Zuckerberg is Building CEO Agent to Let AI Help Him Run Meta

Meta Platforms chief executive Mark Zuckerberg is developing a personal artificial intelligence agent to assist with his day-to-day responsibilities as CEO, according to a report in The Wall Street Journal.

Credit: Mark Zuckerberg / Facebook

The tool, still in early development, allows Zuckerberg to obtain information more quickly by pulling answers that would normally require consulting multiple layers of employees, a person familiar with the project told the Journal.

Zuckerberg has indicated that the effort is the starting point for a wider rollout in which every person inside and outside the company would eventually have their own personal AI agent.

The project fits into Meta's ongoing push to embed AI tools throughout its operations and reduce organizational layers.

As Meta has acquired startups focused on AI agents, including Moltbook earlier this month and Manus, a Singapore-based firm whose technology is already being tested internally, employees now use tools such as Second Brain, an AI system built on top of Anthropic's Claude model that indexes documents and acts like a chief of staff, and My Claw, which can access chat logs, work files and communicate with other agents on users' behalf.

Meta established a new applied AI engineering organization with an ultraflat structure, where as many as 50 individual contributors report to a single manager.

AI usage has become a factor in performance reviews, and internal hackathons and tutorials encourage staff to build their own tools.

In a January earnings call, Zuckerberg described the company's approach:

“We’re investing in AI-native tooling so individuals at Meta can get more done. We’re elevating individual contributors and flattening teams. If we do this, then I think that we’re going to get a lot more done and I think it’ll be a lot more fun.”

The initiative builds on Zuckerberg's public vision for personal superintelligence outlined in a July 2025 letter on Meta's website.

In that document he wrote that Meta aims to deliver AI that empowers individuals to achieve personal goals rather than focusing solely on centralized automation.

No official statement from Meta or Zuckerberg addresses the specific CEO agent project. The company's most recent AI-related announcements have centered on infrastructure expansions and new consumer features in its apps.

What the CEO Agent Actually Does, and What It Doesn't

Based on the WSJ's reporting, the agent is not acting autonomously on Zuckerberg's behalf.

Its current function is intelligence retrieval: surfaces decisions made by different teams, flags relevant signals from across Meta's product portfolio, and compresses information that would otherwise require a chain of human intermediaries.

In effect, it operates as a chief-of-staff and analyst, without being human.

The distinction matters because of a separate and recent cautionary episode at Meta.

In mid-March 2026, Meta's in-house AI agent accidentally shared confidential company and user information with staff who were not supposed to see it.

The leak lasted two hours after an employee brought in the AI to help answer a technical question on an internal forum without checking permissions.

Meta classified the incident as a "Sev 1" breach, just one step below its highest alert level.

The incident's timing (just two weeks before the WSJ's CEO agent report) highlights Meta's bold AI ambitions while revealing the challenges of implementing effective guardrails in real-world settings.

The Broader Bet on Personal Superintelligence

The CEO agent project sits within a much larger financial and strategic commitment.

Meta's fourth-quarter earnings report revealed that its AI-related capital expenditures for 2026 will be between $115 billion and $135 billion, nearly twice the $72 billion spent the prior year, with Zuckerberg telling analysts:

"As we plan for the future, we will continue to invest very significantly in infrastructure to train leading models and deliver personal superintelligence to billions of people and businesses around the world."

Meta's acquisition of Manus, a Chinese AI-agent startup purchased for more than $2 billion at the end of 2025, is part of a broader realignment of the company in the AI space.

Manus, which Meta has said it will continue to operate as a standalone service while integrating into its own products, reached an annualized revenue of $100 million just eight months after its March 2025 launch, with consistent monthly growth of over 20 percent.

Meta also acquired Moltbook, an AI agent-focused social network, on March 10, 2026, with its co-founders joining Meta Superintelligence Labs as part of the deal.

Zuckerberg told investors:

"We're starting to see the promise of AI that understands our personal context, including our history, our interests, our content and our relationships," and added, "A lot of what makes agents valuable is the unique context that they can see, and we believe that Meta will be able to provide a uniquely personal experience."

The development occurs as Meta maintains a workforce of approximately 78,865 employees, down from a pandemic-era peak but up from post-layoff lows in 2023.

Internal experiments with AI agents continue to accelerate as the company positions itself for what executives describe as an AI-native future.