Introduction and Main Facts
The architecture of global card payments is facing its most radical structural evolution since the transition from magnetic stripes to EMV chips. As artificial intelligence moves from a passive tool for product discovery to an active participant capable of executing financial transactions on behalf of humans, global payment standards bodies are racing to adapt. At the center of this transformation is EMVCo—the technical consortium jointly owned by American Express, Discover, JCB, Mastercard, UnionPay, and Visa—which is actively developing a comprehensive technical framework for agentic payments.
Unlike traditional card-present or e-commerce transactions, where a human consumer manually initiates a payment at the point of checkout, agentic payments involve autonomous or semi-autonomous software agents. These AI agents execute purchases based on standing instructions, constraints, and budgets established by the user hours, days, or even weeks prior to the actual exchange of funds.
The core challenge facing merchants, payment networks, and card issuers today is a profound informational gap: while an AI agent may present a valid payment credential, the traditional payment data stream does not routinely transmit what the consumer authorized the agent to do, nor does it indicate whether a specific transaction falls within the boundaries of that delegated authority.
To bridge this chasm, EMVCo closed a public consultation period on September 30 for its foundational blueprint titled Agentic Payments — Framework for Specifications. The framework introduces a standardized layer designed to carry critical metadata through the payment flow, including:
- Consumer Intent: Clear signals outlining the precise parameters, spending limits, and product conditions authorized by the human user.
- Agent Identification: Protocols to consistently identify the software agent executing the transaction.
- Agentic Transaction Indicators: Binary or granular signals notifying the payment ecosystem that an AI agent—rather than a human cardholder—was actively involved in orchestrating the purchase.
Crucially, EMVCo’s proposed framework acts strictly as a data and interoperability foundation. It does not dictate commercial business rules or dictate how individual products are deployed, leaving transaction approval decisions and risk policies firmly in the hands of issuers, networks, and acquiring banks.
Chronology of the Initiative
The push toward standardizing agentic payments is moving at a pace dictated by the rapid commercialization of generative AI and automated retail. The timeline of this regulatory and technical development highlights how quickly the payments industry is responding to consumer behavior shifts:
- June 2024: PYMNTS Intelligence releases data showing early mass consumer readiness for automated purchasing, revealing that over a third of consumers are willing to delegate payment authorization and budget management to AI agents.
- August 2024: Research data tracks a massive consumer migration toward AI-driven discovery, noting that nearly one in five U.S. consumers begin their retail research journeys using artificial intelligence tools.
- September 1, 2025 (or recent cycle): EMVCo officially issues a formal request for public comment on its Agentic Payments — Framework for Specifications, outlining the vision for a "common foundation" to manage consumer intent, agent authentication, and transactional metadata.
- September 30, 2025: The public consultation and feedback window officially closes. Industry stakeholders, technology firms, merchants, and financial institutions submit their technical assessments and critiques.
- Post-September 2025 (Current Phase): The EMVCo Agentic Payments Task Force begins the arduous process of collating feedback, reviewing ecosystem responses, and mapping out structural updates for established EMV technologies such as 3-D Secure, Tokenization, and Digital Payment Credentials.
Supporting Data: The Rise of the AI-Powered Consumer
The urgency behind EMVCo’s technical framework is rooted in hard economic data demonstrating that consumers are rapidly shifting trust from human-driven checkout flows to automated AI systems. According to comprehensive studies from PYMNTS Intelligence, the retail landscape is experiencing structural disruption at every stage of the consumer journey.
The Readiness for Delegation
The Global Digital Shopping Index: The AI-Powered Shopper Has Arrived illustrates a clear hierarchy of trust when it comes to delegating financial power to artificial intelligence:
- 56% of consumers are comfortable allowing an AI agent to search, aggregate, and compare products across digital merchants.
- 37% state they would trust an AI agent to officially authorize and execute payments on their behalf.
- 36% are open to predictive automatic buying, where algorithms replenish household or personal goods without direct human intervention.
- 35% are willing to grant software agents direct access to their saved payment methods and digital wallets.
The Shift in Product Discovery
AI is no longer just closing sales; it is initiating them. Findings from The 50 Million Consumer Migration: The Data Behind Retail’s Shift Toward AI Discovery demonstrate that 19% of U.S. consumers now initiate retail product research directly through AI platforms rather than traditional search engines or marketplace homepages.
For consumers utilizing AI during the discovery phase, the financial impact is tangible:
- 43% successfully discovered a better price point for the desired item.
- 27% ultimately switched retail brands based on AI recommendations.
