🔍 Read the full analysis: Amazon’s AI Shopping Stance Differs From Walmart’s. Here’s Why on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Amazon has blocked Meta’s Muse from shopping on its site, citing its rules for unauthorized agents and concerns about how the tool operates. Walmart, by contrast, announced an Agent Gateway and shopper profiles with spending limits, reflecting competing approaches to AI-assisted purchases. The broader fight is over access, checkout control and the customer relationship; many agents still appear to focus more on product research than completing purchases.
Amazon has blocked Meta’s Muse AI agent from accessing its shopping site, while Walmart is opening a dedicated route for approved shopping agents, putting two retail strategies in direct contrast. The split matters because an agent that shops on a customer’s behalf could influence where purchases happen—and which retailer, platform or publisher retains the relationship with the buyer.
Amazon’s restriction became visible on September 20, when Muse users encountered a notice saying that continued access by an unauthorized AI agent violated Amazon’s Conditions of Use. Amazon said Meta had not told it Muse would shop on the site, that the agent did not identify itself, and that it appeared to capture and store customer credentials. Meta’s launch post said Muse had no visibility into users’ passwords or payment methods. The competing descriptions leave the technical dispute unresolved.
Walmart announced its Agent Gateway and Agent Profiles on October 21. The gateway is designed to let approved agents search Walmart inventory and make purchases through a structured interface, rather than mimicking a person’s clicks in a browser. Walmart also described profiles that can set spending limits. The announcement signals a willingness to host agent-led shopping under defined conditions, rather than simply allowing any automated browser to use the retail site.
Meta said on its October 21 earnings call that more than 10 million people in the United States and Canada had set up Muse in roughly six weeks, and that shopping was its most common use. Meta also said the average purchase was more than $200. Those are company figures, not independently verified measures of completed shopping through agents. Meta previously said it expected to make money by taking a small fee from transactions.
Agentic shopping
Three ways an AI agent buys for you
Every fight between shops and AI agents comes down to which of these three routes the agent uses, and whether the shop agreed to it.
The browser, like a human
The agent opens pages built for people, signs in with your login and clicks through checkout. Works almost anywhere, which is why shops object: they can't tell the agent from you.
Examples: Meta's Muse (from a virtual machine in Meta's cloud), Perplexity's Comet (from your own computer)
The shop's own agent door
Approved agents search and buy through a structured interface instead of clicking through pages. The shop sets the rules.
Examples: Walmart Agent Gateway (MCP, beta), Shopify MCP server and WebMCP checkout tools
Shared protocols
Common rules for who the agent is, what you allowed, and how the checkout and payment are confirmed.
Examples: Personal Agent Protocol (draft), Universal Commerce Protocol, Agentic Commerce Protocol
The protocols competing to be the standard
| Protocol | Backers | Covers | Status |
|---|---|---|---|
| Personal Agent Protocol (Poppy) | Sierra, Meta, Walmart, Shopify, Stripe + 35 design partners | Agent identity, customer permission, one session across site, API and company agent | Draft 0.1 (9 Oct 2026) |
| Universal Commerce Protocol (UCP) | Google, Shopify; site also lists Etsy, Wayfair, Target, Walmart, Amazon, Microsoft | Discovery, cart, checkout, orders; payment mandates via AP2 | Published |
| Agentic Commerce Protocol (ACP) | OpenAI, Stripe | Checkout between agent and merchant | Open source; ChatGPT's own checkout scaled back in March 2026 |
What a careful agent purchase looks like
- You give the task and a limit, for example a $50 cap.
- The agent identifies itself to the shop and gets only the access you approved.
- It prepares the cart and shows you an approval card.
- Payment goes through a token or one-time card, so the agent never sees your card.
- The shop’s normal returns apply if the agent got it wrong.
Sources: Meta (8 Sep 2026), Walmart (21 Oct 2026), TechCrunch on Shopify (28 Sep 2026), Sierra (9 Oct 2026), ucp.dev, agenticcommerce.dev, OpenAI (24 Mar 2026). Purchase steps: Walmart’s agent terms and Meta’s Muse description.
Powered by Thorsten Meyer AI
Who Controls the Agent-Led Sale
The dispute is about more than whether a bot can open a product page. If a shopper delegates research and purchasing to an AI agent, the agent’s operator may influence which products are compared, which retailer gets the order and what information is available at checkout. Retailers want control over access and transactions; agent companies want to make shopping easier across multiple stores.
