AI Shopping Agents Meet A Divide: Amazon Blocks, Walmart Welcomes
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: AI Shopping Agents Meet A Divide: Amazon Blocks, Walmart Welcomes on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get hardware and tech essentials delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

Amazon has blocked Meta’s Muse shopping agent, while Walmart is creating an Agent Gateway and profiles for approved agents. The competing approaches put retailers’ control of customer relationships, agent access and purchase attribution at stake, though most reported agent activity still focuses on product research rather than checkout.

Amazon has blocked Meta’s Muse shopping agent from accessing its site, while Walmart is opening a structured route for approved agents to search products and make purchases. The split, reported in developments through October 2026, shows retailers taking different approaches to automated shopping as they compete to control customer access and transaction data.

A reporter for The Verge asked Muse where granola bars cost the least per piece. The agent recommended Sam’s Club, Walmart and Costco, then said it could not check Amazon because the retailer had blocked it. The example illustrates a practical limit for shopping agents: their answers depend on which retailers allow them to access product information.

Meta launched Muse on September 8, describing it as an agent that can browse, fill out forms and act for users. The source report says Muse runs in a virtual machine in Meta’s cloud and uses Link by Stripe for payments. Amazon blocked the agent on September 20. Amazon said Muse had not identified itself as an agent and appeared to capture and store customer credentials. Meta’s launch material said Muse could not see users’ passwords or payment methods.

Walmart announced Agent Profiles and an Agent Gateway on October 21. Its approach is to let approved agents interact with its shopping systems through structured access, including search and purchasing functions. The source report identifies the gateway as based on the Model Context Protocol, or MCP. The announcement followed Meta’s September plans for retail connectors, which included Walmart and several other retailers.

At a glance
reportWhen: Developments reported in September and…
The developmentAmazon blocked Meta’s Muse shopping agent, while Walmart announced systems for approved AI agents to access its shopping services.
Thorsten Meyer AI
How it works

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.

1

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)

2

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

3

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

ProtocolBackersCoversStatus
Personal Agent Protocol (Poppy)Sierra, Meta, Walmart, Shopify, Stripe + 35 design partnersAgent identity, customer permission, one session across site, API and company agentDraft 0.1 (9 Oct 2026)
Universal Commerce Protocol (UCP)Google, Shopify; site also lists Etsy, Wayfair, Target, Walmart, Amazon, MicrosoftDiscovery, cart, checkout, orders; payment mandates via AP2Published
Agentic Commerce Protocol (ACP)OpenAI, StripeCheckout between agent and merchantOpen source; ChatGPT's own checkout scaled back in March 2026

What a careful agent purchase looks like

  1. You give the task and a limit, for example a $50 cap.
  2. The agent identifies itself to the shop and gets only the access you approved.
  3. It prepares the cart and shows you an approval card.
  4. Payment goes through a token or one-time card, so the agent never sees your card.
  5. 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

The dispute is about more than whether a bot can load a product page. Retailers are deciding who can access inventory and checkout, how an agent identifies itself, and which company controls the customer relationship when software shops on a person’s behalf. A retailer that blocks external agents can direct customers toward its own tools; a retailer that supports approved agents may gain access to new sales routes, while setting conditions for how those routes work.

The choice also affects how purchases are attributed. The source report notes that Amazon Associates commissions depend on customers using tracked links. If an agent reads a publisher’s review but completes a purchase through a different route, the publisher may not receive a commission. The report does not establish how common that outcome is, but it highlights a potential change for affiliate businesses as shopping shifts from links and web pages to agent-mediated transactions.

For shoppers, delegation adds questions about spending authority, payment and returns. Meta says Muse uses one-time-use payment cards, and Walmart’s announced profiles include spending limits. Those safeguards do not by themselves answer which retailer’s return rules apply, what a user must approve, or how disputes are handled. The source report advises users to check purchase controls before relying on an agent; that is consumer guidance, not a guarantee of protection.

Products Worth Considering

Apple MacBook Pro 16.2-inch with M5 Pro, 18-Core CPU, 20-Core GPU, 48GB Memory, 1TB SSD, Nano-Texture Display, Space Black

Apple MacBook Pro 16.2-inch with M5 Pro, 18-Core CPU, 20-Core GPU, 48GB Memory, 1TB SSD, Nano-Texture Display, Space Black

Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo

Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo

Logitech MX Mechanical Full-Size Wireless Mechanical Keyboard, Tactile Quiet Switches, Backlit Keys, Bluetooth, USB-C, Multi-Device, Graphit

Logitech MX Mechanical Full-Size Wireless Mechanical Keyboard, Tactile Quiet Switches, Backlit Keys, Bluetooth, USB-C, Multi-Device, Graphit

As an affiliate, we earn on qualifying purchases.

