TL;DR: Amazon is still actively blocking AI shopping agents from browsing its site and completing purchases, while Walmart has gone the opposite direction, integrating directly with ChatGPT so customers can buy products without ever opening Walmart’s own app.
Two of the largest retailers in the country are running opposite experiments on the same question: when a customer wants an AI assistant to do their shopping for them, does the retailer want to be found by that assistant, or does it want to be the one controlling the checkout no matter who’s asking? Amazon has picked control. Walmart has picked reach.
Amazon’s wall is getting taller, not shorter
Reporting this week laid out just how far Amazon has gone in restricting automated shopping agents, including those built on technology from competitors like Google and OpenAI, from crawling product pages, comparing prices, or completing checkouts on Amazon’s behalf. This isn’t new behavior exactly, Amazon has been quietly tightening crawler access for its own competitive reasons for a while, but the contrast with what Walmart is doing at the same moment makes the strategic divergence much starker than it would otherwise read.
Amazon’s logic isn’t hard to follow. The company has spent two decades building a storefront, a recommendation engine, a review ecosystem, and a loyalty program all designed to keep the customer relationship inside Amazon’s own app and site. An AI agent that completes a purchase on a customer’s behalf, using data scraped from Amazon’s listings but routing the actual transaction and the resulting customer relationship through someone else’s assistant, threatens basically every one of those investments at once. Blocking the crawler is the obvious defensive move if you believe the agent layer is where future retail margin gets captured.
Walmart is betting the opposite is true
Walmart’s calculation looks different, and it comes from a position of being the long-running number two in US e-commerce behind Amazon. Rather than defend a dominant storefront, Walmart has leaned into being wherever the customer already is, which this year increasingly means inside ChatGPT. The company’s integration lets a ChatGPT user complete a Walmart purchase directly in the conversation, without a separate app download, a separate login flow, or a separate checkout page breaking the experience.
Early data circulating from retail analysts suggests ChatGPT-driven traffic already accounts for a meaningful share of Walmart’s referral clicks, a number that would have sounded absurd eighteen months ago and now reads as an obvious consequence of where consumer attention has been moving. If that keeps compounding, Walmart’s bet is that being the easy, frictionless answer inside the assistant someone’s already talking to beats trying to pull that same person back into a dedicated shopping app they have to remember to open.
| Amazon | Walmart | |
|---|---|---|
| AI agent access | Actively blocked for competing assistants | Direct ChatGPT integration, in-chat checkout |
| Strategic logic | Protect existing storefront and customer relationship | Meet customers wherever they already are |
| Risk | Gets bypassed anyway as agents route around the wall | Cedes brand relationship to the assistant layer |
DeepSeek made its own play in this same general space this week, launching search and page-fetch APIs specifically built for AI agents, which we’re covering in a separate piece today, a reminder that the infrastructure underneath agentic shopping is being built out fast on multiple fronts at once, not just by the retailers themselves. It’s a similar dynamic to what we saw when Meta’s own AI agent app started climbing the App Store charts last month: the agent layer is getting crowded fast, and not every company sees that as good news.
Neither bet is obviously right yet
The case against Amazon’s approach is that walls built against today’s agent technology tend to get routed around by tomorrow’s, and that actively alienating the fastest-growing new channel for product discovery risks losing relevance with exactly the customers who’ll matter most in five years. The case against Walmart’s approach is that handing checkout control to someone else’s chat interface means Walmart becomes a fulfillment backend for OpenAI’s relationship with the customer, with less pricing power and less brand visibility than it has today as the front door itself.
Both companies are making a real bet with real money and real engineering effort behind it, not just issuing a press release and waiting to see what happens. That’s usually a sign the people making the decision think the stakes are high enough to justify picking a side now rather than waiting for more data.
Smaller retailers don’t get to sit this one out
The retailers getting attention right now are the two giants with the resources to build custom integrations or enforce sophisticated crawler blocking at scale. Smaller and mid-size e-commerce operators don’t have that luxury, and most haven’t articulated a strategy at all yet, largely because the volume of AI-agent traffic hitting a mid-size store is still small enough to ignore for now. That won’t last. Whatever strategic split wins out between Amazon’s wall and Walmart’s open door, the pattern will filter down to every retailer with an online storefront within a couple of years, and the ones figuring out their position now, rather than reacting once agent traffic becomes impossible to ignore, will have a real head start on everyone still debating whether this is a fad.
The practical question for a smaller store owner today isn’t philosophical. It’s whether your product pages, return policies, and shipping information are even structured in a way an AI agent can parse cleanly if a customer asks one to compare you against a competitor. A lot of small business sites, built years ago and never revisited, would fail that test today regardless of which side of the Amazon-Walmart divide they’d instinctively prefer to be on.
The bottom line
Amazon is defending the house it built. Walmart is betting the house doesn’t matter as much as being wherever the customer is standing. Whichever retailer is proven right probably won’t be obvious for another year or two, but the fact that both are moving decisively, in opposite directions, tells you the AI shopping agent question has stopped being theoretical.
The Teck Hustlers Editorial Team covers technology, business, finance and gaming news for readers who want the story behind the headline. Every article is researched, written and edited by our in-house team, checked against company announcements, filings and reputable reporting, and updated when new facts emerge. We focus on what each development means for founders, professionals, investors and everyday users. Have a tip or a correction? Reach us through our contact page.
