Meta Canceled Plan to Cut Teams 60% and Replace Workers with AI

Meta developed detailed plans earlier this year to reduce certain teams by as much as 60 percent as part of an initiative to become "AI native," according to internal documents and sources reviewed by Reuters. The company ultimately abandoned the effort after AI agents caused significant technical problems and failed to deliver promised productivity improvements.

The initiative, codenamed Project OT (organization transformation), was set in motion by CEO Mark Zuckerberg and directed executives to restructure teams around AI automation. The plan called for two rounds of layoffs, with the first wave occurring in May before Zuckerberg canceled the second round scheduled for November.

The AI native vision

Meta's internal planning documents described an "AI native" company as one where "AI-ready tools and agents interact, workflows are automated, [and] new builds are AI-first." The vision included using AI to perform much of the daily work currently handled by thousands of human employees, with small teams of people overseeing the automated systems.

One HR executive indicated the scenarios would have reduced Meta's overall headcount by approximately 25 percent or more. The company planned to redirect savings from eliminated positions toward compensation for high-performing employees, particularly those with AI engineering expertise.

Meta launched pilot programs that restructured engineering teams, research teams, and at least eight other groups into smaller units. An October internal post titled "AI-Native Playbook" outlined how the pilot would remove middle management layers and use "agent-assisted analysis" to prioritize daily tasks.

When AI agents backfired

The ambitious automation plans encountered serious operational problems. Internal posts documented that AI agents made "large-scale, disruptive actions that humans are unlikely to execute," leading to a 40 percent increase in major technical and security incidents compared to the previous year. Employee time spent resolving these problems jumped as much as 70 percent.

Productivity metrics also revealed disappointing results. While code changes to internal software platforms increased 220 percent year-over-year by early June, changes that actually delivered new or upgraded features to Meta users rose only 36 percent—suggesting substantial churn without proportional value creation.

By July, Zuckerberg acknowledged during a company meeting that "the trajectory of the agentic development over at least the last four months hasn't really accelerated in the way that we expected."

Why it matters

Meta's experience offers a cautionary data point for organizations racing to implement AI automation at scale. Even a technology leader with substantial AI resources encountered fundamental challenges replacing human judgment and workflows with agents. The 40 percent spike in major incidents and the gap between code volume and shipped features suggest current AI systems can generate activity without delivering business value—while creating new operational burdens. For executives evaluating AI workforce strategies, Meta's pivot illustrates the risks of moving too quickly before the technology demonstrates reliable performance in production environments.

Meta confirmed to Reuters that it explored scenarios involving significant team reductions and redeployments but stated it "didn't move forward with every scenario from the exercise." The company emphasized that performance and promotion decisions "were and are made by people, not AI."

Reuters reviewed scores of internal documents, posts, and recordings and spoke with more than 20 people familiar with Meta's operations for its report published today.