🔍 Read the full analysis: AI Agents At Checkout: Amazon Blocks, Walmart Welcomes on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Amazon blocked Meta’s Muse shopping agent, citing its terms and concerns about how the agent accesses customer accounts. Walmart, by contrast, has announced an Agent Gateway and profiles with spending limits for agents. The split highlights an emerging contest over how AI shopping works, who controls the customer relationship and how purchases are attributed.
Amazon has blocked Meta’s Muse from shopping on its site, while Walmart is opening an Agent Gateway for AI shopping agents and introducing profiles that can set spending limits. The opposing moves show how retailers are choosing whether to admit outside agents—and who gets control when software shops on a customer’s behalf.
A Verge reporter asked Muse where granola bars were cheapest per piece. The agent named Sam’s Club, Walmart and Costco, then said it could not check Amazon because the retailer had blocked it. The example illustrates a practical limit on agent-led shopping: an AI may compare only the stores it can access.
Amazon’s block took effect on September 20, according to the supplied account. A message shown to users said 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 disputed the credential concern: its launch post said Muse had no visibility into users’ passwords or payment methods.
Walmart announced its Agent Profiles and Walmart Agent Gateway on October 21. The gateway uses the Model Context Protocol, or MCP, to let approved agents search inventory and make purchases through a structured connection rather than operate a shop’s ordinary webpages. Walmart’s announced profiles include controls such as spending limits. The announcement marks a different approach from Amazon’s restriction, though the supplied material does not establish how widely the gateway is available or which agents can use it now.
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 Checkout
The dispute is about more than whether an AI can find a lower price. It concerns access to inventory, checkout and customer accounts, as well as which company receives credit when an agent completes a sale. A retailer that admits outside agents may make its products easier to find and buy; a retailer that blocks them can keep purchases within channels it controls. The trade-off includes security, customer experience and commercial influence.
Meta says Muse shopping has grown quickly: at its October 21 earnings call, the company reported that more than 10 million people in the United States and Canada had set up Muse in about six weeks. Meta also said shopping was the most common use and the average purchase was more than $200. Those figures are Meta’s own claims, not independent measurements, but they suggest why retailers are paying attention to agent access.
Retailers and publishers also have different incentives. The source material, citing GeekWire, says Amazon earned more than $68 billion in advertising revenue last year. Meta chief executive Mark Zuckerberg said the company expected to earn “a small fee from transactions” through retail connectors. Affiliate publishers face a separate attribution issue: if an agent reads a review but buys through another route, a tracked referral click may not be recorded. These are potential effects of the changing purchase path, not confirmed losses for any specific company.
For shoppers, convenience does not remove the need to understand what an agent is allowed to do. Purchase approval, spending caps and payment controls matter because an agent could act across stores with different terms and return policies. The source material describes Meta’s Muse as using a one-time-use card through Link by Stripe, but that does not answer every question about refunds, disputes or responsibility when a purchase goes wrong.
Popular right now
Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
Shark Navigator Lift-Away Upright Vacuum (NV360)
Dyson Ball Animal 3 Corded Upright Vacuum
As an affiliate, we earn on qualifying purchases.
How Retailers Set Agent Access
Agent shopping can happen through a browser, a retailer’s own connection or shared technical protocols. A browser-based agent interacts with pages much as a person would, potentially using the customer’s login. That can make it work across many sites, but retailers may have difficulty distinguishing an agent from a person and may object to how credentials or page data are handled. The supplied account says Muse runs in a dedicated computer in Meta’s cloud; it also describes Perplexity’s Comet as operating on a user’s own computer.
A retailer-provided gateway offers a more structured route. Walmart’s MCP-based system is intended to let an approved agent search and buy without mimicking clicks on consumer webpages. Shopify has also made tools available for checkout through agents, according to the source material. These integrations can give merchants more control over what an agent can access, though they require retailers and agent developers to support the same interfaces.
