Meta launched Muse on September 8 as a general-purpose AI assistant designed to handle multistep tasks across email, calendars, payments, dining and shopping. The app quickly reached the top of Apple’s US free-app chart, putting its promise of an agent that can act—not merely answer—before a mass audience.

That expansion ran into Amazon’s defenses. Users who sent Muse to the retailer began receiving a warning that continued access by an unauthorized AI agent violated Amazon’s terms. Amazon then asked Meta to remove its marketplace from Muse, saying it had not been notified in advance and had not authorized the service’s activity.

Amazon’s objection goes beyond a technical bot block. The company says Muse does not identify itself, may capture and store customer credentials, and scrapes account information—claims that raise obvious questions about sensitive data and who governs the transaction. Its position is that agents must be transparent, let retailers opt out and create a reciprocal commercial arrangement. “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,” an Amazon spokesperson said.¹

Meta, meanwhile, has said Muse cannot see users’ passwords or payment details, and that credentials are stored securely so the agent can use them without viewing them. The two companies are therefore talking past each other: Amazon is defending control over access, product presentation and customer data; Meta is pitching delegated convenience for consumers.

Amazon has already pursued similar restrictions against agents linked to Perplexity, Google and OpenAI. But shutting out assistants carries a cost, too: as more shoppers use AI to decide and buy, retailers risk putting a gate in front of the next major stream of commerce.