Mastercard and Visa are racing to define how payments work in a world where AI agents, not people, increasingly do the clicking. Both companies have rolled out protocols that let chatbots and shopping assistants browse, select, and pay for products on a user's behalf, turning what used to be a simple checkout button into a new battleground for transaction volume.
The pitch to merchants and banks is straightforward: as more shopping starts with a prompt instead of a search bar, whichever network can move that spending through its rails first stands to keep the fees that come with it. Mastercard has been pushing its agentic-commerce tools through partnerships with AI developers, while Visa has built out its own framework allowing agents to hold spending credentials and complete purchases without a human approving each step.
The land grab reflects genuine uncertainty about who will actually control the checkout moment once AI agents become a normal way to shop. Retailers worry about losing the direct relationship with customers if an assistant becomes the default interface, and both card networks are betting that owning the authentication and settlement layer underneath those agents keeps them essential regardless of which app or chatbot wins the interface war.
Security is the open question hanging over the whole effort. Handing a piece of software the ability to spend money autonomously raises fraud and liability issues that don't map cleanly onto existing card rules, and both companies are still working out how disputes get resolved when an AI agent, rather than a person, made the purchasing decision. How that gets settled will likely shape how quickly banks and merchants are willing to switch on agentic payments at scale.