Meta CEO Mark Zuckerberg reportedly launched a secret internal initiative in 2026 to transform the company into an “AI-native” organisation, with artificial intelligence placed at the centre of how teams build products, write software and make decisions. The project, known as Project OT, explored a dramatic reduction in the size of some teams, according to internal documents and interviews cited in reports.
The ambition was bigger than simply giving employees AI tools. Meta explored a model in which smaller groups of highly skilled workers would supervise increasingly capable AI systems and agents. Scenario planning reportedly considered cutting some teams by as much as 60%, raising the possibility of a much broader transformation of the company’s workforce.
But the experiment also exposed a central problem with the AI-native workplace: producing more digital output is not necessarily the same as producing better products. Reports on Meta’s engineering experiments found that some AI-driven teams dramatically increased code changes, while the increase in shipped features was considerably smaller and incidents also rose.
That distinction is critical. AI can accelerate coding, documentation and routine analysis, but organisations still need people who understand priorities, take responsibility for outcomes and coordinate complex work. Meta’s recent move to reconsider parts of its management structure suggests that removing organisational layers may have created problems of its own.
Meta has nevertheless continued investing heavily in AI. Its engineers are developing systems designed to capture institutional knowledge and make specialist expertise accessible through AI agents, suggesting that the company has not abandoned the broader vision behind its transformation.
The lesson from Project OT may therefore be less about AI replacing employees and more about AI changing what organisations consider productive work. Zuckerberg’s experiment demonstrates that becoming “AI-native” is not simply a technology upgrade. It requires finding the right balance between automation, human judgment and organisational accountability.
For Meta, that balance could ultimately determine whether its enormous AI investment produces a genuine productivity revolution—or simply a more complicated way of working.
