Zuckerberg reportedly testing a ‘CEO bot’
Reports say Mark Zuckerberg is experimenting with a personal AI agent — a ‘CEO bot’ — and exploring flatter hierarchies, a sign Big Tech leadership is pushing AI into executive workflows. Meta calls the reports speculative, but the story underscores growing bets on AI agents for decision support and org design. (the-decoder.com)
The Wall Street Journal’s scoop ran March 22, 2026 and cited “a person familiar with the project” as the source for the CEO‑agent story. (the-decoder.com) Meta closed on its acquisition of AI‑agent maker Manus at the end of 2025 in a deal reported at roughly $2–3 billion, giving the company a production‑grade agent stack to plug into internal workflows. (bloomberg.com) Meta’s headcount stood near 78,000 employees at year‑end 2025 while Reuters reported company planning scenarios that could affect as much as 20% of staff as it balances AI infrastructure costs. (the-decoder.com) Internal policy changes show Meta will make “AI‑driven impact” a formal part of performance reviews starting in 2026, a change described in internal memos reported by Business Insider and summarized by HR outlets. (hrgrapevine.com) Manus has already shipped a “My Computer” desktop agent for Windows and macOS that executes multi‑step tasks across files and apps, demonstrating the kind of execution layer Meta acquired for agentic workflows. (pcmag.com) Meta’s planned AI capital expenditures for 2026 were reported in the $115–$135 billion range, and Mark Zuckerberg has publicly framed 2026 as the year AI “dramatically” changes how the company works. (cnbc.com) Reporting also describes an internal “agents” messaging board and a stated vision that every employee could eventually have a personal agent, signaling an operational push toward agent‑mediated decision paths rather than purely human chains of command. (the-decoder.com)
