A spokesperson for Amazon states, in a statement to Geekwire, that third-party applications offering to make purchases on behalf of customers should operate transparently and respect service providers’ decisions regarding participation. The spokesperson further stated that agent-based applications have the same obligations as other actors and that the company has requested Meta remove Amazon from the experience.
Meta previously announced that Muse does not have access to users' passwords or payment methods, and that these are stored in a protected environment. However, reports have emerged that Muse was able to read the content of users' messages without the necessary permissions being activated. The block is being implemented despite the established business relationship between Amazon and Meta, such as an agreement from last spring to run AI processes via Amazon’s cloud services.
The Battle for Customer Relationship
The incident with Meta is part of a broader strategy by Amazon to maintain control over the customer journey, an issue that has grown in importance for the e-commerce giant. In connection with the company's quarterly report in February, CEO Andy Jassy pointed to independent AI agents as the primary challenger to the company’s market position. This occurred against the backdrop of statistics showing a nearly 700 percent growth in referral traffic from AI platforms to e-commerce retailers during the 2025 holiday shopping season.
During the presentation, Jassy expressed skepticism about the ability of independent agents to deliver a reliable customer experience. He noted that the horizontal agents lack the customer’s shopping history, and that they have errors on many product details and prices. Amazon instead highlights its own closed AI assistant Rufus. The company also relies on factors such as a wide range, its own logistics and pricing. External AI agents, on the other hand, have the ability to search the web and compare offers neutrally.
Billions in Revenue at Stake
Significant economic interests are at the heart of the conflict. Amazon’s business model relies heavily on advertising revenue, which totaled over $68 billion last year. This revenue requires consumers to actively visit the platform and be exposed to sponsored products. If purchases are instead made by AI agents, Amazon risks losing a key revenue source.
The current blocking of Muse follows a similar pattern from last year, when Amazon initiated legal action against the AI startup Perplexity regarding their shopping assistant Comet. Amazon demanded that the bot identify itself, which Perplexity opposed. The AI company argued that its agent acts on direct instructions from a human and should therefore have the same permissions as the user. Perplexity also claimed that Amazon’s primary purpose was to protect revenue from upselling, as an AI bot, unlike a human, is not distracted by advertisements.
The initial legal process between Amazon and Perplexity resulted in a court ruling in August stating that it is the user, not the AI company, who acts on the platform under the federal data law. Since then, Amazon has adjusted its strategy to pursue the issue based on its own terms of use. The company has also begun to limit detailed information in order confirmations via email to make external data collection more difficult.
Andy Jassy has previously opened up to future collaborations with external actors, provided that the parties can find a value exchange that is commercially viable for both sides.