🔍 Read the full analysis: When AI Shops: Amazon Says No, Walmart Says Yes. Who Owns The Customer? on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Walmart says it is opening an Agent Gateway for approved AI shopping agents, while Amazon has blocked Meta’s Muse browser agent. The split puts retailers, AI platforms and publishers in competition over who controls the shopping relationship and receives credit for a sale.
Walmart is opening a structured route for AI shopping agents, while Amazon has blocked Meta’s Muse, making access to retailers a dividing line in the emerging market for automated shopping. The opposing approaches raise a practical question for shoppers and businesses: when an AI finds and buys a product, who controls the customer relationship and gets credit for the sale?
Walmart announced its Agent Gateway and Agent Profiles on Oct. 21. The gateway is designed to let approved agents search Walmart inventory and make purchases through a structured connection, rather than behaving like a person clicking through the store’s website. Walmart says profiles can give agents permissions and spending limits. The announcement describes an access model; it does not establish how widely the gateway is available or how many transactions it has processed.
Amazon has taken a different position with Muse, Meta’s browser-based personal agent. After Amazon blocked the agent in September, users received a message saying that continued access by an unauthorized AI agent violated the site’s Conditions of Use. Amazon said Muse had not identified itself as an agent and appeared to capture and store customer credentials. Meta’s launch post disputed the credentials concern, saying Muse had no visibility into users’ passwords or payment methods. The companies’ statements do not resolve their disagreement over how the agent handled access.
Meta said on its Oct. 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 reported an average purchase above $200. These are company-reported figures, not independently verified transaction data. The cited material does not specify how many users completed purchases or how the average was calculated.
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 Shopping Relationship
The dispute is about more than whether an agent can reach a product page. If shoppers ask an AI to compare goods and make purchases, the agent may become the main interface between the consumer and the retailer. Retailers want control over inventory, checkout and customer data; AI companies want their assistants to remain useful across stores. The system that wins access could influence where shoppers discover products and how merchants compete for them.
There is also a business-model question. Meta’s Mark Zuckerberg said the company expected to profit by taking a small fee from transactions. Amazon, meanwhile, has a large advertising business: GeekWire reported that Amazon generated more than $68 billion in advertising revenue last year. That figure illustrates the commercial stakes, but it does not establish that agent shopping will replace or materially reduce that revenue.
Publishers and affiliate businesses may also be affected. Amazon Associates commissions generally depend on shoppers using tracked links. If an agent reads a review and completes a purchase through another route, the publisher may not receive a commission. The scale of that potential change is not established by the information available, but it highlights how automated shopping could disrupt existing referral systems.
For shoppers, agent access and purchasing authority are separate issues. A tool that can research products may also be able to place an order, depending on its permissions and the retailer’s systems. Spending caps, purchase approval and clear return terms matter because a fast checkout does not determine which party handles a refund or dispute.
Products Worth Considering
Amazon Kindle Paperwhite (12th Generation, 2024), 16 GB, 7-inch Display, Jade
Sonos Era 100 Wireless Smart Speaker, Black
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
As an affiliate, we earn on qualifying purchases.
From Browser Bots to Retail Gateways
The dispute developed quickly in September and October. Meta launched Muse on Sept. 8, describing it as an agent that could open a browser, fill out forms and negotiate for users. The company said it runs in a dedicated virtual machine in Meta’s cloud and uses Link by Stripe for payment through a one-time-use card. Amazon blocked it on Sept. 20. Meta later announced retail connections with companies including Walmart, Best Buy, Gap, Sephora and Wayfair.
Retailers and technology companies are trying several ways to make agent shopping work. One is ordinary browser use, where software clicks through pages as a person would. Another is a retailer-approved connection, such as Walmart’s gateway or Shopify’s tools for agent-assisted checkout. A third is a shared protocol intended to define how agents communicate with stores and what customers have authorized. These systems are at different stages, so their announcements should not be read as evidence of a single settled standard.
One measure of the current limits comes from HUMAN Security data cited in the source report: in its September data, 81% of agent requests went to product and search pages, while 2.4% went to checkout and payment. Those are shares of agent requests in that dataset, not a count of purchases or a market-wide conversion rate. They suggest agents are more often researching than completing transactions, but the source does not provide enough detail to generalize beyond the measured activity.
“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, Purchases and Safeguards
Several practical questions remain open. Walmart has announced its gateway and profiles, but the available information does not say how many agents or shoppers can use them now, what the complete approval process entails, or how purchases and returns will be handled across participating services. It is also unclear whether retailers will adopt common rules or keep separate systems.
The disagreement between Amazon and Meta over Muse’s identification and access has not been independently resolved in the source material. Meta’s user and purchase figures are company-reported, and no independent transaction data is provided. Nor is there enough information to determine how frequently agents complete checkout, how often purchases are reversed, or how commissions and advertising revenues may change as use grows.
Retailers Set the Next Access Rules
The next developments to watch are whether Walmart’s gateway moves from announcement to broader use, whether other retailers create similar approved connections, and whether agent and commerce protocols gain adoption beyond their initial partners. Meta’s reported user numbers may show interest in Muse, but repeat use and completed purchases will be more informative measures of whether shopping agents become a regular route to checkout.
For consumers, the immediate question is what an agent is allowed to do before they enable it: browse only, prepare a cart or place an order. Before authorizing purchases, shoppers should check spending limits, approval settings, payment arrangements and the retailer’s return policy. The unresolved contest is not simply which assistant can find a lower price; it is which company sets the terms when software shops on a person’s behalf.
Key Questions
What did Walmart announce?
Walmart announced an Agent Gateway for approved AI agents and Agent Profiles intended to set permissions, including spending limits. The available information does not establish the gateway’s rollout scale or transaction volume.
Why did Amazon block Meta’s Muse?
Amazon said Muse was an unauthorized agent that had not identified itself and appeared to capture and store customer credentials. Meta disputed the credential concern, saying Muse could not see users’ passwords or payment methods.
Can AI shopping agents already complete purchases?
Some systems are designed to support checkout, but the cited September dataset found that 2.4% of agent requests went to checkout and payment, compared with 81% to product and search pages. Those figures describe requests in that dataset, not all agent shopping or completed sales.
How could agent shopping affect affiliate publishers?
Affiliate commissions often depend on a shopper clicking a tracked link. If an agent researches a publisher’s content but buys through a different path, the referral may not be recorded. The size of any effect is not yet clear.
What should shoppers check before authorizing an agent?
Review whether the agent can purchase without approval, set a spending cap if available, understand which payment method it uses and check which retailer’s return policy applies.
Source: ThorstenMeyerAI.com
