Meta is adding an AI agent to its social media universe. On Tuesday, it launched Muse, a personal AI agent that will sit inside WhatsApp, Instagram and Facebook, and can act autonomously to do everything from messaging friends to planning a night out, to outfit shopping for a holiday.
Muse is a personal AI assistant that users can talk to “just like messaging another person”, like an AI chatbot, that is also capable of acting on the user’s behalf and can even “suggest ideas”, according to the company’s blog post about the news. Muse is rolling out to adults in the US for free first, and Meta plans to add the agent to its AI smart glasses soon, but didn’t provide a timeline or specify which agentic features this would involve.
For shopping, this means Muse can move beyond finding and comparing products to actually navigating checkout and paying on a user’s behalf, presenting the final transaction total in the Muse interface chat for the consumer to approve before completing the purchase. The new Muse AI agent can connect to consumers’ Facebook and Instagram accounts and access their DMs when they grant it permission, the Financial Times reports. WhatsApp is more restricted: because personal messages are end-to-end encrypted, Muse can only draw on information a user explicitly shares with the agent.
Meta isn’t the first tech giant to bring AI into checkout — OpenAI launched (and subsequently pivoted away from) its Instant Checkout last year, and Google launched its “Universal Cart” in May. Both sought to shorten the journey from AI product discovery to purchase, but Muse goes further by positioning the agent to navigate checkout and transact on the consumer’s behalf.
But Meta is betting that years of social signals about what its users follow, watch, save and share could give Muse something other shopping agents lack: an understanding of their taste.
Max Sinclair, CEO of Azoma AI, which works with brands and retailers to improve their agentic commerce discovery, says that the top AI shopping agents right now are Gemini, ChatGPT, Alexa for Shopping and Walmart Sparky.
“But I think Muse could shoot straight to being the fifth — largely because Meta has 3.6 billion people using its apps daily, and this is going to be integrated across all of them,” Sinclair adds. “It’s probably going to be unavoidable for people, because it’s just in front of them all the time.”
Meta enters the AI shopping race #
Meta has been experimenting with AI-powered shopping since earlier this year, allowing users to search and compare products and draw on content from brands and creators. However, it’s shared little evidence of consumer adoption so far. This week, Mastercard forecast that one in 10 shoppers will use a personal AI agent to purchase products and services on their behalf by 2030, but most surveys suggest that consumers are currently skeptical of the tech at best.
Our own Vogue Business AI consumer perception survey, released in April, found that only 31% of consumers would outsource shopping to an AI agent, even if it knew their taste and purchase history. This is largely down to fundamental distrust of how AI companies handle their data: 72% say they would not share card details, 46% would not share browsing history, and 40% say they would not share location data.
For Meta, those concerns come with additional baggage. The company has faced years of scrutiny over how it collects and uses consumer data, while its smart glasses have more recently become a flashpoint in a broader debate around surveillance and consent. Those concerns become more consequential with Muse. While Meta AI, its existing assistant across WhatsApp, Instagram, Facebook and Messenger, could already help consumers find, compare and choose products, Muse asks consumers to go much further: trusting Meta not just to recommend what to buy, but to act — and ultimately spend their money — on their behalf. The Muse announcement suggests Meta is aware of the trust gap and what’s needed for consumers to adopt the new tech. To protect sensitive data, Meta is effectively giving each Muse user their own private computer in the cloud, called a Muse Secure VM, where the agent works in isolation and cannot see users’ passwords or payment details, and must ask for approval before making a purchase.
Payments are handled through Stripe’s Link, which can generate a single-use card for the transaction rather than exposing the user’s actual card details. This Stripe integration also extends some familiar buyer protections to agentic shopping, including returns and coverage for lost or damaged items, which could also help build trust in letting an AI transact on consumers’ behalf.
The bigger gray area is how Muse will use the data consumers give it access to across Meta’s social platforms — including what signals it can draw on to infer their tastes and inform what it recommends or buys. “The more personal the agent becomes, the greater the expectation around privacy, transparency and control,” says Laurence Booth, group CEO at Trust Payments. “Therefore, any company that succeeds in agentic commerce will make consumers comfortable with how that knowledge is being used, particularly at the point where a recommendation becomes a financial transaction.”
What does it mean for fashion? #
Some of Muse’s biggest implications are for commerce, and affect brands, creators, and the consumer discovery process.
Shopify is already integrated with Muse, meaning eligible US merchants are automatically enrolled, rather than having to opt in, with their products shared with Meta by default and direct checkout switched on, unless they choose to disable it, according to the company. This gives Muse access to a significant pool of Shopify’s fashion and luxury brands, provided they don’t opt out.
If users connect their Instagram to Muse and give it permission to draw on information within the app when completing tasks, Meta gives the example of Muse retrieving a recipe reel a user previously saved on Instagram and turning it into a grocery list. In other words, consumers don’t necessarily have to explicitly feed social content into the agent each time — once access is granted, Muse can retrieve relevant context itself. This gives Muse the potential to surface much more relevant recommendations than AI chatbots like ChatGPT and Google Gemini, whose understanding of user preferences begins and ends with their AI chats. Saved posts, followed creators, and other social interactions can contain far richer signals of personal taste than a conventional shopping history. But the unanswered question is how extensively Muse will be allowed to interrogate those signals, and whether consumers will be comfortable giving an AI agent permission to turn years of social behavior into actionable shopping intelligence.
“They have data on our preferences down to when we stop scrolling for half a second and hover on a product, so this is going to be as hyper-personalized as it gets,” says Sinclair.
There’s also an important distinction for creators. If a user connects their Instagram account, Muse can retrieve relevant context from content they’ve previously saved, rather than requiring them to manually feed that content into the agent each time. Imagine a consumer saved a creator’s post featuring a handbag months ago, then later asks Muse to find them a new bag. In theory, that saved post could become part of the context informing what Muse recommends. On the plus side, that could make creator influence potentially more persistent and machine-readable — a post could shape a purchase long after the consumer first encountered it.
But it also raises a thorny commercial question about affiliate links and attribution. If Muse ultimately finds and buys the bag itself, rather than directing the consumer back through the creator’s affiliate link, the creator who originally influenced the purchase could lose both attribution and commission. Meta has not yet explained how creator attribution will work in such scenarios.
Meanwhile, although Meta has only said that Muse is coming to its AI smart glasses “soon”, without detailing the functionality, industry analysts see glasses as a potentially powerful shopping interface — a prediction by fashion futurists in our recent Future of AI report, but something that consumers express little current desire for. In practice, this potential integration could mean the Muse agent could identify a coat a consumer admires on a passer-by in the street and take them directly to purchase within the AI glasses, without ever reaching for a phone.
“The ability to process all the visual data we do when we walk around will only further increase Muse’s value to us, so Meta definitely has the upper hand on this because of the smart glasses lead,” Sinclair says. “Obviously smart glasses weren’t a breakout success yet, but it’s these big technology shifts like having Muse with you all day, every day, that will enable that shift.”
Ultimately, whether Meta can turn that technological advantage into mainstream adoption will depend on trust — an area where years of scrutiny over its social platforms and newer concerns about smart glasses surveillance leave it with considerable ground to make up.
Or as the futurist Magnus Lindkvist puts it in Mastercard’s new future of shopping report this week: “Every great shift in commerce has been, at heart, a shift in trust. From the marketplace to the mall, to the catalog, to the click of a button, each historic evolution of commerce has asked people to place trust somewhere new.”