Meta Brings Muse AI Agent to iPad With Work and Shopping Links
Meta released a dedicated iPad app for its Muse assistant on October 7, one month after Muse debuted on iOS and Android on September 8. TechCrunch
The release follows early adoption on phones. Muse passed 6.6 million installs since launch, according to Sensor Tower estimates cited in that report. In September, Muse ranked as the No. 1 free app on the U.S. Apple iOS App Store and Google Play stores. Bloomberg
Muse is a personal AI agent, meaning software that can take actions for a user, including placing calls. Reuters Users can link accounts to manage email, meetings and bills and complete tasks such as booking reservations, ordering groceries, setting goals and making purchases.
The iPad version adds many more connectors, which are approved links that let Muse work inside other apps, with a focus on work. Meta added Asana, Canva, Dropbox, Figma, QuickBooks, GitHub, Klaviyo and Zoom as Muse connectors aimed at small businesses. It also added Notion and Granola as connectors, plus connectors for Meta ad accounts.
Shopping is the other focus. Meta added Best Buy, Gap, Sephora, Walmart and Wayfair as retail partners to let users shop and complete transactions through Muse.
Two related events surround the release. On October 6, Meta said it is working with industry partners on an open standard to let personal AI agents identify themselves to businesses. On October 2, Apple said it will flag AI requests for Mac data after Meta's Muse drew complaints. Reuters
The broader context here is the change from assistant as chat window to agent as delegate. Connectors for files, design tools, code, accounting and meetings are more useful on a tablet than on a phone. The screen is larger. Multitasking is normal. Work files sit side by side. An agent that can read a brief in Notion, pull images from Dropbox or Canva, check a project in Asana, and match the spending in QuickBooks suits that setup in a way simple question-and-answer never did.
In my view, the list to watch is not the longest one. It is the most transactional one. Retail checkout and ad-account access need delegated logins, limited permissions, records of what was done and clear rules on who is responsible when the agent makes a mistake. Identity for agents, the subject of Meta's October 6 standards effort, is basic infrastructure for that. Businesses need to know if they are talking to a person, to a person working through an agent, or to an agent acting alone.
Worth flagging for enterprise and platform teams is the friction already visible. Agent access to mail, calendars, local files and now Mac data raises questions about permission prompts, system-level notices, and where rules are enforced. Apple's decision to flag AI requests for Mac data points to an emerging pattern. The operating system will describe agent behavior to the user, even when the agent comes from another company.
Looking ahead, there is reason for optimism about what this enables. If identity, permissions and disclosure are worked out, small teams get something they have not had at this price. A single operator with an iPad could sort inbox and calendar, coordinate contractors in Asana and GitHub, prepare designs in Canva and Figma, meet customers over Zoom, and finish orders and books without switching apps every few minutes. The technology fades. The work remains.
