What Mark Zuckerberg’s AI sidekick could teach CEOs about leading by example

Mark Zuckerberg is nothing if not a true believer. Again and again, the Meta CEO and Facebook founder has thrown himself headfirst into his company’s top initiatives. A few years ago, he made himself the face of the company’s since-sidelined metaverse push, and remained steadfast even as the internet mocked how his virtual reality avatar fenced, hydrofoiled, and, at times, looked awkwardly flat. He even ran internal and media meetings inside Meta’s own VR offices, which he argued was a better way to connect than regular video conference calls. He also regularly wears Meta’s bulky AI smart glasses in public, aesthetics be damned. 

The chief executive is now walking the walk on another Meta imperative: AI adoption. According to the Wall Street Journal, Zuckerberg is building an AI agent to help him as CEO. Details are scarce on the still-in-development tool, but the WSJ reports that it’s getting Zuckerberg information faster, expediting processes that normally require him to query multiple people. Meta did not immediately return a request for comment on the tool. 

As Meta spends tens of billions of dollars developing “superintelligent” AI models and building data centers to power them, it’s become borderline obsessed with staff-wide AI adoption. The company has encouraged employees to employ the technology in multiple ways, and incorporated “AI-driven impact” into its performance reviews. It is also reportedly among the tech giants that have established leaderboards that rank employees based on their consumption of tokens—a measure of AI use. But of all the methods of inducing AI adoption, Zuckerberg’s leading by example might be the most effective. 

Data shows an emerging credibility gap in which leaders are mandating and hyping AI but are often only casual users of the technology themselves—sometimes using it less than their rank-and-file employees. Nearly 70% of CEOs, CFOs, and senior executives use AI at work less than an hour a week, including 28% who don’t use it at all, according to a survey of more than 6,000 senior leaders in the U.S., U.K., Germany, and Australia co‑authored by Stanford economist Nicholas Bloom. The disconnect may be blinding leaders to the first-hand experience of using AI, which is causing workload creep and cognitive overload, at least in current use cases.

Separate research from Gallup finds that manager support of AI—including modeling its application—is a strong driver of whether employees use and value AI tools. In organizations investing in AI, employees who strongly agree their manager actively supports their team’s use of AI are more than twice as likely to use AI a few times a week or more, 6.5 times as likely to strongly agree the tools are useful, and 8.8 times as likely to say AI helps them do what they do best every day, Gallup says.

By all accounts, Meta’s AI organization-wide AI push seems to be working. It’s breeding an experimental culture reminiscent of Facebook’s heady early years, the WSJ reports, with employees participating in AI hackathons and deploying personal AI agents that do work on their behalf.

Does every CEO need a Zuckerberg-style AI sidekick? That remains to be seen. What is clear is that leaders who expect AI to be woven into daily workflows can’t stay light users of the tools; if they want credibility—and real adoption—they’ll have to log in and experiment. They’ll need to feel the pain and reap the gains, along with everyone else.

This story was originally featured on Fortune.com