Muse goes to London.
By Mara Quinn | November 9, 2026
Meta's AI assistant Muse reads emails, pays bills, books transport, and shops online. Within weeks it will do all of that in the United Kingdom. Senior Meta executives are travelling to the UK for the launch. The company has declined to give an official date, but reports from The Times and Retail Gazette place the rollout as imminent.
The timing coincides with a feature update that changes what Muse actually is. Meta has begun rolling out Muse Spaces, a mini-app creation tool that converts user commands into persistent, dedicated interfaces stored in the app's sidebar. Each Space sits there waiting, with its context intact, so users never have to re-explain the task. It moves Muse away from a general-purpose chatbot and toward a set of personal tools that live alongside daily life.
Spaces: Turn a Command Into a Tool
The mechanic is direct. A user tells Muse what Space to create, and the agent builds it into a dedicated interface, accessible from the sidebar at any time. Meta demonstrated the range: a family activity log, daily product recommendations drawn from Instagram activity, an EV charging overview, and a package tracking tool.
Mark Zuckerberg's own example was more specific. He created a Space that controls the vents on a meat smoker via Bluetooth. The case illustrates the product's intent well. Spaces are built for recurring tasks, not one-off queries. The agent needs to track and return, not just respond and forget.
Retention has been a problem worth solving. Sensor Tower analysis found usage drops significantly two weeks after download. Social Media Today reported more than 10 million downloads at an earlier stage of the rollout. Spaces gives users a reason to come back that a conversational interface alone cannot manufacture.
Memory and Persistence
Spaces launched alongside two additions drawn from Muse Wonder, Meta's research division. Active Memory lets Muse and Meta AI remember what matters to the user, recognising when information from an earlier session becomes relevant again. Active Focus lets agents continue pushing toward a goal when unexpected obstacles appear. Meta reported a 67 percent improvement in completing everyday tasks when complications arise.
Both updates address the same gap: AI agents that reset with each session cannot build the kind of utility that justifies daily use. Active Memory and Active Focus are Meta's attempt to close it.
11 Million Users, Retail Deals, and a Subscription Tier
Muse launched in the United States and Canada in September, then expanded to Mexico, Singapore, and Japan. The service now has more than 11 million users. Zuckerberg called it "one of our most successful launches ever" and said the company expects Muse to serve billions of people, putting what he described as "the power of personal superintelligence in everyone's hands."
The UK launch will include retail partnerships giving Muse direct purchase capability. No specific UK retailers have been named in reports. The service is free for most users, with a subscription option for heavier use. Meta earns a transaction-based fee when purchases are completed through the assistant.
Shopping as the Business Case
Around a quarter of North American users already use Muse for online shopping, making it the most popular function in the app. Research firm Gartner projects that one-fifth of all e-commerce transactions will pass through AI agents by 2030.
Amazon blocked Muse in September, citing unauthorised access and security concerns. The move showed where the competitive stakes sit. If Muse becomes the interface through which consumers buy, it takes that relationship from platforms that have held it for two decades.
Regulators Have Already Flagged the Risks
The UK rollout arrives against a specific regulatory backdrop. On October 30, the Digital Regulation Cooperation Forum, whose members include the Competition and Markets Authority, the Information Commissioner's Office, the Financial Conduct Authority, and Ofcom, published a joint statement warning that AI agents could "amplify and accelerate risks and introduce new ones." The statement cited fraud and scams, personal data shared in ways users do not expect, and misleading recommendations.
The CMA published separate research on agentic AI and consumers the same day. The ICO opened a call for evidence last month on data-protection risks from agentic AI systems.
Meta has not publicly addressed how it plans to engage UK regulators. Its executives flying in for the launch signals the company is not treating the UK as a simple expansion. They are arriving at a moment when the conversation has already started without them.
