Meta CEO Tests Personal AI Assistant to Streamline Executive Work
Key Points
- Mark Zuckerberg is leading the development of an AI agent to serve as his personal executive assistant.
- The system already functions as an on‑demand information tool that speeds data retrieval.
- Meta employees use internal AI tools like MyClaw and Second Brain to access files, chat logs, and institutional knowledge.
- Engineer output has risen 30 percent, with power users seeing an 80 percent increase year over year.
- Meta plans capital expenditure of $115‑$135 billion for 2026, nearly double the previous year’s spend.
- The company acquired Manus for $2 billion and created Meta Compute to drive AI development.
- The AI assistant is designed to automate information retrieval, not to make independent executive decisions.
- Success of the experiment could shape AI adoption strategies across other large tech organizations.
Meta chief executive Mark Zuckerberg is developing an artificial‑intelligence agent to act as a personal assistant for executive duties. The system, still in development, already serves as an on‑demand information tool that speeds data retrieval compared with traditional hierarchical channels. Internal AI applications such as MyClaw and Second Brain are already in use, giving employees faster access to files, chat logs, and institutional knowledge. Meta reports significant productivity gains, with engineer output up 30 percent and power‑user output up 80 percent year over year. The company is investing heavily in AI infrastructure, including a $2 billion acquisition of Manus and a capital‑expenditure plan that nearly doubles the previous year’s spend.
AI Agent Development
Meta’s chief executive, Mark Zuckerberg, is overseeing the creation of an artificial‑intelligence agent designed to serve as a personal assistant for executive duties. According to reporting, the system remains in development but already functions as an on‑demand information tool that lets the CEO retrieve data faster than traditional hierarchical channels would allow. Zuckerberg has emphasized that this year will be when “AI starts to dramatically change the way” Meta works, and the personal‑assistant project is a concrete step toward that vision. The move underscores a broader admission that valuable information often gets lost or delayed between individual teams and the C‑suite.
Internal AI Tools
Meta employees already have access to internal AI applications that foreshadow the broader re‑organisation underway. One tool, called MyClaw, gives workers access to internal files and chat logs, enabling communication with colleagues or AI‑agent counterparts without navigating bureaucratic handoff points. A second system, known as Second Brain, builds on Anthropic’s Claude infrastructure and functions as a personal chief of staff, organising tasks, surfacing insights, and streamlining the retrieval of institutional knowledge. These tools are intended to reduce the friction of information flow and coordinate across departments more efficiently.
Productivity Gains
Meta’s chief financial officer, Susan Li, reported that output per engineer has risen 30 percent since the start of 2025, a boost primarily driven by AI coding agents. “Power users,” staff who have fully embraced the new AI systems, have seen output increases of 80 percent year on year. These figures suggest that the AI tools are delivering measurable results rather than serving only symbolic purposes, and they illustrate how AI integration can accelerate the amount of work individual employees can complete.
Financial Commitment
Meta has forecast capital expenditure of $115 billion to $135 billion for 2026, nearly double the $72 billion spent in 2025. The company’s investment reflects a belief that AI infrastructure and tools will generate returns sufficient to justify the unprecedented spending. To support the effort, Meta acquired Manus, a general‑purpose AI‑agent developer, for $2 billion in December 2025, and created a new top‑level organisation called Meta Compute, led by Santosh Janardhan and Daniel Gross. These moves signal both priority and resources being devoted to the AI‑driven transformation.
Challenges and Outlook
Building an AI agent capable of assisting a chief executive differs from deploying coding assistants or general information tools. An executive‑level system must navigate competing priorities, weigh strategic decisions with incomplete information, and develop contextual understanding of organisational politics and human relationships. Zuckerberg’s approach is pragmatic: rather than granting the AI independent decision‑making authority, the system is intended to accelerate information access and processing, freeing the CEO to focus on decisions that machines cannot yet make. The experiment’s success or failure will likely influence how other technology leaders approach AI‑driven organisational redesign. If the AI assistant demonstrably boosts executive efficiency, it could encourage broader adoption of similar strategies across large technology firms.