Mark Zuckerberg told Meta employees that the company's AI agent strategy is advancing slower than expected and the organizational restructuring tied to it has not delivered the projected gains. The admission, made at an internal town hall in early June and reported by Reuters, comes after Meta cut 8,000 jobs in May - roughly 10% of its workforce - and reassigned thousands of staff to AI teams in a sweeping reorganisation designed to accelerate AI development.
The CEO's comments reveal a gap between executive optimism at the start of 2024 and the actual pace of progress. "The trajectory of the agentic development over at least the last four months hasn't really accelerated in the way that we expected," Zuckerberg said, according to Reuters. He added that the company's bets on the new structure "haven't come to fruition yet" and acknowledged the rollout was not as "clean" as it could have been, with leaders misjudging the timing.
Why the restructuring missed the mark
Earlier this year, planning was driven by a conviction that Meta would not "move fast enough to adapt" unless it made dramatic changes. Executives were "super optimistic" about AI coding tools, including Anthropic's Claude Code. But the expected acceleration in agentic development did not materialise. Zuckerberg now says he expects meaningful returns from AI Agents & Automation investments within six months.
The gap between ambition and execution is a familiar problem in large-scale AI rollouts. For leaders in strategy roles, the episode underscores that organisational redesign does not automatically translate into faster product cycles. Meta's experience suggests that even well-funded AI pushes can stumble when structural changes outpace the technology's actual readiness.
Employee monitoring and the pushback
The town hall also addressed Meta's internal monitoring programme, which tracked US employees' mouse movements, clicks and navigation patterns to train AI agents for coding and other white-collar work. The software, rolled out in April without an opt-out option, sparked a petition from 1,600 workers and was later paused to review data exposure risks.
CTO Andrew Bosworth said a review found no employee data had been fed into AI training models. He said the mouse-tracking software will return on an "opt-in" basis. "For people who are comfortable, that's great, they can contribute to this kind of great human survey," Bosworth said. "To people who are not, it is not an issue." The shift to optional participation addresses immediate privacy concerns, though the episode raises lasting questions about internal data use for AI development.
Why this matters for Executives and Strategy
Meta's slowdown in agentic development is not just a tech company's internal hiccup. It illustrates a pattern that affects any large organization betting on AI to reshape operations: structural change without commensurate product maturity creates friction, not speed. For executives driving AI for Executives & Strategy, the takeaway is straightforward - align restructuring timelines with verifiable technical progress, not aspirational forecasts. The six-month window Zuckerberg now cites is a public deadline; missing it would force Meta to answer harder questions about whether its AI bets are sustainable.
Your membership also unlocks:
