🔍 Read the full analysis: What Retailers Need To Consider As AI Agents Start Shopping 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
AI agents are beginning to research and buy products, but retailers are responding with different rules: Walmart is building tools for approved agents, while Amazon has blocked Meta’s Muse agent. The stakes include control of checkout, customer relationships, data and referral revenue; the scale and reliability of agent-driven purchases remain uncertain.
Walmart is opening new channels for AI shopping agents, while Amazon has blocked Meta’s Muse agent from its site, sharpening a dispute over how automated services can search, compare and buy products. The split matters to retailers and shoppers because an agent may influence which store gets a sale—and who controls the customer relationship and payment process.
Walmart announced Agent Profiles and an Agent Gateway on October 21. The company says the tools will let approved agents access shopping functions through structured connections, with profiles that can apply customer permissions and spending limits. The move offers a route for agents to interact with Walmart without simply imitating a person clicking through a website.
Amazon has taken a different approach to Muse. The Verge reported that the agent was blocked from Amazon, where a message said continued access by an unauthorized AI agent violated the site’s Conditions of Use. Amazon said Meta had not told it Muse would shop there, that the agent did not identify itself, and that it appeared to capture and store customer credentials. Meta’s launch post said Muse could not see users’ passwords or payment methods. Those accounts differ, and the dispute over the agent’s access and handling of credentials is not resolved by the source material.
Meta reported that more than 10 million people in the United States and Canada had set up Muse within about six weeks, and said shopping was its most common use. The company also said the average purchase was more than $200. These are Meta’s figures, shared on its October 21 earnings call; the source provides no independent verification or breakdown of completed transactions. In a Verge test, Muse named Sam’s Club, Walmart and Costco as places to find cheaper granola bars, but said it could not check Amazon because the site had blocked it.
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
Retailers Set the Rules for Agents
For retailers, the choice is not simply whether to adopt AI. It is whether to let an outside agent search inventory, carry out checkout, or access customer accounts—and on what terms. A retailer-controlled connection can make an agent easier to identify and limit, but may also give the merchant more influence over the transaction. A browser-based agent can work across many sites, yet retailers may have less visibility into what it is doing.
The financial stakes extend to advertising and referrals. The source cites GeekWire’s report that Amazon earned more than $68 billion from advertising last year, while Meta has said it expects to earn “a small fee” from transactions. If agents steer shoppers without using a retailer’s existing ads or referral links, established ways of measuring and paying for a sale could change. That outcome is not settled: it depends on how often agents complete purchases and how merchants, platforms and publishers agree to attribute them.
Shoppers also face practical questions: whether an agent has permission to spend, what happens when it chooses the wrong item, and which company handles returns or disputes. Spending caps and approval steps could limit unwanted purchases, but the source does not establish a common standard for those protections across services.
Products Worth Considering
Sonos Era 100 Wireless Smart Speaker, Black
Bose TV Speaker (2nd Gen) Compact Soundbar with Bluetooth and HDMI Connectivity
Bose SoundLink Flex Bluetooth Speaker (2nd Gen), Portable Waterproof Speaker, Black
As an affiliate, we earn on qualifying purchases.
Three Routes Into Online Checkout
The recent changes show three approaches taking shape. One is the browser route: an agent visits ordinary shopping pages and acts much as a person would. Meta launched Muse on September 8, describing it as a personal agent able to open a browser and fill out forms. Its version runs in a computer environment in Meta’s cloud. This broad access can make an agent useful across sites, but it can also make the agent harder for a retailer to distinguish from a human shopper.
A second approach is a retailer’s own connection for agents. Walmart’s gateway is one example. Shopify has also introduced tools for browser agents to interact with checkout through structured functions rather than relying only on screen controls. A third is shared protocols intended to define how agents identify themselves, receive customer permissions and exchange shopping or payment information. The source describes several efforts, including the Personal Agent Protocol, a draft announced by Meta, Sierra and partners; Google and Shopify’s Universal Commerce Protocol; and the OpenAI- and Stripe-backed Agentic Commerce Protocol.
These standards are still developing, and their roles differ. OpenAI’s earlier Instant Checkout illustrates that a technical connection does not guarantee a lasting buying experience: in March 2026, the company said it was allowing merchants to use their own checkout experiences while it focused on product discovery. The immediate contest is therefore about both access and the commercial terms behind it—not just which protocol gains support.
“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 users, as reported by The Verge
Adoption and Accountability Remain Open
The reported Muse setup and purchase figures come from Meta, and the source does not say how many users completed an agent-led checkout, how often they use the service, or whether purchases were incremental. The granola-bar example shows that access varies by retailer, but it does not establish how consistently Muse or other agents can shop across sites.
There is also no settled answer on how agents should identify themselves, what evidence retailers may require, or who is responsible if an agent mishandles credentials or places an unwanted order. The source describes competing protocols and early retailer tools, not a broadly adopted standard. Amazon and Meta’s differing statements about Muse’s access and credentials remain claims by the parties, rather than an independently resolved account.
For publishers and affiliate businesses, the scale of any lost referral revenue is unknown. An agent may read a review without following a tracked link, but the material provides no measured estimate of the effect on commissions. The extent to which merchants, platforms and publishers will share attribution or fees is also unsettled.
Watch Retailer Access and Checkout
The next test is whether retailers make their agent connections available in practice and whether customers adopt them. Walmart’s gateway and spending controls will show how a major retailer defines approved access. Meta’s retail connections, announced with companies including Best Buy, Gap, Sephora, Walmart and Wayfair, will offer another indication of which merchants are willing to work with Muse.
Retailers and agent developers are also likely to keep negotiating over identity checks, customer consent, payment handling and responsibility for problems. The protocols named in the source may develop, but their current status varies, and a draft or partner list does not establish widespread use. Until access and checkout rules become clearer, shoppers should check an agent’s permissions and spending limits and verify the merchant’s purchase, delivery and return terms.
Key Questions
What did Walmart announce for AI shopping agents?
Walmart announced Agent Profiles and an Agent Gateway on October 21. The tools are intended to give approved agents a structured way to access shopping functions, with profiles that can apply permissions and spending limits.
Why is Amazon blocking Meta’s Muse?
Amazon said Muse did not identify itself as an agent and appeared to capture and store customer credentials, and said access violated its Conditions of Use. Meta said Muse could not see users’ passwords or payment methods. The accounts conflict, and the source does not independently resolve them.
How many people have tried Muse?
Meta said more than 10 million people in the United States and Canada had set up Muse within about six weeks of its launch. The figure is a company-reported setup count, not an independently verified number of completed purchases.
Can AI agents complete purchases everywhere?
No. Access depends on each retailer’s rules and technical setup. The source reports that Amazon blocked Muse, while Walmart announced tools for approved agents. It does not establish that agents can reliably complete checkout across retailers.
What should shoppers check before using an agent?
Review what the agent is allowed to buy, set a spending limit or require approval, and check the payment and return terms that apply to the order. The source does not identify a shared consumer-protection standard across shopping agents.
Source: ThorstenMeyerAI.com
