Stripe, Meta, Walmart, and Shopify are building the infrastructure for AI agents to shop for you

A coalition of major technology and retail companies is quietly rewriting the rules of online commerce, moving from a world where you click “buy” to one where your AI agent does it while you’re doing something else entirely.

Meta, Sierra, Genesys, Rocket, Shopify, and Walmart have announced work on the Personal Agent Protocol, an open standard designed to let personal AI agents interact directly with businesses on behalf of their users.

The numbers behind the shift

The commercial stakes are not abstract. Stripe’s Link digital wallet, which counts approximately 300 million users and connects to around 1 million merchants, has seen agentic purchases surge 38-fold over the past month leading into late September 2026.

Meta’s AI assistant, Muse, launched on September 8, 2026, as the first personal agent to integrate with Stripe Link, giving it the ability to complete purchases through one-time virtual cards or stored payment credentials. Within two weeks of launch, Muse added Shop Pay and PayPal integrations, and it became the first agent to offer purchase protections to users.

Walmart is positioned as a major connector within the Muse ecosystem alongside other large retailers.

A fragmented standards race

Walmart and Shopify are also participants in the Universal Commerce Protocol, co-developed with Google, Target, Etsy, and Wayfair, which is designed to cover the full arc of an agent-initiated checkout process. Stripe, for its part, co-developed the Agentic Commerce Protocol with OpenAI, released under Apache 2.0 in September 2025, to streamline checkout flows for AI-driven transactions. Stripe also launched the Machine Payment Protocol with Tempo in March 2026 to address how machines, rather than humans, handle payment initiation.

Genesys expanded its own collaborations with Meta and others in September 2026, focusing on agentic customer experience orchestration using protocols including A2A and MCP.

What this adds up to is a landscape where several overlapping and competing protocols are vying to become the standard layer through which AI agents transact, with no single framework yet dominant.

Why this matters beyond the tech

The shift from chat-based AI assistants to transaction-capable agents is a structural change in how consumers engage with retail. A chatbot that recommends a product is useful; an agent that buys it, tracks shipping, and handles returns on your behalf is a different category of tool entirely.

Muse’s decision to include purchase protections from the start signals that the companies building these agents understand that trust is the real bottleneck to mass adoption, not the technology itself.