Meta’s ambitious plan to replace workers with AI agents fell apart from the inside

Meta wanted to build the future of work. Instead, it got a case study in how not to roll out AI across an 80,000-person company.

A Reuters investigation revealed that Meta’s internal initiative, code-named Project OT (Organization Transformation), was designed to create what the company called an “AI-native” workforce. The goal was replacing significant portions of its human employees with AI agents, supported by skeleton crews of 3-5 people providing oversight. The plan discussed in January 2026 called for cutting team sizes by up to 60%.

The plan versus reality

On May 20, 2026, Meta laid off approximately 8,000 employees, roughly 10% of its workforce. Another 7,000 were reassigned into AI-focused roles. For context, this came on top of the more than 21,000 positions Meta eliminated during the 2022-2023 period Zuckerberg branded the “Year of Efficiency.”

The company’s projected capital expenditure on AI initiatives in 2026 could reach $145 billion. But CEO Mark Zuckerberg acknowledged in mid-2026 that the AI agents had not accelerated productivity as expected, and admitted to “mistakes” in the restructuring process.

Employee backlash torpedoed the rollout

To train the AI systems meant to replace workers, Meta deployed surveillance technology to monitor how employees performed their tasks. The response was predictable: employees pushed back hard.

Internal sentiment scores tell the story cleanly. Employee satisfaction dropped from 74% to 55%, a 19-percentage-point decline. The backlash forced a dramatic reversal. Zuckerberg had planned a second wave of layoffs scheduled for November 2026. He canceled it just hours before execution.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.