🔍 Read the full analysis: Could AI Shopping Agents Shift 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
Meta says more than 10 million people in the United States and Canada set up its Muse shopping agent within about six weeks, making agent-led shopping a visible commercial issue. Walmart is creating a route for outside agents, while Amazon has blocked Muse; who controls checkout, customer data and purchase referrals remains unsettled.
Meta says more than 10 million people in the United States and Canada set up its Muse personal agent within about six weeks of its September launch, as Walmart prepares a dedicated gateway for shopping agents and Amazon blocks Muse. The competing approaches put a commercial question at the center of agent-led shopping: who controls the customer relationship when software searches and buys on a shopper’s behalf?
Meta shared its Muse figures on its October 21 earnings call, saying shopping was the most common use and the average purchase was over $200. Those are company-reported figures; the source material does not provide independent verification or clarify how Meta calculated an average purchase. Muse can use a browser to search products and complete forms, and it pays through Link by Stripe, which the source says issues a one-time-use card.
Walmart announced Agent Profiles and an Agent Gateway on October 21. The gateway is designed to let approved agents search inventory and buy through a structured connection rather than imitate a person clicking through a website. Agent profiles are intended to give users controls such as spending limits. Walmart’s announcement marks a willingness to admit outside agents through a defined route, though the available material does not establish how widely the system is operating.
Amazon has taken the opposite approach with Muse. On September 20, Muse users reportedly encountered a notice saying unauthorized agent access violated Amazon’s Conditions of Use. Amazon said Muse had not disclosed that it would shop on the site, did not identify itself as an agent, and appeared to capture and store customer credentials. Meta’s launch post said Muse had no visibility into users’ passwords or payment methods. These are conflicting company accounts, and the source material does not resolve the technical dispute.
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 stakes extend beyond whether an agent can find a low price. If an agent mediates product discovery and checkout, it may shape which retailers a shopper sees, where a transaction happens and which company receives the associated data. Retailers want control over access, payment and customer service; AI firms want their agents to be useful across merchants rather than limited to a single store.
That tension also affects advertising and referrals. The source reports that Amazon earned more than $68 billion from advertising last year, citing GeekWire, while Meta has said it expects to profit by taking a small fee from transactions. If a shopper asks an agent to compare products and then buys without clicking a tracked affiliate link, the publisher or referral site that helped inform the decision may not receive credit under click-based arrangements. The scale of that effect is not established by the material.
For shoppers, a more automated purchase can reduce effort, but it also changes who handles the transaction. Users may need to know whether the retailer or agent is responsible for order changes, returns and disputes, and what limits apply when an agent can act without asking about every step.
Products Worth Considering
Sonos Era 100 SL Compact Microphone-Free Speaker, Black
Logitech MX Keys S Wireless Keyboard, Graphite
Amazon Kindle Paperwhite (12th Generation, 2024), 16 GB, 7-inch Display, Jade
As an affiliate, we earn on qualifying purchases.
From Browser Bots to Retail Gateways
The dispute follows a fast series of product and protocol announcements. Meta launched Muse on September 8, describing it as an agent able to open a browser, fill out forms and negotiate for a user. After Amazon restricted access, Meta announced retail connectors with Best Buy, Gap, Sephora, Walmart, Wayfair and Expedia at its September 23 Connect event. Shopify also introduced structured checkout tools for browser agents in September.
There are three broad ways an agent can shop: by using a browser like a person, by connecting to a retailer’s own agent interface, or through shared protocols intended to carry information about products, permissions and checkout. The browser approach can reach sites built for human customers, but retailers may have difficulty distinguishing an agent from the account holder. A retailer-controlled interface offers more structure, while shared protocols are still being developed and adopted.
Several protocol efforts are in play. The Personal Agent Protocol, associated with Meta, Sierra, Walmart, Shopify and Stripe, was published as a draft in October. Google and Shopify are among the backers of the Universal Commerce Protocol, while OpenAI and Stripe developed the Agentic Commerce Protocol. The source also notes 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 OpenAI focused on product discovery. These developments show that a common technical route has not settled the commercial questions.
“Muse has no visibility into people’s passwords or payment methods.”
— Meta, in its Muse launch post
Rules, Reach and Responsibility Remain Open
Meta’s adoption and purchase figures are company-reported, and the source does not specify the counting method behind the reported average purchase. It is also unclear how often Muse users complete purchases, how many retailers support agent access in practice, or whether Walmart’s gateway will attract broad adoption.
The central disagreement over Muse’s access to Amazon remains unresolved in the supplied reporting. The source includes Amazon’s allegations and Meta’s denial, but not independent technical findings. More broadly, the market has no settled answer on how agents should identify themselves, what permission a user has granted, or who bears responsibility for an incorrect order or an account security problem.
One cited measure suggests agents are currently used more for research than checkout: HUMAN Security’s September data says 81% of agent requests went to product and search pages, compared with 2.4% to checkout and payment. The source does not state the full measurement scope or provide a comparison baseline, so the figures should be treated as a snapshot rather than proof of a long-term trend.
Adoption Will Test the New Gateways
The next test is whether retailers, payment providers and agent makers can agree on workable access and checkout arrangements. Walmart’s gateway and profile controls will need to move from announcement to practical use, while protocol developers will have to show that their systems can carry clear user permissions and work across merchants.
Shoppers should check what an agent is authorized to do before enabling purchases. Spending limits, approval requirements and payment controls can reduce the risk of an unwanted order, while checking the retailer’s return and dispute policies can clarify which company handles problems. The source material does not identify a single standard or timetable that would settle those issues; adoption and the terms set by individual retailers remain in flux.
Key Questions
What is Meta’s Muse shopping agent?
Muse is a personal AI agent that Meta says can use a browser, fill out forms and act for a user. The source says it runs in a virtual machine in Meta’s cloud and uses Link by Stripe for payments.
Why did Amazon block Muse?
Amazon said Muse was an unauthorized agent, had not disclosed that it would shop on Amazon, and appeared to capture and store customer credentials. Meta has disputed the credential concern, saying Muse cannot see users’ passwords or payment methods. The supplied material does not independently settle the disagreement.
What is Walmart’s Agent Gateway?
Walmart describes it as a structured way for approved agents to search its inventory and make purchases, rather than navigating its website like a human shopper. Walmart also announced agent profiles designed to support controls such as spending limits.
Are AI agents already completing most online purchases?
The source does not show that they are. HUMAN Security’s September figures say 81% of agent requests went to product and search pages and 2.4% to checkout and payment. Those figures describe requests, not a verified share of all online purchases.
What should shoppers check before letting an agent buy?
Check the agent’s purchase approval settings, spending cap and payment method, and find out which retailer’s return and dispute policies apply. The degree of control available may differ across agents and merchants.
Source: ThorstenMeyerAI.com
