Meta's AI-Native Workforce Plan Hit a Wall After Testing Showed Chaos

Meta seriously pursued a strategy in early 2026 to reimagine itself as an AI-native company, with internal planning suggesting certain departments could shrink by up to 60 percent if autonomous systems took over much of the routine work. The initiative, known internally as Project OT (organization transformation), ultimately fell short of its initial scope, according to reporting from Reuters.

The Restructuring Plan#

CEO Mark Zuckerberg green-lighted Project OT in January, tasking executives with modeling how the company might operate with far fewer people and significantly more automation. The proposal centered on deploying AI agents to handle substantial portions of day-to-day tasks performed by thousands of employees, with minimal human oversight teams managing the systems. Meta did carry out an initial round of workforce reductions in May aligned with this framework, though a planned second phase never materialized.

Why the Plan Stalled#

Meta's own framing of events suggests the scenario planning exercise was exploratory in scope. The company noted that not all proposals from the exercise were intended to move forward and stressed that many possibilities remained theoretical. This language masks a more telling reality: testing the system revealed significant problems. The AI agents triggered what Reuters sources characterized as "large-scale, disruptive actions" during trial runs, raising serious questions about whether such automation could function reliably at scale.

The Broader Challenge#

Meta's experience underscores a persistent tension in AI deployment across industries. Identifying where automation genuinely improves operations versus where it creates risk and instability remains genuinely difficult. The gap between theoretical efficiency gains and real-world execution has emerged as a critical constraint on how aggressively companies can pursue AI-driven workforce transformation.