The news
Three developments in late September 2026 pointed at the same gap in agentic payments, where AI software buys things on a person's behalf: banks and merchants need a reliable record of what the customer actually authorized.
On September 29, Federal Reserve Governor Christopher Waller gave a speech titled "Payments in the Age of AI Agents" at the Sibos conference, according to PYMNTS. He said banks and networks must be able to verify that an agent has authority to pay for the buyer, which PYMNTS described as a machine-readable record of delegated authority covering spending limits, approved payment methods and expiry. Waller named three barriers: authentication, liability for agent mistakes, and fraud systems tuned to human behavior.
On September 30, Mastercard (MA) announced new trust and intelligence services for Agent Pay, its agentic payments program. The first, rolling out for testing in the United States, is a probability score estimating how likely it is that an AI agent initiated a transaction, intended to help card issuers approve legitimate agent purchases. Mastercard said it will add signals on behavior, merchant risk, transaction patterns and credential risk over time. It named Cloudflare as a partner on privacy-preserving analysis of AI-driven activity and Skyfire, a Know Your Agent (KYA) identity provider, as a partner on recognizing trusted agents.
Mastercard said one in 10 consumers is projected to routinely use agents to make purchases by 2030. Capital One's Todd Kennedy, executive vice president of payments strategy, was among those quoted backing the effort.
EMVCo, the body that manages the EMV chip and related card standards, has published a draft called Agentic Payments: Framework for Specifications, PYMNTS reported on October 2. It proposes Intent Services that let participants register, reference, retrieve and manage what a consumer authorized an agent to do before, during and after a transaction. Public comment closed September 30. EMVCo director of engagement Oliver Manahan told PYMNTS that intent must be managed across its lifecycle because parties may need historical records for dispute resolution, and that the framework does not set business rules.
The numbers
- Consumers projected to routinely use agents to buy by 2030
- 1 in 10 (per Mastercard)
- Consumers letting agents authorize payments
- 37% (cited by PYMNTS)
- Consumers letting agents search and compare
- 56% (cited by PYMNTS)
- EMVCo comment deadline
- September 30, 2026
Why CEOs should care
For merchants, the practical question is chargebacks. If a customer disputes a purchase an agent made, the outcome will depend on whether a recorded intent shows the agent stayed within its instructions. Payments leaders should ask their acquirers and processors which agent signals they will pass through, and whether Mastercard's agent score or EMVCo-style intent records will change dispute outcomes.
For issuers and bank risk teams, Waller's three barriers are a checklist. Fraud models built on human behavior will flag legitimate agent activity, such as one approval leading to bookings at several merchants, the example Mastercard gave. CISOs should map how agent identity, delegated authority and spending limits will be verified, logged and audited.
For boards and product leaders building agent features, the standard is not yet settled. EMVCo explicitly leaves business rules to others, so liability allocation will come from network rules, contracts and possibly regulators. Build intent logging now rather than retrofitting it.
The bigger picture
Card payments rest on proving that the cardholder was present or consented. Agentic commerce breaks that assumption, and the industry is racing to replace it with registered agents, intent records and risk scores. Visa's Intelligent Commerce effort, cited by PYMNTS, pursues similar goals, so whether networks converge on one standard through EMVCo or compete with their own frameworks will matter for merchants that must support both.
What’s next
Watch for EMVCo's response to public comments and a final specification, results of Mastercard's U.S. test of its agent score, and any Fed or network guidance on who bears liability when an agent errs.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
Companies in this story