The commercial stakes include advertising and referral revenue. GeekWire reported that Amazon earned more than $68 billion from advertising last year. The source material does not specify the precise accounting period or provide a comparable figure for Meta, but the scale illustrates why control of product discovery matters. Affiliate publishers also face a potential attribution problem: if an agent reads a review and completes a purchase without using a tracked link, the publisher may not receive a commission under traditional click-based arrangements.
For shoppers, an agent could reduce the work of comparing prices, but convenience does not settle practical questions about authorization, payment and returns. A spending cap or purchase approval may help limit unwanted orders. Buyers also need to know which retailer’s return policy applies when an agent moves between platforms or checks out through a third-party system.
Products Worth Considering
Sonos Era 100 Wireless Smart Speaker, Black
Bose SoundLink Flex Bluetooth Speaker (2nd Gen), Portable Waterproof Speaker, Black
Apple iPad Air (4th Generation, 10.9-inch, Wi-Fi, 256GB) – Space Gray
As an affiliate, we earn on qualifying purchases.
Browser Access Versus Retailer Gateways
Shopping agents can reach stores in different ways. A browser-based agent acts through pages designed for people: it may search, click and fill forms. That approach can work across many websites, but retailers may have difficulty distinguishing the agent from a customer and may object to automated access. Muse uses a dedicated virtual machine in Meta’s cloud, according to the source material. Amazon’s objection centers partly on how Muse accessed its site and handled credentials; Meta disputes the concern about visibility into passwords and payment details.
A retailer-controlled gateway offers another model. Walmart’s Agent Gateway uses the Model Context Protocol, or MCP, to give approved agents a structured way to search and transact. Shopify has also introduced tools intended to let browser agents use structured checkout functions. Separately, Meta and Sierra announced the Personal Agent Protocol, also called Poppy, with Walmart, Shopify and Stripe among its backers. The draft, published October 9, describes ways to communicate an agent’s identity and the customer’s permissions; payments are left to extensions.
These efforts are still developing and do not amount to one settled industry standard. The source material also describes Google and Shopify’s Universal Commerce Protocol and the OpenAI-Stripe Agentic Commerce Protocol. Their existence shows that technology companies and retailers are trying to define how agents discover products and complete orders, but protocol support does not mean every retailer accepts every agent.
“Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use, to which our customers have agreed.”
— Amazon, in a notice shown to Muse users
Access, Orders and Buyer Protections
It is not clear how many Muse users completed purchases through the agent, how often shopping agents reach checkout, or how much revenue they currently generate for retailers and platforms. Meta’s user and purchase figures are company-reported, and the source material does not establish whether the average purchase represents a completed transaction or a broader measure of activity.
Data cited from HUMAN Security’s September analysis said 81% of agent requests went to product and search pages, while 2.4% went to checkout and payment. Those figures describe requests in that dataset, not all shopping agents or all online purchases; the source does not provide a comparison baseline. It is also unclear how Walmart will approve agents, enforce profile spending limits, or resolve disputes when an agent acts incorrectly. The competing protocols may change, and their adoption across retailers remains uncertain.
Retailers Set the Rules for Agents
The next test is whether Walmart’s gateway moves from announcement to dependable, widely used shopping infrastructure—and which agents and merchants it supports in practice. Amazon’s current position, as described in the source material, is to require agents to identify themselves under its terms and to restrict access it considers unauthorized. No change to that policy is confirmed here.
Shoppers and businesses will also be watching whether protocol groups settle on compatible ways to convey identity, permissions and payment instructions. Until access and accountability are clearer, consumers may want to review an agent’s permissions, set a spending limit and check the retailer’s payment and return terms before allowing it to place an order.
Key Questions
Why did Amazon block Meta’s Muse?
Amazon said Muse was an unauthorized agent, had not identified itself and appeared to capture and store customer credentials. Meta disputed the credentials concern, saying Muse could not see users’ passwords or payment methods.
How is Walmart’s approach different?
Walmart announced an Agent Gateway for approved agents to search inventory and buy through a structured interface. It also described Agent Profiles with spending limits, rather than relying only on agents acting through ordinary web pages.
Are AI agents already completing many online purchases?
The figures cited from HUMAN Security’s September data suggest agent requests were much more common on product and search pages than at checkout: 81% versus 2.4% in that dataset. Those numbers do not represent every agent or the share of all retail purchases.
What should shoppers check before letting an agent buy?
Review what the agent can access, set a spending cap or require approval for each order, and check which retailer handles payment, delivery and returns. The specific protections available depend on the agent and retailer.
Source: ThorstenMeyerAI.com
