Visa’s AI Shopping Problem Is No Longer Search — It Is Who Gets Trusted to Pay

Consumers are already using AI tools to compare products and choose what to buy. The harder question is whether those tools should be allowed to spend money. Visa chief executive Ryan McInerney told the Goldman Sachs Communacopia + Technology Conference that the company is seeing AI shopping adoption, but not yet consumer comfort with fully autonomous payments.

That distinction is the fresh part of the story. Agentic commerce is not waiting for people to ask chatbots for product ideas; it is waiting for a trust layer that lets merchants, issuers and networks decide whether an AI agent is legitimate. Visa’s Trusted Agent Protocol and Intelligent Commerce pages frame the problem as identity, consent and transaction safety, not only checkout convenience.

Trust moves upstream

Traditional card fraud controls examine a payment after it is presented. Agentic commerce moves the risk earlier. A merchant may need to know whether an agent is authorised by a customer, whether the buying instruction is genuine and whether the agent can be held to a payment rule. That makes the verification layer as important as the payment credential itself.

The same shift is visible in other consumer technology stories. Karmactive has covered how AI queue systems are entering transport and charging decisions, while surveillance pricing rules show regulators are watching how data changes consumer prices. Visa’s push sits between those two forces: AI convenience on one side and consumer-protection questions on the other.

Stablecoins and onchain credit remain part of Visa’s wider strategy, but they are adjacent to this story. The immediate bottleneck is not whether a payment rail exists. It is whether shoppers and merchants trust an autonomous software actor enough to let the rail operate.

If Visa’s framework becomes a practical standard, smaller fintechs and AI-shopping start-ups may have to build around it. If it does not, agentic payments could remain stuck at the edge of checkout while humans keep clicking the final payment button.