- 140+ payment service providers, acquirers, merchants, and B2B platforms have joined Visa's Agentic Ready program, including Adyen, Stripe, PayPal, and J.P. Morgan Payments.
- Visa's Trusted Agent Protocol (TAP) introduces agentic network tokens to securely identify AI-driven transactions.
- B2B platforms like Billtrust and HighRadius are already piloting AI agents for automated invoice payments.
Experts agree that Visa's Agentic Ready program represents a critical step toward standardizing autonomous AI commerce, though legal and liability frameworks must evolve to support this shift.
When Algorithms Buy: Visa Lays the Tracks for Autonomous AI Commerce
SAN FRANCISCO, CA – November 19, 2026 — For years, the promise of artificial intelligence in commerce was largely conversational: a chatbot recommending a sweater or a predictive algorithm suggesting a timely inventory restock. But the paradigm is shifting from recommendation to execution. Autonomous AI agents are no longer just browsing; they are buying. This transition from human click to machine transaction represents a profound shift in digital commerce, one that requires an entirely new set of foundational rails to ensure trust, security, and accountability.
Enter Visa. The global payments giant has announced a massive expansion of its Agentic Ready program, shifting its focus to the acceptance ecosystem. Building on an initial cohort of more than 175 card issuers across 45 markets, the expanded initiative brings together over 140 payment service providers, acquirers, merchants, and B2B platforms. Heavyweights including Adyen, Stripe, PayPal, J.P. Morgan Payments, and Checkout.com have joined the fold, signaling a broad industry consensus that the era of "agentic commerce" is imminent.
"Agentic commerce is already taking shape, but scaling it requires more than innovation," noted Rubail Birwadker, SVP of Growth Products & Partnerships at Visa. "Merchants need confidence in who they're transacting with, and payment providers need to know these transactions can move securely and reliably through existing infrastructure."
The Infrastructure of Intent
How do you distinguish a legitimate AI assistant buying office supplies from a malicious bot attempting credential stuffing? The answer lies in identity and tokenization. Visa's approach centers on the deployment of "agentic network tokens" and the Trusted Agent Protocol (TAP).
Standard network tokens revolutionized e-commerce by replacing sensitive card numbers with cryptographic equivalents, ensuring that a compromised token could not be used elsewhere. Agentic tokens take this a step further by embedding metadata that specifically identifies the transaction as originating from a recognized AI agent. This allows for granular control, risk assessment, and policy enforcement specific to machine-led activity.
Companies like Basis Theory and VGS are working alongside Visa to serve as the enablement layer for these credentials. By establishing a clear, cryptographic proof of identity, TAP allows merchants to apply specific risk models to machine-led transactions. "As AI agents change how people shop and pay, merchants need a simple, secure way to support these new experiences," explained James Armstead, CTO for Basis Theory.
This infrastructure is not about blocking bots; it is about transparently identifying legitimate automation. Major payment service providers like Adyen and Stripe are actively building out the APIs and risk engines to parse these agentic signals without adding friction to the checkout process. As Karan Katyal, Global Head of Agentic Commerce at Adyen, pointed out, defining clear, actionable standards is an essential step toward empowering merchants to let this channel scale.
Navigating the Liability Labyrinth
Yet, building the technical rails is arguably the easier half of the equation. The legal and liability frameworks surrounding autonomous purchases remain a murky frontier. What happens when an algorithm makes a mistake?
Consider a scenario where a consumer instructs their AI agent to purchase a specific brand of electronics, but due to an AI "hallucination" or a prompt injection attack, the agent orders thousands of dollars of incorrect inventory from a different vendor. Under current payment card network operating regulations, unauthorized transactions are well-defined, usually involving stolen credentials. But an AI hallucination blurs the line between "unauthorized" and "unintended."
Legal experts and regulatory attorneys point out that current frameworks, including the Consumer Financial Protection Bureau's (CFPB) guidance on automated financial transactions, are ill-equipped for this nuance. Is the consumer liable for poorly prompting the agent? Is the developer of the large language model responsible for the hallucination? Or does the merchant, who faithfully fulfilled the digital order, bear the cost of the return and chargeback?
Acquirers and financial institutions act as the critical bridge in navigating this new ecosystem. J.P. Morgan Payments, one of the largest merchant acquirers in the United States, is collaborating with Visa to define processing requirements and standards to mitigate these exact risks. "As one of the largest merchant acquirers in the U.S., we have an obligation to advance the trust and safety of the ecosystem," stated Mike Lozanoff, global head of merchant services for J.P. Morgan Payments. "That requires us to participate in all viable approaches to agentic commerce so that we can provide unbiased, specific, and timely advice and solutions to our clients."
The Quiet Revolution in B2B
Because of the consumer-side complexities and liability questions, the true vanguard of agentic commerce isn't found in retail—it is hidden in the back offices of corporate finance. While the idea of an AI buying groceries captures the public imagination, B2B platforms are quietly deploying autonomous agents to handle accounts payable, reconciliation, and invoice settlements. This is where the tangible difference of agentic commerce is currently most visible.
The B2B environment offers a highly controlled, high-value proving ground. Platforms like Billtrust, HighRadius, and Versapay are utilizing the Agentic Ready program to test agentic bill pay. In these workflows, AI agents can access approved payment flows and submit Visa credentials while preserving strict business visibility and supplier controls.
"With Visa we launched the first live pilot of trusted AI agents completing approved invoice payments end to end through supplier portals," said Kunal Patel, SVP of Payments at Billtrust. "Agentic Ready is how that scales responsibly across the ecosystem. Suppliers keep their acceptance policy and full visibility, and they release trapped cash faster."
This controlled automation eliminates manual portal logins and emailed credit card numbers, seamlessly streamlining the order-to-cash flow. As Vikram Gollakota, Senior Vice President at HighRadius, observed, the ultimate vision is a world where AI agents take over the repetitive transactional work traditionally done by humans.
The Race for the Future of Checkout
Visa is not alone in recognizing this monumental shift. Across the broader payment ecosystem, competitors are mobilizing. Mastercard continues to invest heavily in AI-powered secure digital commerce, while open banking initiatives in Europe present a formidable alternative. Through APIs, open banking could allow AI agents to initiate account-to-account transfers directly, potentially bypassing traditional card networks altogether for certain B2B transactions. Furthermore, smart contracts on blockchain protocols offer yet another decentralized paradigm for machine-to-machine payments.
However, Visa's aggressive push to unite the acceptance ecosystem under a standardized framework is a calculated move to ensure its existing infrastructure remains the bedrock of future commerce. By bringing acquirers, PSPs, and merchants to the same table, the network is attempting to solve the interoperability puzzle before fragmentation takes hold.
The transition to agentic commerce will not happen overnight. It requires a delicate balancing act between rapid technological innovation, the establishment of clear legal boundaries, and the meticulous construction of trust. But as the lines between human intent and machine execution continue to blur, the foundational work being done today by coalitions like the Agentic Ready program will dictate how the global economy transacts tomorrow.
Topics & Related
📝 This article is still being updated
Are you a relevant expert who could contribute your opinion or insights to this article? We'd love to hear from you. We will give you full credit for your contribution.
Contribute Your Expertise →