Shopify lets Meta's Muse agent in via Shop Pay – Amazon denies access
Retailers' bot defenses and commercial reluctance are becoming the practical bottleneck for consumer AI agents: Amazon has shut the door on Muse, Delta says no to third-party agents, Walmart claims its blocks were unintentional — while the industry now works on an open standard for agent-to-agent communication in commerce.
Amazon has begun blocking Meta's AI agent Muse from its store, according to TechCrunch in an article published October 6, 2026. That means the agent can no longer browse or make purchases from Amazon's product catalog. At the same time, the case reveals an industry pattern: some sites are deliberately closing the door to agents, others are doing so by accident, and several of the industry's major players have already begun building a standard for who gets let in.
What has happened
According to TechCrunch, Amazon started the blocking recently — the article says only "most recently," so the exact date of the decision is not known. After the block, Muse can neither explore Amazon's selection nor complete purchases on a user's behalf.
Amazon's reasoning is known through Business Insider, republished via Yahoo Finance on October 5, 2026: the company reportedly raised concerns about how Muse obtained access to and stored customer credentials and sensitive data, as well as scraping of account data. This is the company's claim as relayed by secondary sources — not verified with Amazon directly.
The dispute over credentials and data
The core conflict concerns what an agent actually does when it shops for you. From the website's perspective, agent traffic looks like machine traffic against a user account — the same category as the bots that anti-bot systems are built to stop. From the agent's and user's perspective, the agent is acting on behalf of a real customer, using the customer's credentials. Amazon's concern, as reported, points precisely at this intersection: who stores the login data, how sensitive data is handled, and whether the agent collects more account information than necessary.
TechCrunch notes that the problem does not concern Amazon and Muse alone. User complaints on social media suggest that AI agents frequently run into similar blocks — some intentional, like Amazon's, others a side effect of traditional anti-bot measures designed to protect against spam and malicious activity.
Two opposing strategies
Where Amazon shuts the door, Shopify has opened it. The payments platform has entered a partnership with Meta that allows Muse users to check out with Shopify merchants through Shop Pay, according to Business Insider.
The effect may already be noticeable to individual users. Leif Abraham, co-founder and co-CEO of Public, says that several of his orders are now processed via Shopify instead of Amazon, precisely because Amazon blocks Muse purchases. This is one person's anecdote from a single source — Business Insider itself notes that it does not prove Amazon has a sales problem. But it illustrates the mechanism: when a large retailer blocks agents, agent-driven commerce is steered toward retailers that allow it.
It is this tension that makes the story bigger than a single block. If consumer agents gain traction, the partnership choices of platforms like Amazon and Shopify will determine where agent-driven purchases are completed.
Travel: no, and a "we're evaluating"
The airlines differ from retail. Delta has, according to TechCrunch, no partnership or integration that allows a third-party AI agent to search for or book Delta flights on a customer's behalf on the company's digital platforms. "We continue to evaluate how new technology, including external AI agents, could enhance the way customers interact with Delta," Heena Chavda, Delta's general manager for Global Communications, told TechCrunch.
United points to its terms of use, which state that the user agrees not to use "any robot, spider, other automatic device or manual process to monitor or copy this website or the content contained therein, or for any other unauthorized purpose, without United's prior written permission." The company did not, however, respond to a follow-up question about whether it actually blocks specific personal AI agents like Muse. It is therefore unclear whether United actively blocks agents — what is established is that the company offers no integration and cites terms that in principle prohibit automated tools.
The gray zone: Walmart
The Walmart case shows how messy the landscape is. The company is officially a Muse partner — the partnership was announced at Meta's Connect developer conference in September — yet users have still experienced agent purchases being blocked. A Walmart spokesperson told TechCrunch that the blocks were not intentional.
That means even a declared agent-friendly partner can see its own anti-bot systems stop legitimate agents. For the industry, this may be the most telling example: the obstacle is not only commercial reluctance but also technical infrastructure that does not yet distinguish "good" agent bots from malicious ones.
The industry's answer: an open standard
To solve the problem, a number of industry players — Meta, Walmart, Stripe, Sierra, Genesys, Rocket, NiCE and Decagon — have begun working on an open standard intended to determine how AI agents can communicate with businesses. The goal, as Meta describes it in a blog post relayed by TechCrunch, is to help distinguish good bots acting on a user's behalf from malicious ones. The protocol will specifically focus on agent-to-agent communication in online commerce.
The details remain thin: neither the name, specification, nor timeline for the standard has been made public beyond the Meta blog post's description. But the participant list is significant — it includes commerce platforms (Walmart), payments players (Stripe), agent and customer service infrastructure (Sierra, Decagon, Genesys, NiCE) and finance players (Rocket).
What remains unclear
Several key details are missing. Neither Amazon's nor Meta's own primary sources — statements or blog posts about the block or the standard — are available; both are known only through TechCrunch's and Business Insider's reporting. Amazon's rationale is the company's claim, not verified. The scope of the blocks is unknown: the reporting rests on user complaints on social media and a single CEO anecdote. And it is unclear exactly when Amazon made the decision, or whether United actively blocks agents.
What is nonetheless clear is that the practical bottleneck for consumer AI agents is no longer model capability, but access: a map of who lets them in — Shopify and Walmart on the partnership side — and who does not, like Amazon and Delta, with an industry standard taking shape that may determine what the map looks like going forward.