Story
September 21, 2026
Muse Hits No. 1, but Meta Still Has to Earn the Keys to Users’ Lives
Muse’s rapid rise suggests consumers are ready for AI that acts rather than merely chats, while Meta argues users remain in control of what the agent can access. The company’s distribution advantage is powerful, but its privacy record leaves it with a harder job: persuading people to hand over the context an agent needs to be useful.
Meta launched Muse as a personal AI agent built to do more than answer prompts: research products, fill out forms, make bookings, shop, manage subscriptions and link with services including email, calendars, Spotify and Instagram. Meta says the agent remembers preferences and seeks approval before sensitive actions.1
Within roughly a week to 10 days, that pitch had pushed Muse to No. 1 among free iPhone apps in the United States, ahead of ChatGPT, Gemini, Claude and Instagram. The climb is an early win for Mark Zuckerberg’s belief that Meta can turn AI into an everyday consumer product through design, distribution and its existing network of users.1
Early anecdotes have helped fuel the buzz. One user said Muse found equivalent car insurance that was $3,500 cheaper annually, bought the policy and canceled the old one in about five minutes. Others said it identified an overlooked recurring charge or handled a ticket cancellation and refund by text.2
Meta is now widening the agent’s reach. Zuckerberg said Muse for Mac works across “apps, files, calendar, notes, and messages,” while insisting that users control its access.
3 He also announced connectors for developers, pitching a system in which Muse brings the browser and the user’s intent to outside services.4
That is also where the risk sharpens. A genuinely useful agent may need visibility into email, purchases, financial accounts, contacts and even calls made in a user’s name. Early users have already reported limits: one said Muse could not summarize an Instagram workout video because it saw only the first frame. Retailers, meanwhile, have incentives to curb bots that turn their sites into little more than checkout pipes.1
For Meta, the central tension is clear. Muse’s mass-market momentum shows the appeal of an AI assistant that gets things done. Yet every new permission asks users to extend trust to a company still carrying deep privacy and safety baggage — including a recent settlement over claims that Facebook and Instagram were designed to addict children.2