A Fast-Moving Agent Rollout

The reported sequence began with Muse’s launch on September 8. Amazon blocked it on September 20, saying the agent violated its Conditions of Use. On September 23, Meta announced retail connectors, including Walmart, and CEO Mark Zuckerberg said the company expected to earn a small fee from transactions. Shopify later introduced structured tools for browser agents to complete checkout, rather than relying only on screen interactions.

On October 6 and 9, Sierra and Meta announced the Personal Agent Protocol, also called Poppy, with Walmart, Shopify and Stripe among its participants. A draft followed, along with additional design partners. The source material describes Poppy as an early draft focused on identifying agents and recording what customers allow. It also names other commerce protocols, including Google’s Universal Commerce Protocol and the Agentic Commerce Protocol from OpenAI and Stripe. These initiatives are at different stages and should not be treated as a single settled industry standard.

The source report cites HUMAN Security data from September showing 81% of agent requests went to product and search pages, while 2.4% went to checkout and payment. Those figures describe the measured requests in that dataset and period; they do not establish that the same proportions apply to all agents or retailers. They suggest that, at least in that measurement, agents were more often researching products than completing purchases.

“Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use, to which our customers have agreed.”

— Amazon, in the notice users saw after access was blocked

Access, Safeguards and Adoption

Several practical and technical questions remain open. The source material does not spell out the full approval process for Walmart’s Agent Gateway, which agents will be able to use it, or what limits will apply to each kind of transaction. It also does not provide a full account of how spending controls, payment disputes, returns and customer support will work across participating retailers.

The claims about credential handling are disputed: Amazon said Muse appeared to capture and store credentials, while Meta said the agent could not see passwords or payment details. The supplied material does not include an independent technical finding resolving that disagreement. It is also unclear how frequently consumers will use agents to complete purchases, rather than research products, or how retailers will measure sales generated by agents.

The protocols described are not interchangeable, and some remain drafts or have changed direction. The source report says OpenAI and Stripe’s Instant Checkout later gave merchants more control over their own checkout experiences after OpenAI said its initial version lacked the flexibility it wanted. Wider adoption, technical compatibility and the commercial terms between agents and retailers remain developing issues.

Retail Access Moves to Implementation

The next test is whether retailers and agent providers turn announcements and protocol drafts into working services used by shoppers. Walmart’s gateway and agent profiles will need to show how approved agents are identified, what users can authorize, and how purchases are completed. Meta’s retail connectors and the participating services will also indicate whether Muse can offer useful shopping access without relying on a retailer’s ordinary web pages.

Users should watch for clear controls over purchase approval, spending caps, payment methods and return policies before delegating transactions. Retailers, publishers and payment providers will be watching whether agent referrals can be attributed and whether the different protocols can work together. The source material does not give a timetable for broader rollout or establish which approach will prevail.

Key Questions

Why can’t Meta’s Muse check Amazon prices?

Amazon blocked Muse in September 2026, saying the agent was unauthorized and raising concerns about its handling of credentials. Meta said Muse could not see users’ passwords or payment methods. The supplied material does not include an independent finding resolving the competing claims.

How is Walmart’s approach different?

Walmart announced an Agent Gateway and Agent Profiles to provide approved agents with structured access to shopping services. The source report says the gateway uses MCP and that profiles include spending limits. Full access rules and rollout details were not provided.

Are shopping agents already completing many purchases?

In HUMAN Security’s September data cited by the source report, 81% of agent requests went to product and search pages and 2.4% to checkout and payment. Those figures describe that dataset and period, not all shopping agents or retailers.

What should shoppers check before using an agent?

Check what purchases require approval, whether there is a spending cap, which payment method is used, and which retailer’s return and dispute policies apply. Agent safeguards and terms can vary by service.

Source: ThorstenMeyerAI.com

Nothing in this article is financial or investment advice. Cryptocurrency and precious-metal investments carry significant risk — do your own research and consider a licensed advisor.
You May Also Like

Technology operations signal monitor: Show HN: Kage – Shadow any website to a single binary for offline viewing

Kage, a new tool allowing users to shadow websites into a single binary for offline viewing, is gaining interest among product and engineering leads at small software firms.

Explore In-Region Anthropic Model Inference With Amazon Bedrock

Anthropic says its models on Amazon Bedrock can be used for in-region inference in Seoul and Singapore; model coverage and data-handling details remain unspecified.

2026’S Finest 4K Webcams Enhanced By Artificial Intelligence

Discover the latest 2026 4K webcams with AI-driven features, offering improved image quality, autofocus, and noise reduction for streaming and professional use.

The Delegation Ladder: The Four Agentic Loops, And What Each One Lets You Stop Doing

A detailed analysis of the four agentic loops in AI engineering, explaining their functions, significance, and implications for automation and control.