- 26% pivoted to entirely different product categories guided by algorithmic insights.
This data underscores a fundamental commercial reality: consumers are utilizing AI to navigate complex purchasing decisions, establish multi-variable constraints, and optimize spending. Consequently, the downstream payment infrastructure must evolve to understand and process the complex instructions generated upstream.
Official Responses and Stakeholder Perspectives
Navigating the intersection of artificial intelligence and legacy financial rails requires delicate coordination between technical standard-setters and commercial institutions.
Oliver Manahan, Director of Engagement and Operations for EMVCo, has been vocal about the boundaries and ambitions of the new framework. Speaking on the nature of the initiative, Manahan emphasized that the standards body is purposefully avoiding regulatory overreach into commercial policy.
"It does not define the business rules, in other words, how a product is used," Manahan explained in an interview with PYMNTS.
Instead, the framework focuses purely on technical interoperability. A major pillar of this technical approach is the concept of "Intent Services," which envisions a shared infrastructural layer where ecosystem participants can register, reference, retrieve, and manage consumer-authorized intent before, during, and after a transaction takes place.
Manahan also highlighted that consumer intent cannot simply vanish once an authorization is granted. Because payment disputes, chargebacks, and fraud investigations often occur weeks or months after a purchase, the ecosystem must maintain access to historical intent data.
"EMVCo does recognize that intent must be continuously managed from a lifecycle perspective, as entities within the ecosystem may need to retrieve historical information for dispute resolution," Manahan noted.
Furthermore, EMVCo is exploring Know Your Agent (KYA) capabilities. Much like Know Your Customer (KYC) regulations designed to verify human identity, KYA would provide a consistent mechanism to identify software agents, bind them to verifiable metadata, and establish an unbroken chain of accountability. According to Manahan, these agentic transaction indicators will be vital for:
- Enhanced risk assessment and predictive fraud prevention.
- Streamlining complex dispute resolution processes.
- Ensuring operational transparency, auditability, and regulatory reporting.
Implications for the Payments Ecosystem
The integration of EMVCo’s agentic payment framework carries profound operational, security, and strategic implications for every stakeholder in the global payments value chain.
1. For Issuers: Redefining Fraud and Risk Models
Traditional card-issuer fraud models rely heavily on behavioral biometrics, device fingerprinting, and location data to determine if a human cardholder is legitimately interacting with a merchant. When an autonomous AI agent executes a transaction at 3:00 AM from a server farm halfway across the world, standard velocity checks and geofencing alarms might falsely flag the activity as malicious.
By incorporating agentic transaction indicators and verifiable consumer intent metadata, issuers gain visibility into the context of the transaction. An issuer can instantly verify that while the agent is operating autonomously, its actions strictly comply with a $50 budget limit established by the human cardholder on Tuesday. This dramatically reduces false-positive declines and preserves conversion rates.
2. For Merchants: Adapting to Algorithmic Buyers
Merchants must prepare for a future where a significant percentage of their digital storefront traffic consists not of human browsers, but of programmatic AI agents executing high-speed, programmatic queries and transactions.
Merchants will need to ensure their checkout APIs and payment gateways are capable of ingesting and passing along agentic metadata. Furthermore, the nature of marketing will shift: rather than optimizing web pages exclusively for human eyeballs and search engine optimization (SEO), brands will increasingly need to optimize product data, pricing APIs, and inventory feeds for AI agents scanning the web for optimal value.
3. For Payment Networks and Standards Bodies
The close of the public comment period marks the transition from conceptual design to technical implementation. The EMVCo Agentic Payments Task Force is currently evaluating how the feedback gathered will reshape foundational technical suites.
Future enhancements may be integrated into several core EMV technologies:
- EMV 3-D Secure (3DS): Adapting authentication protocols to authenticate the agent’s delegation rights rather than just prompting a human for a one-time passcode (OTP).
- EMV Payment Tokenization: Issuing specialized tokens tied directly to agentic credentials with embedded programmatic restrictions.
- EMV Secure Remote Commerce (SRC) and Digital Payment Credentials: Standardizing how intent and KYA metadata travel alongside tokenized card data in digital wallets.
Conclusion
As artificial intelligence transitions from a novelty to an indispensable commercial assistant, the financial plumbing of the global economy must modernize to match its speed and complexity. EMVCo’s framework for agentic payments represents a vital first step toward creating a secure, transparent, and interoperable ecosystem. By standardizing how consumer intent and agent identity travel through the payment lifecycle, the industry is laying the groundwork for a frictionless future where humans and machines can transact with absolute trust.
