Amazon and Meta have always been business partners, but September 2026 brought a clash few people saw coming.
The spark was Muse, Meta’s new personal AI agent, launched on September 8, 2026.
In less than a week, the app was already the top free download on the U.S. App Store, surpassing even ChatGPT.
It seemed like a perfect launch, but it didn’t last long.
Amazon decided to cut Muse off from its e-commerce platform, and what had been a partnership turned into a public standoff over privacy, security, and, at its core, a much bigger question:
When an AI starts shopping for you, who really owns the customer relationship?
That’s exactly what we’re going to break down here. 👇
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
What is Muse and why it spooked Amazon
Muse isn’t your average chatbot. It was designed by Meta to work as a full-fledged personal assistant, capable of carrying out multi-step tasks instead of just answering one-off questions. The agent connects to services like email, calendar, payments, restaurants, and, of course, shopping. Think of an agent that logs into your account, compares products, reads reviews, and completes orders without you ever needing to open a single app. In promotional images released by Meta, Muse appears comparing strollers and asking the user for approval before finalizing an order. That’s the level of autonomy it aims to deliver, and that’s exactly where the tension with Amazon started to build.
The big problem, according to Amazon, is that it never agreed to any of this. Meta didn’t notify Amazon that Muse would access its store, the agent doesn’t identify itself when browsing pages, and, according to the e-commerce company, it appears to capture and store customer credentials, which could create privacy and security risks. In other words, from Amazon’s perspective, there’s an unidentified third party roaming around inside user accounts, processing transactions, and handling sensitive data, all without the platform’s knowledge or consent.
The core issue wasn’t just technical. It was strategic. On top of running its flagship e-commerce operation, Amazon pulled in more than 68 billion dollars in advertising revenue last year, a business that depends directly on people browsing pages and seeing sponsored products. Allowing another company’s AI agent to do the shopping instead of the customer would threaten exactly that browsing flow and the data intelligence on which Amazon has built a huge chunk of its empire.
The block and the market reaction
Amazon’s decision to block Muse came after the company tried, unsuccessfully, to convince Meta to voluntarily exclude the e-commerce site from the agent’s experience. When no agreement was reached, on Sunday evening users who tried to shop on Amazon through Muse started seeing a popup with a message saying that continued access by an unauthorized AI agent violates Amazon’s Conditions of Use, which customers had already agreed to.
In an official statement, an Amazon spokesperson got straight to the point, saying the company considers it pretty straightforward that third-party applications offering to shop on behalf of customers at other businesses should operate openly and respect the decisions of service providers about whether or not to participate. The language was formal, but the message behind it was crystal clear: Amazon doesn’t want an unidentified middleman between itself and its customers, especially not one controlled by one of the biggest tech conglomerates on the planet.
The detail that makes this clash even more interesting is that the two companies work together on other fronts. Amazon products have been purchasable inside Facebook and Instagram since 2023, and in April Meta signed a multibillion-dollar deal to run its agentic AI workloads on Amazon’s cloud using Graviton chips. So heavyweight business partners now find themselves on opposite sides of a fight over who controls the online shopping experience. 🔥
On Meta’s side, the company had previously stated that Muse has no visibility into people’s passwords or payment methods, and that credentials a user shares go into secure storage so Muse can use them without actually seeing them, including passwords the person types into the browser itself. The company did not immediately respond to requests for comment on the block, but Amazon said it is in direct conversation with Meta about the matter.
Privacy or control? The question everyone is asking
At the heart of this entire standoff is a discussion that goes way beyond the two companies involved. When an AI agent like Muse starts acting as a middleman in e-commerce transactions, it can inevitably access account pages and order history, as long as the customer asks it to do so. Because the agent doesn’t identify itself, Amazon sees this as an undisclosed third party moving through customer accounts, processing purchases, and handling sensitive data without its knowledge or consent.
Amazon argues that agents operating openly help keep the customer experience safe and trustworthy, and that other services that shop on behalf of consumers typically do so with the retailer’s agreement. The company gave very concrete examples: food delivery apps and the restaurants they take orders for, or online travel agencies and the airlines they book flights with. According to the spokesperson, agentic third-party applications like Muse have exactly the same obligations, and that’s why Amazon asked Meta to remove Amazon from the experience.
