Muse launched in September as Meta’s personal AI agent, promising to tackle chores such as emails, travel bookings and shopping. Within its first 40 days, it logged 10.9 million downloads—10.2 million of them in the US—putting it in a small club of apps to reach that mark so quickly.¹

The acceleration was not entirely organic. Sensor Tower said Meta had devoted 39% of its October advertising spend to Muse, while the company used its vast family of apps, with 3.6 billion daily active users, to put the new assistant in front of potential customers.¹ Still, the early engagement numbers offer Meta more than a marketing victory: Muse reached 2.5 million daily active users on October 16, and its US retention outpaced ChatGPT, Claude and Grok in the first two weeks.¹

By Wednesday’s third-quarter earnings call, Mark Zuckerberg was framing that reach as the beginning of a business. Meta said more than 10 million people in the US and Canada had set up a Muse agent, with about a quarter using it on a given day.² Shopping has emerged as the key test case: more than a quarter of engaged shoppers were making purchases averaging above $200. Zuckerberg said Plaid had not seen comparable activity through its connector “on any other service before.”²

Meta’s plan is to keep Muse free for most people, take a cut of commerce and eventually charge power users. It also wants Muse on smart glasses next month and imagines a “Muse network” in which agents interact with friends, family and businesses.²

That vision demands deeper access to emails, calendars and financial connections—and brings Meta’s old weakness back to the foreground. Forrester analyst Kate Winick said, “The more information Muse has, the more effective it can be, and overall trust in Meta remains low.”³ The warning arrived as a New Mexico judge ordered Meta to pay $25 billion over allegations including misleading users about data handling. Meta has won the download race; earning permission to become a constant personal assistant is the next, tougher contest.³