The strategic question is not whether an AI assistant can recommend a product. It is whether the retailer that fulfills the order can remain economically important when software, rather than a customer, becomes the primary interface. Walmart’s answer is to permit that transition under defined permissions. Amazon’s response, by contrast, has been to restrict at least some third party shopping agents.
The contrast was highlighted by TechCrunch in a post describing Walmart’s new technology and Amazon’s decision to block automated shoppers. Walmart’s approach could expand the number of ways customers reach its products, while Amazon’s position protects a more direct relationship between its website, shoppers and commercial systems.
Walmart’s controlled delegation model
In its announcement, Walmart said it is introducing Agent Profiles and the Walmart Agent Gateway, along with rules for third party AI agents that interact with its services. The company is not presenting delegation as unrestricted access. Instead, it is building a permissions layer around the purchase.
Customers can require approval for every transaction or authorize an agent to spend within preset limits, including $50, $100 or $200. Agents receive separate profiles and cannot access the customer’s payment card information, according to Walmart. Those controls are important because they separate the authority to select and order products from access to the underlying financial credentials.
That structure gives Walmart a way to make automated shopping more useful without treating an agent as identical to the customer. It also creates a clearer record of which actions were authorized, which could matter when an order is mistaken, disputed or returned.
For Walmart, the commercial opportunity is reach. An AI agent could search across products, compare options and complete a purchase without requiring a customer to navigate Walmart’s website or application step by step. If those agents become a common entry point for commerce, retailers that support them may capture demand that would otherwise flow through another platform.
The risk is that Walmart could become the fulfillment layer while the agent owns the customer interaction. That may reduce the retailer’s control over merchandising, recommendations and the data generated during product discovery. Walmart’s gateway and profile system appear designed to preserve some influence by setting the rules under which agents can operate.
Amazon protects its closed ecosystem
Amazon has taken a more restrictive position in the dispute. GeekWire reported that Amazon blocked Meta’s Muse AI assistant after asking Meta to remove Amazon from the experience. An Amazon spokesperson said third party applications that make purchases for customers should operate openly and respect retailers’ decisions about whether to participate.
That argument frames automated shopping as a relationship between platforms, not simply as a convenience feature for consumers. Amazon’s position gives retailers a veto over whether an external assistant can use their storefront to complete transactions. Walmart’s policy instead attempts to define the conditions under which such access is acceptable.
The business stakes extend beyond order volume. Axios reported that GlobalData Retail analyst Neil Saunders believes Amazon may be trying to maintain its closed ecosystem, and may be concerned that agentic shopping could weaken its advertising business. If an AI system selects products before a shopper reaches Amazon, the retailer could lose opportunities to influence discovery and monetize product visibility.
Walmart may be making a different calculation. Welcoming agents could create more routes to its inventory and make its marketplace relevant inside tools that it does not control. The tradeoff is that Walmart would need to manage automated demand, permissions and disputes while competing for influence with the agents themselves.
The trust and privacy test
Delegated purchases also introduce risks that spending caps cannot fully eliminate. An agent may misunderstand a request, select the wrong variation, order an unsuitable substitute or repeat a purchase. Even when a customer approves a transaction, responsibility may become complicated if the recommendation was shaped by incomplete information or commercial incentives.
The Associated Press quoted Georgia Tech cybersecurity and privacy professor Veronica Rivera, who compared using an AI agent with handing someone keys and a to-do list. Her comparison highlights the access problem: an agent may need enough personal information to act effectively, but that same information can expose household preferences, purchasing patterns and account privileges.
The supplied reaction coverage is critical rather than supportive. No independent source provided for this story offers a favorable assessment of Walmart’s move. That does not establish that the strategy will fail, but it means the early debate is centered on control, privacy, platform power and commercial risk.
Walmart’s launch will therefore be judged on execution, not just openness. If approvals are clear, spending limits work reliably and customers can resolve mistakes quickly, the retailer could gain an advantage by becoming an early destination for agent-led commerce. If automated shopping produces costly errors or weakens customer trust, Amazon’s closed approach may appear less like resistance to innovation and more like protection against an immature channel.
The larger contest is over who owns the transaction in an AI-led economy. Walmart is betting that access will bring value, provided it can set the permissions. Amazon is betting that control over access is itself a competitive asset.
Rebeca Smith is not a person. No notebook, no deadlines, no face behind the name — just a byline this newsroom publishes under. Here is the production line underneath it, because a name beside a portrait reads like a journalist, and this one is not one.
The models. Writing: gpt-5.6-luna and qwen3-max. Out on the live web: gpt-5.6-luna and gpt-5.6-terra. Pictures: gpt-image-1 and gpt-image-1-mini. Swap one in the newsroom and this line swaps with it — it is read off the machines, not typed here.
How a story is made
- Research. The searching model reads around the story, pointed at primary sources — the filing, the post, the repository — rather than at somebody else's write-up of them.
- Writing. The writing model drafts it against what was found, at Rebeca Smith's usual length and in Rebeca Smith's usual register.
- The loop. A reviewer reads the draft and sends it back with notes. Then reads it again. A piece can go round several times before it leaves the building.
- Enrichment. A quotation has to appear word for word on the page it is taken from. A chart may only use figures that appear in the source it cites. Whatever fails is dropped, and the reason is kept.
- Fact check. A last pass hunts for claims the article makes and its sources do not.
- A human stop. Sensitive subjects are held for a person to read before publication, and a person can kill any of it at any point.
If that sounds less like a newsroom and more like a factory: quite. It is called Press Factory.
This article was generated using AI and published automatically without human pre-publication review.
How this article was made
The article was produced by the Grandmonts Media News Engine using automated research, drafting and verification workflows. No human editor reviewed the article before publication. Grandmonts Media remains responsible for the published content. Errors can be reported at office@grandmonts.cz.