On the other hand, blocking an AI agent based on privacy and compliance justifications raises another equally relevant question: to what extent can major e-commerce platforms decide which technologies their users are allowed to use? The block can also be read as Amazon protecting its own shopping ecosystem. The company launched Alexa for Shopping in May, an AI agent that searches products and makes recommendations, and it also has Buy for Me, an agentic feature that finds items on external brand websites. The difference, according to Amazon itself, is that Buy for Me identifies itself and allows brands to opt out. The debate over who really controls the customer relationship is just getting started. 🤔
The legal battle behind the dispute
This standoff isn’t Amazon’s first against external AI agents. Over the past year, the company has been trying to keep these agents off its site. It sued Perplexity over the Comet browser and moved to block shopping agents from Google and OpenAI. In the case of ChatGPT, for instance, Amazon blocked OpenAI’s new shopping agent, a move that generated plenty of buzz about the future of AI-assisted commerce.
The dispute with Perplexity helps illustrate the legal terrain Amazon is navigating. The company obtained a preliminary injunction against Perplexity in March, but lost it on August 4 when the Ninth Circuit ruled that it was the user, not the AI company, who was accessing Amazon’s computers under the federal anti-hacking law. The court denied Amazon’s request for a rehearing on September 10. Still, the decision left one door open: claims based on contracts and terms of service. It’s no coincidence that the message Muse users are seeing now doesn’t accuse anyone of hacking but instead cites Amazon’s Conditions of Use.
How Muse works under the hood
It’s also worth understanding how Muse operates on a technical level. Meta launched it on September 8, 2026 as a personal agent that handles multi-step tasks, connecting to services like email, calendar, payments, restaurants, and shopping. It’s free with paid subscription tiers and is available on iOS, Android, the muse.ai website, and WhatsApp.
According to Meta, the agent runs in a secure virtual machine with its own browser and confirms with the user before taking sensitive actions like sending an email or making a purchase. There’s also a separate monitoring agent called Sentinel that must approve anything Muse sends out to the internet. The connection mechanism is right at the center of the dispute: if a service has a public API, Muse connects to it using credentials provided by the user. If there’s no API at all, Meta says the agent can use the service through a browser in the same way you would.
This browser-based approach helps explain Muse’s lightning-fast popularity. One week after launch, it became the number one free app on the U.S. App Store, ahead of ChatGPT. Early users reported very practical uses, like switching a car insurance policy, hunting for coupon codes at checkout, and building shopping carts on online grocery stores.
What this changes for the future of AI agents in e-commerce
This episode between Amazon, Meta, and Muse will likely mark a turning point in how the market thinks about integrating AI agents into e-commerce platforms. For the first time, we saw in a very concrete and very public way what happens when an autonomous shopping agent tries to operate on territory belonging to another company with equally massive interests in data and customer relationships. Amazon’s block isn’t just a response to Meta. It’s also a signal to the entire AI development ecosystem that major platforms won’t open their doors without very clear rules about who can come in and under what conditions.
For developers and companies building AI agents with the ability to carry out transactions, the Muse case serves as an important wake-up call. An agent’s technical ability to shop autonomously is one thing, but the viability of operating inside the walled gardens of major e-commerce platforms is something else entirely. It will take formal agreements, layers of compliance with each platform’s policies, and, most likely, pushback that goes far beyond purely technical concerns. The battlefield is now political, commercial, and legal just as much as it is technological.
And for the end user, who feels the impact of all this the most in everyday life, the picture taking shape is one of growing fragmentation: different AI agents will work well in some environments and be blocked in others, depending on the commercial interests behind each platform. This could create a user experience far less seamless than what Muse promised to deliver. The idea of a truly universal AI assistant that shops for you anywhere without friction is still a long way from reality. And cases like this show that the biggest obstacles to getting there aren’t just technical, but also power struggles between giants that refuse to give up control over their digital territories. 🚧
