Mark Zuckerberg is developing a personal AI to help him lead Meta

At Meta, AI is no longer just used to launch products or boost advertising. It is also beginning to reach the core of internal power, where decisions are made.

An AI connected to the boss’s desk

Mark Zuckerberg is developing a personal AI agent to help him manage Meta faster and with fewer intermediaries. This tool is still a work in progress, but it is already used to speed up information search by bypassing several hierarchical levels. The idea is simple, almost brutal: to fetch the right data without making the request travel throughout the company.

This detail says much more than it seems. For years, the promise of AI in business was mainly about customer support, coding, or office productivity. At Meta, it now rises all the way to the command post. AI is no longer just a work tool. It becomes a layer of management.

This choice aligns with an obsession Zuckerberg has displayed for several months: making Meta faster, more direct, and more competitive compared to much lighter AI startups. When a company has nearly 79,000 employees, each added level slows down the machine. The personal agent envisioned for the CEO thus looks like a technological shortcut, but also a very clear cultural signal.

Meta wants to flatten its own machine

At the end of January, Zuckerberg had already set the scene. During the release of the quarterly results, he presented 2026 as the year when AI would begin to deeply transform how Meta operates. At the time, this sentence might have seemed abstract. It is much more concrete today.

The Wall Street Journal also describes a broader use of agents within the group. Tools like MyClaw provide access to work files and discussion logs. Another system, Second Brain, is presented internally as a sort of AI chief of staff and relies on Anthropic’s Claude infrastructure. In other words, Meta is not experimenting with an isolated gadget for its founder. The company is testing a new way of circulating information.

This is where the subject becomes strategic. If employees consult agents to retrieve data, organize a project, or get context in a few seconds, the value of intermediate layers changes. It does not automatically disappear. But it must now justify itself differently than by simple information transmission. In many large groups, this is precisely where AI begins to disrupt the established order. This reading is an inference based on how Meta deploys its internal tools and the stated objective to streamline work.

The promised efficiency also opens an area of tension

This shift comes in a sensitive context. Meta could be preparing a new layoff plan that might exceed 20% of the workforce, even if the final scope and timeline are not yet decided. The group rejected this reading by calling it a speculative and theoretical scenario, without confirming such a plan.

Why is this suspicion gaining so much traction? Because Meta is spending massively on AI. Reuters mentions up to 135 billion dollars in investments in 2026, in a context where markets still accept the effort but closely monitor the real return of this race. If AI improves productivity, the question of the number of employees mechanically comes back to the table.

So the crux of the story is less technological than organizational. Zuckerberg is not only looking for a more powerful AI. He is looking for a more agile company, with less friction, less delay, and perhaps fewer human relays between idea and execution. It is an appealing promise for investors. For employees, it can also resemble an era of constant sorting, where AI becomes at once copilot, filter, and silent judge of each person’s usefulness. That is why Jamie Dimon warns about the real impact of AI on global employment.