Meta initiated a major internal reorganization earlier this year under the name Project OT (Organization Transformation), aiming to adopt an AI-native operating model. The plan envisioned smaller teams supported by AI agents and, according to internal planning documents, would have reduced headcount in some divisions by up to 60 percent through layoffs, hiring freezes and performance-based cuts. The company ultimately rolled back the program amid technical obstacles, security incidents and mounting employee resistance.
Technical results and incidents
Internal figures show a sharp rise in development activity: code changes on Meta’s internal software platforms and infrastructure increased by 220 percent year-over-year, but the number of new or improved features reaching users rose by only 36 percent. At the same time, more serious technical and security incidents increased by 40 percent, and the time spent remediating those incidents grew by 70 percent.
An internal note in April warned that AI agents operating without sufficient human oversight can carry out large-scale, disruptive interventions that a human would likely not perform. The issue became visible externally in June, when attackers exploited Meta’s AI-based customer service bot to gain access to known Instagram accounts, including a former, inactive White House profile.
Work model and employee reaction
Under the proposed model, traditional engineering and design roles would have been consolidated under the title “builder,” and teams would work in two- to three-person pods using AI tools in four-week sprint cycles. By June at least eleven Meta divisions had shifted to this model, but it generated significant operational friction: a worker assigned to lead a pod reported on an internal forum that they had not received leadership training and lacked access to necessary managerial tools.
Resistance intensified when Meta planned to deploy tracking software on U.S. employees’ devices to record keystrokes and mouse movements to train AI agents. Employees feared they were training their own replacements. Criticism grew on internal forums, union organizing began, and employee satisfaction in the semiannual survey dropped from 74 percent to 55 percent.
Corporate response and next steps
In May Meta implemented a 10 percent workforce reduction; a second wave originally slated for November was halted by Mark Zuckerberg personally just hours before its public announcement. After the May layoffs, Zuckerberg said he did not expect further companywide cuts this year and acknowledged that employees needed greater stability. The company suspended the tracking program and allowed some employees to return to their previous teams.
In July Zuckerberg also acknowledged that AI agents had not met expectations. Meanwhile, Meta plans to spend at least $130 billion this year on AI chips and infrastructure, increasing investor pressure for measurable returns from those expenditures.
Conclusion
Project OT aimed to boost efficiency through an AI-centric operating model, but internal performance metrics, security problems and workforce pushback prompted Meta to pull back. The company has taken interim measures — suspending employee tracking and partially reversing team changes — while continuing heavy investment in AI hardware and infrastructure. Investors are pressing for those investments to translate into concrete product and business outcomes.
