The shopping cart battleground: what Amazon's shopping agent standoff means for UX

Amazon blocks third-party AI shopping agents to keep control of the customer journey and its advertising revenue. For customers, the result is a failed task, which makes it a UX problem as well as a legal one.

User experience is more than a design discipline. It's now the primary competitive battleground in digital commerce, and Amazon is proving it in real time by blocking AI shopping agents from Perplexity, Google, OpenAI and, most recently, Meta.

How we got here

In November 2025, Amazon accused Perplexity's Comet browser, an AI agent that can search and buy on a user's behalf, of browsing Amazon.com disguised as human traffic. Amazon blocked Comet and took the dispute to court. It won an early injunction in March, but lost it on appeal in August, when the Ninth Circuit ruled that the user, not the AI company, was the one accessing Amazon's systems. Amazon has kept blocking Comet, and has applied the same approach to Google's and OpenAI's shopping agents.

Meta's Muse, which launched on 8 September and has been topping the App Store charts, is the latest and most striking example. Despite the two companies being commercial partners, Amazon blocked it from shopping on Amazon.com on 20 September, citing violations of its terms of service, after Meta declined to exclude Amazon from the experience. Amazon says Muse doesn't identify itself when it browses and appears to store customer credentials. Meta disputes the security claims, and the outcome could shape how far AI agents are allowed to act in areas that have traditionally been human web activity. Blocking a commercial partner's agent looks less like rivalry and more like defending control of the human interaction itself, even when that interaction now happens through an agent.

The principle is simple. A user asks an AI agent to buy something on their behalf, and the agent can't do its job because Amazon blocks it. From the user's point of view, they can't complete their task, through no fault of their own. Whatever the legal justification, that is by definition a broken and frustrating user experience. The question this raises for every organisation is whether protecting your business model is worth the cost of visibly failing the customer you're supposed to be serving.

Why this matters

Amazon built its dominance on UX, and now faces its own logic turned against it. One-click ordering, "Look Inside", review sorting, personalised recommendations: these were the innovations that made Amazon the default. An AI agent buying a stroller won't get diverted into a bestseller or a flash sale on the way. That's the threat. These agents strip out the curated, monetised journey Amazon built its entire business model on.

Amazon isn't against agent shopping, either. It runs its own, including Rufus and Buy for Me. What it objects to is someone else sitting between it and the customer.

Platform control equals business control, and the gap between Amazon's words and actions is telling. Almost a year ago, Amazon's CEO Andy Jassy, told investors that Amazon expected to partner with third-party agents over time, as long as it could make the customer experience good, which he said it wasn't at that point. Since then, Amazon has blocked four major agents. Whoever controls the human interaction layer controls the commercial relationship, and that control is now being actively defended through legal mechanisms, to the detriment of the user's desired experience.

User appetite isn't waiting for the dispute to settle. Muse topping the App Store charts tells its own story. Trust still has some way to go: only 16% of shoppers are comfortable letting an AI assistant find and buy products for them, according to Coveo's 2026 Commerce Relevance Report, and most agents are still used for research rather than purchases. But the interest is clear and growing. The experience that best serves the business and the one that best serves the user are pulling apart, and whatever friction Amazon introduces, people want to see what agent-mediated shopping can do for them.

What this means for your organisation

1. Audit your experience for agent vulnerability. If an AI agent bypassed your curated journey entirely, what commercial value would you lose, and what friction would customers be glad to see removed?
2. Design for transparency and choice. Amazon distinguishes between agents that identify themselves and those that don't, and its own Buy for Me agent identifies itself. Is that a workable principle for resolving the conflict between user experience and commercial defences?
3. Decide now, not later, how you want to be found by agents. Resisting the shift may buy time, but it may not be a long-term strategy.

Twenty years ago, user experience was a nice-to-have. A decade ago, it became a differentiator. Today, it's the battleground itself. Amazon can block the agents, but it can't block the shift in how people expect to shop.

At User Vision, we help organisations navigate exactly this kind of strategic UX challenge, from understanding how AI agents might disrupt customer journeys to designing experiences that hold their value regardless of how users choose to access them. If your organisation needs to rethink its experience strategy for an agent-mediated future, let's talk.