Meta’s newest artificial-intelligence offering, the Muse personal agent, has crossed the ten-million-user threshold in the United States and Canada after just a little over a month on the market. The milestone was confirmed during the company’s third-quarter earnings call, where CEO Mark Zuckerberg highlighted that roughly one-quarter of the app’s adopters engage with it on a daily basis. This level of stickiness is especially noteworthy given the crowded AI-assistant landscape, where many rivals struggle to retain users beyond the first week.
The daily active user count for Muse surged to 2.5 million on 16 October, marking a 15 % weekly increase. Retention metrics also outpace competitors: after seven days, 30 % of U.S. users remain active, and after fourteen days the figure stands at 23 %. By comparison, platforms such as ChatGPT, Claude and Grok typically hover around 19 % and 16 % for the same intervals. These numbers suggest that Meta’s cross-promotion strategy—leveraging its 3.6 billion-strong daily active base across Facebook, Instagram and WhatsApp—combined with a sizable advertising push, is effectively converting casual browsers into repeat users.
Commerce engine fuels dramatic spend
Beyond mere conversation, Muse has quickly become a powerful shopping conduit. More than a quarter of the app’s engaged audience initiates purchases through the AI agent, with the average transaction exceeding $200. The underlying payment infrastructure relies on Plaid’s connector, which links users’ bank accounts to the platform securely. According to Zuckerberg, Plaid has never observed comparable activity on any other service, underscoring the novelty of Muse’s agentic commerce model. This high-value basket size is unusual for a mobile AI assistant and points to a monetization pathway that could eclipse revenue per user from Meta’s legacy family of apps.
Monetization roadmap
Meta plans to keep the core Muse experience free for the majority of its audience while extracting a commission from every transaction the AI facilitates. Power users seeking access to more advanced language models or larger token allocations will encounter a subscription tier. CFO Susan Li also mentioned a forthcoming “Muse Confidential VM” that will seal off even Meta’s own visibility into user content, aiming to bolster trust and justify premium pricing. The company projects that, once scaled, Muse will generate more annual revenue per user than Facebook, Instagram, WhatsApp and Messenger combined, which currently sit at several hundred dollars each.
Scaling challenges and hardware integration
Running a sophisticated generative-AI model at consumer scale is compute-intensive. Zuckerberg noted that Meta is engineering a ten-fold improvement in cost efficiency over the next twelve months, targeting profitability once transaction volume reaches a critical mass. In parallel, the AI agent is slated to become the centerpiece of Meta’s upcoming smart-glasses initiative, with a first-generation device expected to ship next month. The “Muse network” concept envisions a collaborative layer where a user’s personal agent can interact with the agents of friends and family, turning commerce into a shared conversational experience.
Analysts stress that the ultimate success of Muse hinges on consumer trust. While the platform embeds features like the “Muse Secure VM” to protect data, lingering skepticism about Meta’s handling of personal information remains a hurdle. Nevertheless, the rapid adoption curve, robust spending patterns, and aggressive integration with hardware suggest that Muse could become the flagship product of Meta’s AI-first strategy for the foreseeable future.
