Mark Zuckerberg is building an AI agent to help him run Meta, turning the company’s AI push into something far more personal.
First reported by The Wall Street Journal, the project shows the company experimenting with AI as a leadership tool. It also opens a window into Meta’s wider internal changes, where speed, access to information, and decision-making are becoming a bigger part of the company’s AI agenda.
The CEO project also fits a much bigger idea Zuckerberg has been talking about: a future where everyone has a personal AI agent, with him starting by building one for himself.
A personal AI built for the top job
According to WSJ, the AI agent is being designed to help Zuckerberg pull in the information he needs without waiting for it to travel through the usual chain. Instead of relying as heavily on briefings, intermediaries, or multiple layers of managers, the system can surface what matters directly and quickly.
That makes the idea notable not because it sounds futuristic, but because of what it is meant to replace. The promise here is less filtering, fewer handoffs, and a shorter path between the CEO and the details he wants in front of him.
For a company as large and sprawling as Meta, even a small change in how the chief executive gathers information can say a lot about how the company wants power and decision-making to flow.
More than a CEO experiment
The AI agent fits into a wider change already underway across the company. Zuckerberg has described it as an investment in “AI-native tooling” that is “elevating individual contributors and flattening teams.”
Meta is also changing how some teams are organized around that effort. It recently created a new applied AI engineering group with an ultra-flat structure, including setups in which as many as 50 individual contributors report to one manager.
The internal tools already in play
Employees are sharing AI use cases and homegrown AI tools on Meta’s internal message boards, while tools such as My Claw can tap into chat logs and work files and even reach out to colleagues or their own agents on a user’s behalf.
Another tool, Second Brain, is gaining traction as a way to index and query project documents. Built internally on top of Claude, it was introduced to staff as something “meant to be like an AI chief of staff,” a detail that neatly echoes the logic behind Zuckerberg’s own CEO agent.
The culture around these tools is also becoming more active and more social. There is even an internal group where employees’ personal agents talk to each other, showing that these systems are starting to move from isolated experiments into everyday collaboration.
When the experiment meets real consequences
The bigger stakes come into view when these systems start shaping real work.
In a recent incident, an AI agent reportedly responded to an employee’s request for help on an internal forum, and the employee followed that guidance, exposing sensitive user and company data internally for about two hours. Meta said no user data was mishandled, but the episode still showed how quickly an internal tool can create a serious problem once it starts influencing decisions.
That tension is also showing up more broadly within the company. While some employees have described the current moment as fun and empowering, others have said the rapid pace of change and intense focus on AI use are feeding anxiety about potential layoffs.
The same changes meant to speed decisions and streamline work are also forcing tougher questions about oversight and responsibility when something breaks.
Also read: A reported multibillion-dollar solar purchase from China suggests Elon Musk is still looking overseas for scale and speed.