Visa’s Chief Product and Strategy Officer just said the quiet part loud. Jack Forestell told attendees at the Wolfe Research FinTech Forum last week that “the agentic web is the biggest opportunity that I’ve seen in my 20-plus years in payment technology.”

That’s not a throwaway line from a mid-level product manager. It’s the strategic thesis of a $592 billion company that processes roughly $16.7 trillion in payment volume annually. When Visa says something is the next big thing, the payments industry tends to pay attention — or scramble to catch up.

What agentic commerce actually means

Here’s the concept in plain terms: instead of you browsing a website, comparing prices, entering your card number, and clicking “buy,” an AI agent does all of that for you. It shops, negotiates, selects, and pays — autonomously. Think of it as giving your credit card to a very competent personal assistant who never sleeps and never impulse-buys snacks at checkout.

The market opportunity is substantial. By 2030, AI agents are projected to handle between 15% and 25% of all US e-commerce purchases. That translates to somewhere between $300 billion and $500 billion in transaction volume, up from roughly $3 billion in 2025. For context, $500 billion is larger than the entire GDP of Norway.

Amazon is already seeing early proof that this works. Its AI assistant Rufus reached 250 million active users in 2025, and shoppers who used it were 60% more likely to complete a purchase compared to those browsing without agent assistance. Less friction means more conversions. More conversions mean more revenue flowing through payment networks.

Forestell laid out four specific ways agentic commerce expands the payments ecosystem, and each one maps directly to Visa’s business model.

The four pillars of Visa’s agentic thesis

First, agents reduce payment friction. Digital transactions still suffer from surprisingly high fall-off rates and declines. An AI agent can optimize payment routing, retry failed authorizations, and select the best payment method in milliseconds. When success rates climb, transaction volumes climb with them — a rising tide that lifts everyone from consumers to card issuers.

Second, transaction density accelerates. Visa’s transaction count has more than tripled over the past decade, driven by subscriptions, streaming, and other low-ticket, high-frequency payment models. Agents push this further by breaking purchases into even smaller increments. Imagine paying for cloud computing by the second or streaming by the minute rather than subscribing monthly. Each micro-transaction is another swipe through Visa’s network.

Third, B2B payments finally modernize. This is arguably the sleeper opportunity. Business-to-business payment flows remain shockingly manual — supplier onboarding, invoicing, reconciliation, and payment execution all involve friction that agents can eliminate. The B2B payments market dwarfs consumer commerce, and it’s been stubbornly resistant to digitization. Agents might be the force that finally drags it into the modern era.

Fourth, overall economic activity expands. Forestell’s argument here follows a historical pattern: every major wave of payments innovation — from credit cards to e-commerce to mobile wallets — has grown the total pie rather than just redistributing existing slices. Agentic commerce, he believes, will do the same through pure efficiency gains.

The competitive chess match

Visa isn’t making this bet in a vacuum. Mastercard announced its Agent Pay platform and integrated it with PayPal’s wallet back in October 2025, signaling that the two largest card networks are racing toward the same destination. The question isn’t whether agentic commerce happens — it’s who controls the infrastructure.

Visa has some built-in advantages. The company’s network connects over 14,500 financial institutions globally, and its AI-powered fraud detection systems analyze roughly 300 billion transactions per year. Last year alone, Visa detected and prevented more than $40 billion in fraudulent activity, with an 85% improvement in fraud detection attributed to machine learning capabilities.

That fraud angle matters more than it might seem at first glance. E-commerce fraud is projected to balloon from $56 billion in 2025 to $131 billion by 2030. Autonomous agents making purchases without direct human oversight create new attack surfaces. A network that can verify whether an agent is legitimate — and whether its instructions are authorized — has a massive competitive moat.

To that end, Visa has developed what it calls a Trusted Agent Protocol, designed specifically for secure agent-mediated transactions. Think of it as a digital handshake that confirms an AI agent is who it claims to be before any money moves. In a world where bots are buying things on your behalf, trust verification becomes the product.

Look, the broader implication here goes beyond just Visa and Mastercard. Around 30% to 45% of US consumers already use generative AI tools for product research. That adoption curve is steep. Retailers who don’t integrate agentic capabilities risk losing customer data and purchasing control to third-party platforms — a dynamic that mirrors what happened when Amazon became the default shopping destination a decade ago.

The average payments company now employs over 30% more AI-focused workers than traditional financial institutions, according to industry data. That talent gap tells you where the sector thinks the future lies.

What investors should watch

For investors, the agentic commerce narrative introduces both opportunity and risk. On the upside, increased transaction density and higher success rates directly benefit Visa’s revenue model, which earns fees on each transaction processed. More transactions per dollar spent means more fee revenue without requiring consumers to spend more money overall.

The risk side is more nuanced. AI agents are inherently cost-optimizers. They’ll route payments through the cheapest available channel, which could pressure interchange fees over time. If agents consistently steer transactions toward lower-cost alternatives — say, account-to-account transfers or stablecoin rails — traditional card networks could face margin compression even as volumes grow.

There’s also the question of who captures the customer relationship. Today, Visa sits between the consumer and the merchant. In an agentic world, the AI platform might become the new intermediary, potentially commoditizing the payment rail underneath. Visa’s Trusted Agent Protocol is partly a defensive play against this scenario — an attempt to remain essential rather than invisible.

The competitive landscape will reward companies that establish robust agent-to-merchant protocols early. First-mover advantage in setting standards tends to be durable in payments infrastructure. Visa’s scale gives it leverage, but the company will need to execute quickly. Mastercard is already live with Agent Pay, and fintech startups unencumbered by legacy systems could move even faster.

Bottom line: Forestell’s declaration isn’t just corporate enthusiasm — it reflects a genuine structural shift in how commerce gets executed. The transition from human-initiated to agent-initiated transactions could be as transformative as the shift from cash to cards. Visa is betting its strategic future on being the trusted plumbing for autonomous commerce. Whether that bet pays off depends on execution, but the $300-500 billion opportunity by 2030 suggests the stakes are high enough to warrant close attention.