In January, Mark Zuckerberg and his top executives gathered at his Hawaii compound for their annual leadership retreat. What they planned there never made it to the public until now.
Reuters published a special report on August 26, based on dozens of internal documents, recorded conversations, and more than twenty in‑depth interviews, that details what Zuckerberg was actually attempting and why the initiative faltered.
Project OT aims to replace human workers with AI agents
The initiative was codenamed Project OT—short for Organization Transformation. Its vision was an “AI‑native” future where artificial intelligence takes on a substantial share of tasks previously performed by human staff, with smaller, highly specialized teams supervising virtual workers, the report states.
Certain units faced potential cuts of up to sixty percent. The restructuring unfolded in two phases: the first kicked off in May, the second scheduled for November to target a broader suite of roles across the company, Reuters reported.
Why Project OT unraveled
The reversal materialized as internal data highlighted serious uncertainties about the technology’s true impact.
Employees were generating substantially more software through AI assistance. Modifications to Meta’s codebases and infrastructure surged 220 percent year over year, yet tangible product improvements visible to users increased only 36 percent, according to IBTimes.
The productivity disconnect was only part of the issue. Meta had also rolled out surveillance tools on U.S. staff devices, capturing keystrokes and mouse interactions to train AI agents that replicate human workflows, Reuters noted.
This approach fueled an open revolt among employees, who felt the AI transformation was largely intended to displace rather than support their careers.
Zuckerberg builds a personal AI assistant for executive decision‑making
Meanwhile, Meta scaled back its most aggressive layoff agenda, yet Zuckerberg was advancing his own AI-driven initiatives.
Internal communications describe a “CEO agent” — a personal AI capable of retrieving information instantly, bypassing traditional multi‑layer review processes, as highlighted by Euronews.
If a billionaire executive can condense decision pipelines that previously demanded extensive team involvement, the same model is being applied company‑wide. During a January earnings call, Zuckerberg hinted that projects once requiring large groups were now being completed by a single exceptional individual, Barchart reports.
Meta’s dedicated AI‑native units now report under structures allowing up to fifty individuals per manager. AI tool adoption has become a factor in performance assessments, underscoring a clear strategic trajectory despite the reduced tempos.
Implications for employees and investors
For staff, the experience with Project OT demonstrates that AI‑driven job replacement remains a real and persisting risk at Meta, even when the most severe edition of the plan was retired. The corporation continues to reconfigure operations under the assumption that AI will ultimately absorb many existing human functions.
Investors, however, should focus on the stark productivity challenge exposed by internal data. While Meta has poured resources into AI — code‑generation volumes have risen sharply — the products consumed by end‑users advanced barely 36 percent, creating a decisive divide between capital outlay and market impact that will dominate the upcoming earnings release.
Ongoing legal proceedings, specifically accusations of an addictive design targeting minors, have also weighed on shares, per TheStreet.
A workforce strategy that provoked internal friction and delivered mixed productivity outcomes adds another layer of difficulty to Meta, a firm simultaneously battling a major litigation case and attempting to reinvent itself around technologies it has not yet definitively proven.