Meta reportedly developed an ambitious plan to use artificial intelligence agents to automate large amounts of employee work and potentially reduce the size of some teams by as much as 60%. However, the plan was eventually scaled back after the company’s AI systems failed to deliver the productivity gains executives had anticipated.
According to a Reuters investigation based on internal documents, posts and recordings, Meta spent the first half of 2026 working on an internal initiative known as Project OT, short for Organization Transformation.
The project was designed around the idea that AI agents could take over routine tasks currently performed by thousands of employees. Smaller groups of workers would then supervise the systems and focus on higher-value responsibilities.
The strategy reflected a broader push by major technology companies to integrate AI into everyday operations and reduce the amount of repetitive work handled by human employees.
Project OT reportedly included two major restructuring waves. The first was planned for May, while another round was scheduled for November.
The proposed changes would have combined layoffs with the closure of vacant positions and the removal of employees whom Meta considered poor performers.
An HR executive reportedly estimated that the overall workforce reduction could reach or exceed the roughly 25% reduction Meta carried out three years earlier.
However, the second phase of the restructuring was reportedly called off shortly before the first round began.
On the night of May 19, Meta CEO Mark Zuckerberg cancelled planning for the November phase, according to the Reuters investigation. The reason for the decision could not be independently determined.
Meta nevertheless proceeded with significant workforce changes the following day. Around 8,000 jobs were eliminated, representing roughly 10% of the company’s workforce.
At the same time, approximately 7,000 employees were moved into AI-related positions, indicating that the company continued to prioritize artificial intelligence despite scaling back parts of its workforce transformation strategy.
The developments suggest that Meta’s approach to AI may have shifted from simply replacing employees toward combining AI systems with workers who can manage, supervise and improve those systems.
AI agents are designed to perform multi-step tasks with limited human intervention, making them potentially more powerful than traditional software automation. Technology companies have increasingly explored their use in areas such as coding, research, customer support and business operations.
However, deploying AI agents at scale can be more complicated than demonstrating their capabilities in controlled environments. Agents may struggle with reliability, complex decision-making and tasks that require context or judgment.
The reported difficulties at Meta highlight the gap that can exist between expectations surrounding AI and its practical performance inside large organizations.
For employees, the situation also illustrates the uncertainty created by rapid AI adoption. Companies may consider reducing teams or restructuring roles based on expected productivity improvements before the technology is fully capable of delivering those gains.
Meta’s decision to move thousands of employees into AI-related positions suggests that the company continues to view artificial intelligence as central to its future strategy.
The reported scaling back of Project OT does not necessarily mean Meta has abandoned AI-driven automation. Instead, it could indicate that the company is adjusting its plans based on the performance and limitations of its current AI systems.
The experience may also provide a broader lesson for companies considering major AI-led workforce transformations: technological capabilities need to match business expectations before large-scale organizational changes can deliver their intended results.
