Amazon has cut off Meta’s new Muse AI agent from shopping on Amazon.com, deepening an industry-wide standoff over who controls the customer relationship as AI agents purchase on consumers’ behalf.
The e-commerce giant began blocking Muse on Sunday night. Before the radical action, Meta reportedly declined Amazon’s request to voluntarily exclude the retailer from the agent’s shopping platforms list, according to Bloomberg. So now, users attempting to buy products through Muse on Amazon (which agent description says is possible) face a pop-up stating: “Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use, to which our customers have agreed.”
Amazon says Meta never disclosed that Muse would access its store. The e-commerce giant is concerned the agent does not identify itself while browsing, and that it appears to capture and store customer credentials. These allegations only aggravate trust issues voiced by fintech community and users from the very beginning of the Muse launch.
“We think it’s fairly straightforward that third-party applications that offer to make purchases on behalf of customers from other businesses should operate openly and respect service provider decisions about whether or not to participate,” Amazon spokesperson Lara Hendrickson said in a statement.
The company added that agents built to buy on a customer’s behalf typically operate with the merchant’s consent. This is a common procedure for instance in food delivery apps and online travel agencies. “Agentic third-party applications such as Muse have the same obligations, and we’ve requested that Meta remove Amazon from the experience,” the spokesperson said. However, Muse apparently doesn’t offer an opt-in function.
Amazon also says Muse can reach account pages and order history when prompted. That practically means that an unidentified third party could be moving through customer accounts and handling sensitive data without Amazon’s knowledge. Meta did not immediately respond to requests for comment. Yet, the company has previously said Muse “has no visibility into people’s passwords or payment methods,” noting that credentials shared by users “go into secure storage, so Muse can use them without seeing them.”
The clash is especially notable not only in AI commerce context, but also given the companies’ existing business ties. Amazon products have been purchasable inside Facebook and Instagram since 2023, and Meta signed a multibillion-dollar deal in April to run agentic AI workloads on Amazon’s Graviton chips.
Nevertheless, this is not Amazon’s first move against outside shopping agents. The company sued Perplexity over its Comet browser, winning a preliminary injunction in March that was later overturned by the Ninth Circuit in August, which ruled that the user, not the AI company, was technically accessing Amazon’s systems. Amazon’s petition for rehearing was denied September 10, leaving contract-based claims, like the Conditions of Use cited against Muse, as its main legal lever going forward. Amazon has also moved to restrict shopping agents from Google and OpenAI.
Meanwhile, Amazon operates its own agentic tools: Alexa for Shopping, launched in May, and Buy for Me, which finds items on external sites but, as Amazon notes, identifies itself and lets brands opt out. However, when the products just launched, small retailers had also voiced plenty of concerns as they discovered their products were listed and sold through Amazon AI solutions without their consent.
Muse launched September 8 as a personal AI agent capable of multi-step tasks, e.g. shopping, booking appointments, managing email and calendar. It functions across iOS, Android, WhatsApp, and muse.ai. The app sparked immediate public interest, briefly overtaking ChatGPT as the top free app in Apple’s U.S. App Store.
Meta says the agent runs on a secure virtual machine with its own browser and that a separate “Sentinel” system must approve any outbound actions. At the same time, Meta separately advertises the upcoming function, Muse Confidential VM, “intended to cryptographically and verifiably prevent Meta from accessing data in your VM.” The company adds that current Muse architecture “does not prevent Meta from accessing data when necessary to support, secure or operate the service.” Those statements are quite vague to understand exactly which data can be accessed and how the tech giant would use it.
Industry data suggests the immediate commercial stakes remain modest. At present, only 16% of shoppers are comfortable letting an AI assistant complete purchases on their behalf, per Coveo’s 2026 Commerce Relevance Report. But McKinsey estimates AI-mediated search and purchase automation could account for up to $1 trillion in U.S. retail revenue by 2030. This projection could force retailers to decide whether locking out third-party agents protects their business and whether they should consider building their own in this case, such as Walmart’s in-development “Sparky” agent.
