The global financial services industry is currently witnessing a fundamental shift in the conceptualization of automated digital interactions. For decades, automated "bots" were viewed by financial institutions primarily as security threats—vectors for credential stuffing, scraping, and fraudulent transactions. However, Visa, the world’s largest payment processor, has formally signaled a strategic pivot, embracing a future where autonomous artificial intelligence agents act as legitimate consumers and intermediaries. This emerging paradigm, termed "agentic commerce," represents a significant evolution in how value is exchanged across the global economy.
According to Ryan McInerney, Chief Executive Officer of Visa, the advent of agentic commerce is poised to expand the company’s total addressable market by accelerating the digitization of commerce, increasing transaction frequency through micro-payments, and streamlining the historically friction-laden B2B payment sector. This transition marks the third major wave of digital commerce evolution, following the initial rise of e-commerce in the late 1990s and the mobile commerce revolution of the 2010s.
The Strategic Shift: From Bot Mitigation to Agent Enablement
The historical stance of the payments industry toward automated scripts was one of containment. Oliver Jenkyn, Visa’s Group President of Global Markets, recently noted that for over 20 years, the industry’s primary goal was to stop bots. The shift in perspective stems from the increased sophistication of Large Language Models (LLMs) and their ability to perform complex, goal-oriented tasks. Unlike the "dumb" bots of the past, modern AI agents possess the reasoning capabilities to make purchasing decisions, optimize for price, and manage logistics on behalf of human users.
Visa’s leadership argues that this "agentic" shift will benefit the company in four distinct ways. First, it acts as a catalyst for global digitization. Much like how smartphones brought millions of unbanked and underbanked individuals into the digital economy, AI agents are expected to bring even the most granular daily tasks into the sphere of digital payments.
Second, the volume of transactions is expected to surge. In a traditional commerce model, a consumer might make a single bulk purchase. In an agentic model, an AI could split that purchase into multiple smaller transactions to optimize for shipping times, price fluctuations, or supply chain availability. Furthermore, these agents may begin to pay for their own operational costs—such as data usage or computing power—on a per-event basis, creating a massive new category of micro-transactions that were previously economically unfeasible.
Chronology of Innovation: Visa’s Path to Autonomous Payments
Visa’s preparation for this shift did not occur in a vacuum. The company has spent the last decade building the infrastructure necessary to support delegated authority—the process of allowing a third party (in this case, an AI) to spend money on behalf of an account holder.
- 2014 – The Launch of Visa Token Service: Visa introduced tokenization to replace sensitive 16-digit card numbers with a digital identifier. This technology is now the cornerstone of agentic commerce, as it allows users to issue "restricted" tokens to AI agents that can only be used at specific merchants or within certain spending limits.
- 2020-2022 – Pandemic-Driven Digitization: The global shift to contactless and card-not-present transactions provided Visa with a massive data set to train its fraud detection algorithms, which now process over 900 million transactions daily.
- Late 2023 – The Pivot to Agentic Commerce: Leadership began publicly discussing the transition from "defensive" AI to "productive" agentic commerce.
- 2024 – Product Launches: Visa introduced "Intelligent Commerce Connect," a platform designed to serve as a bridge for developers building AI agents. This was followed by the proof-of-concept for "Visa CLI" (Command Line Interface), allowing developers to integrate payment credentials directly into coding environments and automated workflows.
The B2B Frontier and Macroeconomic Impact
While consumer-facing AI agents garner the most media attention, the most significant immediate impact of agentic commerce may be in the Business-to-Business (B2B) sector. Global B2B payment volume is estimated to exceed $120 trillion annually, yet a staggering portion of this remains tied to manual processes, paper checks, and slow wire transfers.
McInerney highlights that AI agents can remove the "enormous friction" inherent in these transactions. An autonomous agent can receive an invoice, cross-reference it with a digital contract, verify that goods or services were delivered, and initiate a payment via a virtual card or tokenized credential without human intervention. This level of autonomy requires a high degree of trust and a robust set of rules, which Visa intends to provide through its existing network of 14,500 financial institutions.
