Over the past year, AI-driven shopping traffic to U.S. retail sites has surged by 4,700%. This spike left merchants scrambling to distinguish legitimate AI agents from malicious bots - a difficult task that threatens security and erodes user trust.

Visa’s answer is the Trusted Agent Protocol (TAP): a system to cryptographically verify in real time that an AI agent making a purchase is indeed legitimate and truly acting on a consumer’s behalf.

TL;DR

• AI-driven commerce is exploding, but merchants can’t tell which AI agents to trust.

• Visa’s Trusted Agent Protocol (TAP) cryptographically verifies that an AI agent is legitimate and truly acting on a user’s behalf.

• Platforms like Oscilar add an intelligent risk layer on top of TAP - using real-time AI to detect fraud, anomalies, and compliance issues.

• Together, TAP and Oscilar form the new “trust stack” of agentic commerce - keeping AI-led payments smooth, secure, and transparent.

How Visa TAP Works: Real-Time AI Agent Verification Using Cryptographic Proofs

TAP adds a digital proof-of-identity to every agent-initiated transaction. It uses cryptographically signed HTTP messages to transmit an agent’s intent, verified user identity, and payment details. Merchants then validate the signature using Visa’s public keys, confirming the agent’s authenticity.

In this way, TAP creates a standard trust layer that operates within existing web infrastructure (leveraging the HTTP Message Signatures standard) and Visa’s card network. The protocol is openly available (via Visa’s Developer Center), with early adopters including Nuvei, Adyen, and Stripe.

Agentic Finance Explained: Why AI-Driven Payments Are Visa’s Next Strategic Frontier

As commerce shifts toward agentic finance - where autonomous agents handle shopping and payments - controlling the trust layer is vital. Visa sees TAP as a way to stay central even when AI agents (not humans) initiate purchases. By acting as the verification backbone between consumers’ AI agents and merchants, Visa aims to preserve its role as payments become more automated and invisible to users.

If AI agents handle most purchases by 2030, the power will lie with whichever standard verifies those agents. Visa is betting that TAP will be that universal standard - the “plumbing” of agent-driven commerce that keeps Visa indispensable.

Visa TAP vs Google AP2 vs OpenAI Checkout: Competing Protocols for AI-Powered Payments

Visa isn’t alone in tackling AI-led commerce. Google’s Agent Payment Protocol (AP2) is an open, payment-agnostic standard for agents to transact via cards, bank transfers, or even stablecoins, using cryptographic user mandates to prove consent. Earlier, OpenAI and Stripe, meanwhile, have launched a conversational checkout (ChatGPT’s Instant Checkout) where an AI assistant can recommend and sell products entirely within a chat.

Visa’s TAP therefore differs by focusing on verifying the actor rather than reinventing the interface. In other words, Google and OpenAI are reimagining how AI-driven payments begin, while Visa is ensuring that whoever (or whatever) initiates a payment can be trusted as legitimate.

Oscilar and Visa TAP: Building the Trust Stack for Secure AI Commerce and Fraud Prevention

Cryptographic protocols like TAP are necessary but not sufficient - they must be complemented by intelligent fraud prevention. Platforms like Oscilar provide real-time risk infrastructure to validate an agent’s behavior and context, not just its credentials. More importantly, Oscilar’s unified platform spans fraud detection, identity/credit risk, and compliance checks. In practice, its AI-powered system monitors agent-led transactions for anomalies or red flags, adding a dynamic machine-learning defense layer on top of TAP’s identity proof.

Visa itself acknowledges that as AI-driven fraud evolves, networks will need ML-based verification layers on top of signatures. Oscilar already answers this by deploying its own AI agents in a risk decisioning platform that continuously learns from new fraud patterns and adapts controls.

By integrating solutions like Oscilar, banks and merchants form a layered “trust stack”: TAP assures an agent is who it claims to be, while Oscilar confirms in real time that the agent’s actions are safe. This combination keeps agentic transactions smooth yet secure, preserving trust as commerce becomes more and more automated.

The Future of Agentic Commerce: How Visa TAP and Oscilar Will Shape AI-Driven Payments by 2030

In the coming years, agentic commerce - with AI assistants handling increasingly more transactions - could fundamentally reshape digital markets.

Protocols like TAP, coupled with real-time risk intelligence from players like Oscilar, will form the backbone of this new infrastructure. Businesses that embrace this trust stack early can thus ensure that as commerce becomes more automated and invisible to consumers, the underlying payments remain secure and transparent.

Visa’s TAP and similar efforts from OpenAI/Stripe, Coinbase or Google may well become essential plumbing for the AI-driven economy, defining how value is exchanged in the next decade. But this won’t work unless trust is ensured.