Retail giant Walmart is diving headfirst into the era of agentic commerce, announcing a sweeping strategic move to accommodate artificial intelligence agents on its digital platforms. Speaking on Wednesday at the Masters of Scale Summit in San Francisco, Walmart President and CEO John Furner revealed that the company is launching its own proprietary Agent Gateway alongside a suite of new features specifically designed for consumers utilizing AI assistants. The initiative spans both the flagship Walmart ecosystem and its membership warehouse division, Sam’s Club, positioning the retail titan as a trailblazer in an industry where digital storefronts are increasingly wrestling with the rise of automated web traffic.
This forward-leaning strategy sharply contrasts with the defensive postures adopted by many other online businesses. While consumer adoption of autonomous AI agents has surged, various travel portals, booking sites, and e-commerce competitors have responded by locking these agents out of their domains entirely. Notably, this announcement comes on the heels of a high-profile move by Walmart’s chief rival, Amazon, which recently began blocking AI agents from making purchases on behalf of online shoppers. Rather than shutting its doors to the burgeoning technology, Walmart is leaning into it, building upon earlier announcements such as its partnership with Meta’s Muse AI agent, alongside existing integrations with popular generative AI applications including ChatGPT and Gemini.
At the heart of Walmart’s new ecosystem is the Agent Gateway, a specialized Model Context Protocol (MCP) server developed to facilitate secure, streamlined connections between approved AI agents and the retailer’s digital infrastructure. The Model Context Protocol, originally pioneered by Anthropic, acts as a standardized bridge allowing AI agents to interface seamlessly with external data sources, specialized software tools, and complex operational workflows. Through Walmart’s implementation of this protocol, an authorized AI agent gains the capability to comb through the retailer’s vast online inventory and securely execute transactions on behalf of the human user.
Crucially, Walmart has engineered the Agent Gateway with data privacy in mind. The retailer emphasizes that the system enables customers to shop for their regular brands and routine household items without forcing them to hand over their sensitive personal information or complete purchase histories directly to the third-party AI agents. By keeping the transactional bridge secure and compartmentalized, Walmart aims to alleviate consumer anxieties regarding data exposure while still delivering the convenience of automated shopping.
To further manage the complexities of automated purchasing, Walmart has introduced technological frameworks designed to clearly distinguish between shopping actions initiated directly by a human customer and those executed autonomously by an AI agent. Central to this oversight mechanism are Agent Profiles, a consumer control feature slated to roll out next month. When a shopper uses a specific AI agent to make a purchase through the Walmart Agent Gateway for the first time, a dedicated Agent Profile is automatically established within their main account.
This profile serves as a centralized management dashboard, allowing customers to view all agent-driven purchases separately from their personal shopping history. Furthermore, the system puts complete budgetary and operational authority back into the hands of the consumer. Shoppers can configure their preferences to require explicit human approval for every individual transaction attempted by an agent, or they can establish preset spending limits that give the AI autonomous clearance up to a specific dollar threshold.
Walmart likens these newly introduced controls to the familiar financial practice of adding an authorized user to a traditional credit card account. Just as a primary cardholder can track secondary purchases and cap spending limits for another person, Walmart customers can monitor and restrict the financial activity of their digital assistants. The creation of these profiles requires nothing more than linking the agent to a standard Walmart account during initial setup, and shoppers retain the freedom to delete any Agent Profile at their discretion. Notably, this agent support infrastructure is being made available to any registered Walmart account holder, meaning a paid Walmart+ membership is not required to access the service.
Alongside these technological tools, Walmart is rolling out a specialized set of operating terms specifically tailored for AI agents. These guidelines explicitly dictate how autonomous systems must behave and outline their precise operational rights and limitations when interacting with the retailer’s digital storefronts. Among the most consumer-friendly provisions within these terms is the application of Walmart’s standard returns policy to agent-driven transactions. This ensures that even if an AI agent malfunctions, misunderstands a prompt, or executes an unwanted purchase, the customer remains fully protected and will not find themselves stuck with unwanted goods due to software errors. This policy is expected to play a critical role in reassuring hesitant shoppers who might otherwise view themselves as reluctant early testers of agentic commerce.
The new terms also enforce strict operational boundaries to protect consumer security. Walmart has stipulated that AI agents are strictly prohibited from taking actions that exceed the explicit intent of the shopper. Additionally, agents are barred from accessing or storing the customer’s underlying payment information and are restricted to retrieving only the specific data points required to complete a designated task. To ensure compliance, Walmart has stated that any AI agent that violates these mandated terms of service will be promptly blocked from accessing the retailer’s website and digital infrastructure.