The macroeconomic implications are equally substantial. External economic analyses cited by Visa suggest that AI integration could provide an incremental boost to global GDP of 80 to 150 basis points. As GDP grows, the total volume of consumer and business spending naturally follows, creating a virtuous cycle for payment processors. By positioning itself as the "rails" for these AI-driven transactions, Visa seeks to capture a percentage of this new economic growth.
Infrastructure for an Agentic World: Scale and Security
To capitalize on this trend, Visa is leveraging its massive scale. The company’s network currently encompasses over 175 million seller locations and 5 billion credentials across 200 countries. However, the move to agentic commerce introduces new risks, particularly regarding identity and intent.
The challenge lies in distinguishing between a "good" agent authorized by a user and a "bad" bot attempting to exploit a system. Visa’s strategy relies heavily on its data advantage. By processing 300 billion transactions annually, the company uses AI to protect the ecosystem. In an agentic world, Visa argues that tokens will become the primary "identity" of the transaction. Because tokens can be limited in scope, they provide a layer of security that traditional credit card numbers cannot.
Furthermore, Visa is positioning itself as a more stable alternative to decentralized finance (DeFi) solutions like stablecoins. While some proponents of AI commerce suggest that agents should use cryptocurrencies to avoid traditional banking fees, McInerney argues that merchants and buyers prefer the established legal protections, liquidity management, and "Know Your Customer" (KYC) protocols that come with the Visa network. Unlike many blockchain-based transactions, Visa transactions offer dispute resolution and fraud protection—features that are essential for building the "trust" necessary for users to let agents spend their money.
The Developer Ecosystem and the CLI as a Commerce Platform
A unique aspect of Visa’s recent initiatives is its focus on the developer community. The company has observed the rise of agentic coding assistants, such as Claude Code and GitHub Copilot, which enable developers to build and deploy software at unprecedented speeds. Visa’s view is that the Command Line Interface (CLI)—traditionally a tool for technical experts—is becoming a new commerce platform.
By launching the Visa CLI proof-of-concept, the company is allowing developers to pay for API access, cloud storage, or digital assets directly through their terminal. The goal is to make card acceptance as ubiquitous among API endpoints as it is among physical storefronts. This "developer-first" approach ensures that as new AI services are built, the payment mechanism is already integrated into the foundation of the code.
Challenges and Future Outlook
Despite the optimistic projections, the road to widespread agentic commerce is fraught with hurdles. The primary limiting factor is consumer trust. Entrusting an AI agent with financial autonomy requires a level of reliability that current LLMs have yet to fully demonstrate consistently. Issues such as "hallucinations" (where an AI provides incorrect information) or security vulnerabilities in the agent’s code could lead to unauthorized spending.
In response, Visa has indicated that it will continue to evolve its rules and standards. Just as the company developed "3-D Secure" and other protocols for e-commerce, it is now working on standards for "authenticated tokens" that carry data on user intent. This would allow an issuer to see not just who is paying, but why the agent is initiating the transaction.
Industry analysts suggest that the competition in this space will be fierce. Competitors like Mastercard are also investing heavily in AI-driven fraud detection and digital identity, while fintech giants like Stripe and Adyen are simplifying the integration of payments for AI startups.
Conclusion: A New Era of Value Exchange
Visa’s pivot to agentic commerce reflects a broader recognition that the nature of the "consumer" is changing. In the coming decade, a significant portion of economic activity may be initiated by algorithms rather than humans. By providing the infrastructure—the tokens, the security, and the network—Visa aims to ensure that it remains as central to the AI-driven economy as it was to the era of plastic cards.
As McInerney summarized, the goal is to create a system where buyers, sellers, and now agents all recognize the value of a standardized, secure, and globally accepted payment method. If Visa succeeds in scaling its "Intelligent Commerce Connect" and "Visa CLI" initiatives, it may well transform the command line into the most powerful cash register in history, ushering in an era of micro-transactions and autonomous B2B efficiency that could redefine global trade.