Walmart leans into AI agents as Amazon shuts them out

The new Agent Gateway utilizes the Model Context Protocol, an open standard created by Anthropic that enables AI to interface with external databases. By adopting this framework, Walmart is effectively creating a secure pipeline for third-party bots to navigate its digital shelves. Unlike many competitors that view autonomous agents as a security threat or a disruption to their traffic metrics, Walmart is treating them as a new customer segment.

To manage this interaction, the retailer is introducing Agent Profiles next month. These profiles function like authorized user accounts, allowing shoppers to set specific spending caps and transaction requirements for each bot. This granular control ensures that users do not need to share sensitive personal information or full purchase histories with the AI tools. Furthermore, Walmart has established clear terms of service for these digital entities, explicitly stating that standard return policies apply even when an agent makes an error. By formalizing these operational guardrails, the company aims to mitigate consumer anxiety regarding autonomous shopping while maintaining a clear distinction between human and machine-initiated transactions.