🔍 Read the full analysis: Amazon And Walmart’s Split On AI Agents Raises A Customer Question on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Walmart has announced tools that let approved AI agents search and shop through a dedicated gateway, while Amazon has blocked Meta’s Muse agent from accessing its site. The dispute highlights unresolved questions about who controls the customer relationship, how purchases are authorized and whether agents can be held to store policies.
Walmart is opening a dedicated gateway for approved AI shopping agents, while Amazon has blocked Meta’s Muse from accessing its store, setting up a dispute over how automated systems can browse and buy on behalf of customers. The split matters because shoppers may increasingly delegate product searches and purchases to agents, but retailers differ on whether those agents can enter their sites and under what rules.
Walmart said on October 21 that it was introducing an Agent Gateway based on the Model Context Protocol, along with Agent Profiles that can set permissions and spending limits. The gateway is designed to let approved agents search Walmart’s inventory and make purchases through structured connections rather than behaving like a person clicking through ordinary webpages. The announcement does not establish that every agent can use the service or that all purchases will occur without customer confirmation.
Amazon, by contrast, has restricted Muse, Meta’s browser-based personal agent. A notice shown to users said continued access by an unauthorized AI agent violated Amazon’s Conditions of Use. Amazon said Muse had not identified itself as an agent and raised concerns about how it handled customer credentials. Meta has said Muse cannot see users’ passwords or payment details. Those statements describe the companies’ positions; the source material does not independently resolve the disagreement.
Meta reported 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. The company said shopping was its most common use and that the average purchase exceeded $200. These are Meta’s figures, not independently verified measures of completed agent purchases. A separate HUMAN Security analysis cited in the source said 81% of agent requests went to product and search pages, while 2.4% went to checkout and payment pages, suggesting that browsing is currently more common than completed checkout.
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 an Agent-Led Purchase
The dispute is about more than whether software can open a retailer’s website. Retailers want control over access, checkout and customer data; agent makers want to make shopping easier and may seek a share of the value their services generate. If a shopper asks an agent to compare products, choose one and complete an order, the retailer may have less direct contact with the customer—and less control over how the choice is presented.
That could affect businesses that rely on referrals. Amazon’s Associates program pays commissions tied to tracked links, according to the source material. If an agent reads a review and completes a purchase without a credited click, the publisher that informed the decision may not receive a commission. The impact will depend on how retailers and agent providers design attribution and payment systems; the available information does not show how widespread such lost commissions are.
For shoppers, the new convenience brings practical questions about authorization and accountability. A spending limit or approval step can reduce the chance of an unwanted order, but it does not by itself explain whose return policy applies, what happens if an agent selects the wrong item, or how a payment can be disputed. Those rules remain part of the product experience, not just a technical detail.
Products Worth Considering
Sonos Era 100 Wireless Smart Speaker, Black
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
Anker SOLIX C200 DC Portable Power Station, 192Wh, 200W
As an affiliate, we earn on qualifying purchases.
A Short Race to Set Shopping Rules
The move follows a rapid sequence of product and protocol announcements. Meta launched Muse on September 8, describing an agent that can use a browser and fill out forms. On September 20, users encountered Amazon’s access block. Three days later, Meta announced retail connectors that included Walmart, Best Buy, Gap, Sephora, Wayfair and Expedia, and Mark Zuckerberg said Meta expected to earn a small fee from transactions.
Other companies are trying to make agent purchases work through structured tools rather than ordinary browser activity. Shopify has introduced tools for agents to work with checkout, while Walmart’s gateway is another retailer-led route. Meta and Sierra have also promoted the Personal Agent Protocol, known as Poppy, with Walmart, Shopify and Stripe among its backers. The source describes its October 9 release as a draft; the protocol is intended to communicate an agent’s identity and the customer’s permissions across services.
There is no settled standard. The source also identifies Google and Shopify’s Universal Commerce Protocol and OpenAI and Stripe’s Agentic Commerce Protocol as competing approaches. OpenAI’s earlier Instant Checkout illustrates the limits of a checkout-focused model: in March 2026, the company said it was stepping back from that initial version so merchants could use their own checkout experiences while OpenAI focused on product discovery. The competing efforts show that agreeing on technical connections is only one part of the problem; retailers, payment providers and agents also need rules for consent and responsibility.
“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, Consent and Liability Remain Open
It is not yet clear how many completed purchases Muse has made, how Meta calculated its reported average purchase value, or whether that figure represents all users or a narrower group. Meta’s setup and usage figures are company-reported, while the cited traffic percentages measure requests to page types rather than the share of shoppers who ultimately buy.
Walmart’s announcement does not specify which outside agents will be approved, how customer authorization will be verified in every transaction, or how disputes and returns will work when an agent places an order. Amazon’s stated concerns about Muse and Meta’s response remain competing company claims in the material provided. Nor is there evidence that one protocol will prevail: the named standards cover different parts of commerce and remain at different stages of development.
It also remains uncertain how retailers will attribute sales generated by agents and whether publishers or other referral sources will be paid when there is no conventional tracked link. The source describes a potential problem for affiliate commissions, but provides no market-wide estimate of lost revenue.
Watch the First Agent Transactions
The next test is whether Walmart’s gateway becomes usable by a broad set of agents and whether customers can understand and control the permissions they grant. Retailers and agent providers will also need to explain what approval is required before an order, how spending caps operate, and which company handles refunds or errors.
Further developments are likely to come through product rollouts and protocol revisions rather than a single industry-wide decision. Readers can watch for named integrations, published terms for agent access, and clearer reporting on completed purchases—not only searches or agent setups. For now, shoppers considering an agent purchase should check the payment method, set limits or require approval where available, and confirm the retailer’s return terms before relying on an automated checkout.
Key Questions
What did Walmart announce?
Walmart announced an Agent Gateway using the Model Context Protocol and Agent Profiles that can set permissions and spending limits. The tools are intended to let approved agents search and shop through a structured connection.
Why is Amazon blocking Meta’s Muse?
Amazon said Muse was an unauthorized agent that did not identify itself and raised concerns about customer credentials. Meta says Muse cannot see users’ passwords or payment methods. The available information does not independently settle the dispute.
Are AI agents already making many online purchases?
The cited HUMAN Security data said 81% of agent requests went to product and search pages and 2.4% to checkout and payment pages. Those figures describe requests, not completed orders. Meta separately reported more than 10 million Muse setups in the United States and Canada, but did not establish how many resulted in purchases.
What should shoppers check before using an AI shopping agent?
Check whether purchases require approval, set a spending limit if available, and review the payment method and retailer’s return policy. The terms for handling an agent’s mistakes or disputed orders are not yet consistent across services.
Source: ThorstenMeyerAI.com
