cut through automation

Companies replacing employees with artificial intelligence are discovering the limits of automation, with a new survey highlighting a reversal in workforce decisions.

32% of US hiring managers surveyed by staffing firm Robert Half said their organisations had brought back the exact roles previously eliminated because of AI or automation gains. The findings point to an emerging trend in which employers are reconsidering job cuts after technology fails to deliver the expected operational benefits.

The phenomenon, described as the “AI boomerang”, is appearing in functions such as customer service and engineering, according to International Business Times, which reported on the findings. Fast Company also reviewed the Robert Half research, highlighting lost institutional knowledge, additional oversight requirements and weaker-than-expected productivity gains as reasons behind the reversals.

Why companies are reconsidering AI-led job cuts

The survey findings suggest that automating individual tasks does not always remove the need for the people responsible for managing wider processes. Employees often handle responsibilities beyond their formal job descriptions, including resolving unexpected problems, maintaining client relationships and applying knowledge built through years of experience.

Management expert Brooks Holtom of Georgetown University's McDonough School of Business told Fast Company that organisations can underestimate how employees actually work when deciding which roles AI can replace.

The research identified three prominent reasons for restoring positions among hiring managers who reported bringing back roles:

• 40% cited organisational knowledge and context beyond AI's capabilities.

• 38% said they needed more supervision or quality checks than anticipated.

• 35% reported productivity improvements falling short of expectations.

The findings indicate that cost savings from automation can be offset by the need for human oversight, operational knowledge and problem-solving.

Klarna and Ford illustrate the limits of automation

The experience of Swedish buy-now-pay-later company Klarna offers one example of a company reassessing its reliance on AI for customer service.

Chief Executive Sebastian Siemiatkowski acknowledged that the company had gone too far in relying on chatbots to reduce costs. Klarna had previously promoted its AI assistant's capacity as comparable to hundreds of customer service agents, but returned to recruiting human customer service staff in 2025.

The company continues to use automation for routine queries while keeping human employees available for more difficult customer problems. These include complaints involving unusual circumstances, emotional responses or situations in which a chatbot cannot reliably determine the appropriate action.

A similar distinction has emerged in manufacturing at Ford. The automaker recruited or promoted experienced engineers, including former employees and people from suppliers, after automated quality checks failed to meet expectations.

These engineers help identify manufacturing defects, investigate their causes and use the findings to strengthen quality checks. A Ford representative credited the combination of AI and engineers' technical expertise with improvements in quality.

Human oversight remains important in AI adoption

The examples highlight a distinction between automating repetitive tasks and replacing the judgement required to manage complex situations. AI systems can support employees, but organisations may still need experienced workers to check outputs, resolve exceptions and maintain operational standards.

The trend also comes as companies across sectors announce workforce reductions linked to AI and restructuring. Microsoft, Barclays and Kroger are among the companies cited in reports of AI-related job cuts.

For employers, the Robert Half findings offer a reason to assess automation decisions beyond immediate headcount savings. The effectiveness of AI deployment also depends on whether organisations account for supervision, quality control and knowledge that may not be captured in formal job descriptions.

As businesses expand AI adoption, the challenge will be to identify which tasks can be automated reliably and where human expertise remains necessary. The survey suggests some employers are already revisiting those decisions after experiencing the operational consequences.