🔍 Read the full analysis: Amazon Vs. Walmart: Two Approaches To AI Shopping Agents on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
Walmart has announced tools that let approved AI agents search and buy from its store, with customer profiles and spending limits. Amazon has blocked Meta’s Muse agent and says outside agents must identify themselves under its terms. The approaches reflect a wider dispute over customer control, payment, and how shopping agents access retailers.
Walmart is opening a route for approved AI shopping agents to search its products and make purchases, while Amazon has blocked Meta’s Muse and says outside agents must identify themselves under its Conditions of Use. The contrasting moves sharpen a developing fight over who controls the shopping relationship when software, rather than a person, researches and buys products.
Walmart announced its Agent Gateway and Agent Profiles on October 21. The gateway uses the Model Context Protocol, or MCP, to let approved agents access shopping functions, including inventory search and purchasing. Agent Profiles are intended to let customers set permissions and spending limits. The announcement describes a structured connection between agents and Walmart rather than an agent operating the store’s ordinary website as if it were a person.
Amazon has taken a more restrictive approach to Meta’s Muse. After blocking the agent in September, Amazon said Muse had not identified itself as an agent, that Meta had not notified Amazon it would shop there, and that the software appeared to capture and store customer credentials. Meta disputed the credentials concern: its launch post said Muse had no visibility into users’ passwords or payment methods. The accounts describe opposing positions, not an independently established finding about how the system handled credentials.
Meta’s figures show the potential scale, though they are company-reported. At its October 21 earnings call, Meta said more than 10 million people in the United States and Canada had set up Muse in about six weeks, shopping was its most common use, and the average purchase was over $200. Those numbers suggest meaningful early interest, but do not establish how often users complete purchases or how the agent affects retail sales.
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 Contest Agent Access
The divide matters because an agent can become the intermediary between a shopper and a retailer. If the agent chooses where to search, what to compare, and where to pay, the retailer may lose control over parts of the customer relationship and the information around a purchase. Walmart’s gateway offers a way to permit access through defined tools and permissions; Amazon’s restrictions preserve a tighter boundary around its store and its own shopping services.
There are also questions about who receives credit and revenue. The source material points to Amazon’s advertising business, which GeekWire reported generated more than $68 billion last year, and to Meta chief executive Mark Zuckerberg’s statement that Meta expects to earn a small fee from transactions. Neither figure settles how agent-led shopping will affect retail economics. But the competing models could shift value among retailers, agent providers, payment companies, and publishers that currently earn referral commissions.
For shoppers, convenience comes with practical choices about spending authority, payment details, and returns. Walmart’s announced controls address some permission questions, while an agent using a browser may encounter different protections and policies at each store. A smooth demonstration does not by itself establish who is responsible when an order is wrong or a purchase exceeds what the user intended.
Products Worth Considering
Sonos Era 100 Wireless Smart Speaker, Black
Bose SoundLink Flex Bluetooth Speaker (2nd Gen), Portable Waterproof Speaker, Black
DJI Osmo Mobile 8 Phone Gimbal with Mic Mini Transmitter, Infinity Black
As an affiliate, we earn on qualifying purchases.
From Browser Bots to Store Gateways
The dispute developed quickly in September and October. Meta launched Muse on September 8, describing it as an agent able to open a browser, fill out forms, and negotiate on a user’s behalf. On September 20, Amazon cut off Muse. Three days later, Meta announced retail connectors involving Walmart and several other merchants. Walmart’s October 21 gateway announcement followed a series of efforts to create more structured ways for agents to interact with shops.
These systems use different routes. A browser-based agent can navigate pages made for people, but a retailer may find it difficult to distinguish the agent from the customer. A retailer’s own gateway instead exposes selected functions to approved software. Shared protocols aim to standardize how agents and merchants communicate, but they remain at different stages of development and do not guarantee that retailers will accept every agent.
Early usage data also puts the checkout claims in perspective. HUMAN Security reported in September that 81% of observed agent requests went to product and search pages, while 2.4% went to checkout and payment. That dataset indicates that agents were largely researching rather than completing purchases during the measured period; it does not show that all agents or retailers have the same pattern.
“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
Purchase Rules Still Vary
It is not yet clear how many purchases Muse or other agents complete, how representative Meta’s reported user figures are, or what share of agent-assisted transactions will take place through retailer gateways. Walmart’s announcement does not establish how quickly the gateway will be available to outside agents, which agents will qualify, or how the new controls will work in every transaction.
Responsibility for disputed orders, refunds, and returns also remains dependent on the retailer and payment setup. The source material describes Amazon’s and Meta’s conflicting statements about Muse’s access and credentials, but does not independently resolve them. Protocols such as Poppy, Google’s Universal Commerce Protocol, and OpenAI and Stripe’s Agentic Commerce Protocol differ in scope and maturity; their existence does not mean they are interchangeable or widely deployed.
Agent Access and Checkout Tests
The next test is whether retailers and agent makers turn announced tools into working shopping experiences with clear permissions and reliable checkout. Walmart’s gateway and profiles will show whether retailers can allow agents in while setting limits on what they can do. Amazon’s position will show whether it continues to require agents to identify themselves and channels customers toward its own services.
Shoppers and merchants will also be watching how the competing protocols develop and whether agents move beyond product research to completed transactions. Until adoption, safeguards, and return procedures become clearer, users should check what an agent is authorized to buy, set spending limits where available, and confirm which retailer’s policies apply.
Key Questions
What did Walmart announce?
Walmart announced an Agent Gateway using MCP for approved agents to access shopping functions, alongside Agent Profiles intended to let customers set permissions and spending limits.
Why did Amazon block Meta’s Muse?
Amazon said Muse had not identified itself as an AI agent, that Meta had not notified it about the shopping activity, and that Muse appeared to capture and store customer credentials. Meta said Muse could not see users’ passwords or payment methods. The conflicting claims have not been independently resolved in the source material.
Are AI agents already completing many online purchases?
The cited HUMAN Security data suggests agent requests were mostly for research: in its September data, 81% went to product and search pages and 2.4% to checkout and payment. Those figures describe that dataset and period, not every retailer or agent.
What should shoppers check before using an agent?
Check what purchases the agent is allowed to make, use a spending cap or approval step when available, and confirm the payment method and retailer’s return policy. Protections can differ between a retailer’s own agent gateway and a browser-based agent.
Source: ThorstenMeyerAI.com
