The Continued AI Growth in Retail Transactions

As artificial intelligence increasingly influences purchasing decisions, Visa introduces its Trusted Agent protocol to enhance merchants’ insight into these processes.

In the emerging agentic commerce landscape, merchants must be able to evaluate and distinguish between genuine AI agents and fraudulent entities. Visa’s framework leverages unique identifiers like cryptographic signatures to provide this capability.

These identifiers can offer insights such as the specific products an agent is seeking or evidence of previous interactions between consumers and merchants. The system also verifies whether an agent supports payment methods compatible with a merchant’s preferred checkout options.

Guarding Agentic Commerce

This solution was driven by the increasing prevalence of AI in retail, highlighted by Adobe data showing generative AI traffic surged 4,700% year-over-year as of July. Many consumers who have used AI reported improved shopping experiences.

In response, more businesses are integrating AI across various applications. For instance, Klarna partnered with Google to use its AI models in creating personalized visuals and tailored marketing campaigns within the Klarna app.

Agentic commerce takes this a step further by allowing AI agents to initiate and complete payments, thus sparking concerns over transaction safety and security.

In Google’s Agent Payments Protocol (AP2), safeguards are implemented via mandates that ensure an AI agent adheres to user instructions, detailing the parameters and timing of purchases.

Regardless of the specific protocol used, robust security and fraud mitigation measures are essential for agentic commerce to progress. This presents a challenge: beyond detecting bots and fraudulent activities, systems must also minimize false positives while maintaining transparency throughout the process.