Several protocol efforts are competing to define those connections. The source material describes the Personal Agent Protocol, or Poppy, backed by Sierra, Meta, Walmart, Shopify and Stripe among others, as a draft published October 9. It says Google and Shopify’s Universal Commerce Protocol covers a broader shopping journey, while OpenAI and Stripe’s Agentic Commerce Protocol focuses on checkout. Shared standards could make connections easier to build, but the existence of a protocol does not mean every retailer or agent has adopted it.
The timeline is moving quickly. Meta launched Muse on September 8, then announced retail connectors with Walmart and other businesses on September 23. Shopify added structured checkout tools on September 28. The October 21 Walmart announcement followed those steps. Separately, the supplied source says OpenAI had previously reduced the role of its ChatGPT Instant Checkout approach, saying the initial version lacked the flexibility it wanted and that it would focus on product discovery while merchants used their own checkout experiences.
“Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use, to which our customers have agreed.”
— Amazon, in a message shown to Muse users
Access, Liability and Adoption Remain Open
It is not clear how many customers are actively completing purchases through Muse, how often agents reach checkout, or how much revenue those purchases generate. Meta’s figures count people who set up Muse and report shopping as its most common use; they do not establish the share of users who have completed a purchase. The source material cites HUMAN Security data from September saying 81% of agent requests went to product and search pages, while 2.4% went to checkout and payment. Those figures refer to the requests measured in that dataset, not all AI shopping activity, and the material does not provide a broader market baseline.
The source material does not specify the gateway’s rollout schedule, its approval process, or whether Walmart will permit every outside agent to connect. It is also unclear what standards will prevail as protocol projects develop, whether retailers will support several at once, and how agents will disclose purchases or preserve referral attribution. The source describes Poppy as a draft, not a finalized standard.
Responsibility for mistakes also remains unsettled in the information provided. The practical questions include who handles an unauthorized purchase, how refunds and returns work, what happens if an agent misreads a product listing, and which company is accountable when a payment or account is compromised. Amazon and Meta have offered conflicting accounts of Muse’s access and credential handling; the supplied material does not independently resolve that dispute.
Watch Gateway Rollout and Standards
The next indicators will be which agents Walmart approves, how its spending controls work in practice, and whether customers can review or cancel orders before payment. Retailers may also clarify how they handle returns, disputes and data access for purchases made through an agent.
Protocol adoption will be another measure to watch. Poppy’s draft and the other commerce standards will need participation from retailers, payment firms and agent developers before they can support consistent shopping across services. For now, the clearest confirmed difference is that Amazon has blocked Muse while Walmart is building a structured access route. Whether that divide persists—or shifts as technical and commercial terms change—remains to be seen.
Key Questions
Why did Amazon block Meta’s Muse?
Amazon said Muse was an unauthorized agent that did not identify itself and appeared to capture and store customer credentials. Meta said Muse had no visibility into users’ passwords or payment methods. The supplied material does not independently resolve the disagreement.
What is Walmart’s Agent Gateway?
Walmart announced an MCP-based gateway intended to let approved AI agents search inventory and make purchases through a structured connection. Walmart also announced Agent Profiles with controls such as spending limits. The rollout details are not specified in the source material.
Can Meta’s Muse still compare Amazon prices?
In the reported granola-bar example, Muse said it could not check Amazon because Amazon had blocked it. The agent compared Sam’s Club, Walmart and Costco instead.
Does the reported Muse usage mean millions of people bought something?
No. Meta said more than 10 million people in the United States and Canada set up Muse in about six weeks and that shopping was its most common use. Those figures do not state how many users completed purchases.
What should shoppers check before using an AI agent to buy?
Review which purchases require approval, set a spending limit, understand the payment method, and check the retailer’s return and refund rules. The source material also points to one-time-use payment cards as a possible control, but does not describe protections for every agent or retailer.
Source: ThorstenMeyerAI.com
