🔍 Read the full analysis: How Amazon And Walmart Differ On AI-Powered Shopping on ThorstenMeyerAI.com
TL;DR
Walmart has announced an Agent Gateway and Agent Profiles that let approved AI agents search and purchase through structured tools. Amazon has blocked Meta’s Muse agent and says outside agents must identify themselves and comply with its terms. The split puts control, shopper permissions and attribution at the center of emerging AI-powered shopping.
Walmart is opening a route for approved AI shopping agents while Amazon has blocked Meta’s Muse, setting up a clear divide over how automated assistants can browse and buy from major retailers. Walmart announced its Agent Gateway and Agent Profiles on October 21, 2026; Amazon says outside agents must identify themselves and follow its Conditions of Use.
Walmart’s Agent Gateway uses the Model Context Protocol, or MCP, to let approved agents search inventory and make purchases through structured connections rather than navigating the retailer’s consumer-facing pages. Its Agent Profiles are intended to give an agent defined permissions, including spending limits. The announcement describes infrastructure for agent access; it does not establish how widely the tools are available or how many shoppers are using them.
Amazon has taken a more restrictive approach to Muse, Meta’s browser-based personal agent. The Verge reported that Muse was blocked from Amazon on September 20. Amazon said the agent was unauthorized, did not identify itself and appeared to capture and store customer credentials. Meta disputes the credential concern: its launch materials said Muse has no visibility into users‘ passwords or payment methods. The disagreement has not been resolved in the information provided.
Meta reported on its October 21 earnings call that more than 10 million people in the United States and Canada set up Muse in about six weeks, and said shopping was its most common use. Meta also said the average purchase was more than $200. These are company-reported figures; the source material does not give an independent measure of completed purchases or explain 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
Control of the Shopping Journey
The dispute is about more than whether an automated assistant can open a product page. Whoever controls the agent may influence which products shoppers see, where they buy them and how retailers recognize or compensate the parties involved. Walmart’s approach could give it a role in setting the terms for agent access, while Amazon’s restrictions keep more of the shopping process within its own systems.
There are also consequences for publishers and referral businesses. The source cites Amazon Associates terms that tie commissions to tracked customer clicks. If an agent reads a review and completes a purchase without following its tracked link, the publisher may not receive credit. HUMAN Security data for September, as reported in the source, found 81% of agent requests went to product and search pages, compared with 2.4% to checkout and payment. Those figures describe request traffic in that dataset, not the share of all shopping agents that complete purchases.
For shoppers, the convenience comes with practical questions: what an agent is authorized to spend, whether a purchase needs approval and which retailer’s return rules apply. A smoother buying process does not, by itself, settle responsibility for an order or protect users from an unwanted purchase.
Popular right now
Samsung 42-Inch S90H OLED 4K Smart TV Bundle
LG C2 Series 65-Inch OLED evo 4K Smart TV (OLED65C2PUA, 2022)
Apple AirPods Pro 3 Wireless Earbuds
As an affiliate, we earn on qualifying purchases.
How Retailer Access Is Changing
AI agents can shop in several ways. A browser-based agent such as Muse can use pages designed for people, potentially signing in and checking out as a user would. That flexibility can make it work across many sites, but retailers may have difficulty distinguishing the agent’s actions from a customer’s and may object to automated access. The source says Muse operates through a dedicated virtual machine in Meta’s cloud and uses Link by Stripe, which issues a one-time-use card.
A second route is a retailer’s own agent connection. Walmart’s MCP-based gateway and Shopify’s structured checkout tools are examples of systems that let agents interact through defined functions instead of simulating clicks. A third route is shared protocols intended to carry information about the agent, customer permissions or checkout. The source lists the Personal Agent Protocol, backed by Meta, Sierra, Walmart, Shopify and Stripe among others, as a draft published October 9. Google and Shopify’s Universal Commerce Protocol and the OpenAI-Stripe Agentic Commerce Protocol are other efforts, with different scopes.
Recent events show that these arrangements are still taking shape. Meta announced retail connectors including Walmart on September 23. Shopify added tools for browser agents to complete checkout, according to the source, and protocol proposals followed in September and October. The source also reports that OpenAI stepped back from its initial ChatGPT Instant Checkout approach in March 2026, saying it would let merchants use their own checkout experiences while focusing on product discovery. That history cautions against treating protocol announcements as proof of widespread, routine agent purchases.
„Muse has no visibility into people’s passwords or payment methods.“
— Meta, in its Muse launch materials
What Agent Shopping Can Do
Walmart’s announcement does not show how many retailers or agents are already using the gateway, or how often it leads to completed purchases. Meta’s figures count people who set up Muse, according to the company; they do not establish how many became regular users or how many purchases were completed. The source also does not provide a common, independent measure for comparing shopping activity across Amazon, Walmart and other services.
The dispute over Muse’s access and handling of credentials remains contested in the supplied material. It is also unclear how retailers will handle refunds, returns, disputed orders and customer support when an agent participates in a purchase. Protocols under development may help identify an agent and record a customer’s permissions, but their adoption and practical protections remain unsettled.
Finally, the source reports differing approaches to agent traffic and payment, but not the commercial terms Walmart will apply to all partners or how Amazon may change its policy. It is too early to conclude which model will attract more shoppers or generate more sales.
Retailers Set the Next Rules
The next test is whether Walmart’s gateway and profiles move from announcement to regular use by approved agents and shoppers. That will depend on which partners connect, what permissions are available and whether transactions can be completed reliably. The Personal Agent Protocol is described as a draft, while other commerce protocols are at different stages; their adoption and compatibility will be worth watching.
Amazon’s policy toward outside agents, and any response to retailer-led gateways, will help show whether its restriction of Muse is a specific dispute or part of a broader access strategy. Shoppers weighing agent purchases should check spending approvals, payment settings and the retailer’s return terms before allowing an agent to place an order. The source material provides no confirmed schedule for further Amazon or Walmart changes.
Key Questions
What did Walmart announce?
Walmart announced an Agent Gateway using MCP and Agent Profiles with permissions such as spending limits. The tools are designed to let approved agents search inventory and make purchases through structured connections.
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 has no visibility into users‘ passwords or payment methods. The supplied information does not settle that dispute.
Does Meta’s user figure mean millions of purchases were completed?
No. Meta reported that more than 10 million people in the United States and Canada set up Muse in about six weeks, and said shopping was its most common use. That figure is not a count of completed purchases; the source does not give an independently verified transaction total.
How could AI agents affect shopping websites and publishers?
Agents may visit product and search pages without following the tracked links publishers use for referral commissions. HUMAN Security’s September data cited in the source found 81% of agent requests went to product and search pages and 2.4% to checkout and payment, but those figures do not measure all shopping activity or prove that a sale occurred.
What should shoppers check before letting an agent buy something?
Review whether purchases require approval, set a spending limit and check which payment method is used. Shoppers should also confirm the retailer’s return and refund terms, since the source material does not establish how every agent-assisted transaction will be handled.
Source: ThorstenMeyerAI.com
