Enterprise

Walmart opens an Agent Gateway for AI shoppers, forcing small brands to make their catalogs legible to bots

A new MCP server and forthcoming Agent Profiles let ChatGPT, Gemini, and Meta's Muse buy on customers' behalf — turning product data quality into the new storefront for Marketplace sellers.

By Asher Glick· Enterprise software desk·2026-10-26·3 min read

Photo: Unsplash / rupixen — A laptop and smartphone on a desk being used for online shopping

On October 21, at the Masters of Scale Summit in San Francisco, Walmart CEO John Furner unveiled an Agent Gateway that lets third-party AI assistants, ChatGPT, Gemini, Meta’s Muse, search inventory and complete purchases on behalf of shoppers at Walmart.com and Sam’s Club. The gateway is built on the model context protocol, the Anthropic-originated standard that has quietly become the plumbing of agentic commerce. Agent Profiles, with per-agent spending caps around $50, $100, or $200 and transaction-level approval controls, roll out next month. No Walmart+ membership required.

The contrast with Amazon, which blocked Meta’s Muse in September for what it described as unauthorized account access, data scraping, and transaction processing, is the whole story. One incumbent is treating agents as a hostile scraping surface. The other is handing them a documented API and a billing relationship.

For small brands on Walmart Marketplace, that posture change is the thing to internalize. Agents don’t browse. They parse. A ChatGPT instance evaluating three equivalent SKUs reads structured attributes, current price, inventory status, and description fidelity, then picks. The merchandising polish that wins human eyeballs is invisible. The catalog metadata most sellers treat as back-office hygiene is now the storefront.

Walmart has pre-empted the obvious objection by stating that “the prices that an AI agent sees are the same prices that a human would see on Walmart.com, with applicable offers and fees shown before purchase.” Lindsay Owens of Groundwork Collaborative, speaking on Bloomberg’s Odd Lots, warned that agents won’t always act in the consumer’s best interest. Both things can be true, and small sellers should assume agents will optimize ruthlessly on whatever signals the catalog actually exposes.

This isn’t a one-off listing refresh. It’s the same shift Stripe flagged when agentic commerce went vertical in September, and it rhymes with the discovery-collapse dynamic playing out as TikTok is pulling its feed into a full-funnel sales machine. Attributes, prices, and stock counts have to stay accurate because an algorithm you’ll never see is comparing them against a competitor’s, silently, every time a customer asks their assistant to buy something.

The strange part isn’t that Walmart did this. It’s that most sellers still don’t know their listings are being auditioned.

Sources