Mark Zuckerberg spent millions hiring the team of engineers who would lead Meta to compete head-to-head with the AI giants. While we are still waiting for that to happenthe company is advancing in another area that they consider critical and that is the adoption of AI within its own company. This includes a personal agent for your CEO.
A personal agent. Zuckerberg imagines a future in which all Meta employees have their own AI agent to help them be more productive, and he’s starting with himself. According to an exclusive from Wall Street JournalMeta is developing a specific AI agent for its CEO. The goal is for Zuckerberg to be able to obtain information faster, avoiding having to go to different people in different departments to get it, like a kind of AI secretary.
Target: AI-native. Meta has 78,000 employees spread across countless departments. Having such a complex structure is a disadvantage compared to startups that have much smaller staff in which, in addition, the adoption of AI is present from day one. Zuckerberg believes that having his employees use AI in their work is critical to the company’s future success. During the last earnings call, he said that “We are investing in tools designed specifically for artificial intelligence, so that Meta employees can be more productive. We are enhancing the role of individual collaborators and simplifying the structure of teams. If we do that, I think we will achieve much more and it will be much more fun.”
It is evaluable. As reported Business InsiderMeta has included the use of AI within employee performance evaluations starting in 2026. Although it was not mandatory in 2025, employees were highly encouraged and even compensated for those who achieved exceptional results thanks to the use of AI. Additionally, they introduced an “AI Performance Assistant” that helped them write their evaluations.
AI Culture. Meta has an internal messaging platform where employees share how they are leveraging AI to be more productive. Some of the tools they are using are My Claw, a kind of OpenClaw who has access to your chat history and can communicate with other colleagues’ My Claw on your behalf. There’s even a group on the internal messaging platform for agents to communicate with each other, which is very reminiscent of Moltbook, the social agent network that Meta recently bought.
Another of the tools they are integrating into Meta is called ‘second brain’ and is a hybrid between chatbot and agent. It was created by an employee with Claude and is a kind of “AI staff manager” with whom you can consult documents for projects.
Enthusiasm and doubts. On top of all this, the company is also providing AI training and holding hackathons where employees are encouraged to create their own tools to boost their productivity. They say in the Wall Street Journal that while some employees find it “stimulating and fun,” others believe that this insistence and so many changes could be the prelude to new layoffs.
And the models? That’s what we would like to know. Meta spent a real million on sign the best AI talentsis building gigantic data centers and at the moment they have launched a total of zero models. The latest information points because the launch of their models has been delayed because their performance is not yet at the level they want. Yet, company results continue to improvemainly thanks to its advertising business.
Images | Goal, Unsplash
In Xataka | Meta has tried to kill the metaverse after a resounding failure. To his surprise, he has met resistance
GIPHY App Key not set. Please check settings