← Back to news

Mastercard Adds a Trust and Intelligence Layer to Agent Pay

Mastercard is testing a probability score that tells issuers when a card transaction was initiated by an AI agent, part of a new intelligence layer for its Agent Pay program built with Cloudflare and Skyfire.

Author
A2AWire
Published
Category
Industry
Reading time
2 min read

On September 30, Mastercard expanded Agent Pay, its agentic payments program, with new trust and intelligence services aimed at AI-initiated card transactions. The first service, now in testing with U.S. issuers, is a probability score estimating how likely a transaction was started by an AI agent, so banks can approve legitimate agent-led purchases instead of declining them out of caution.

Highlights#

  • The services form the intelligence layer of Mastercard's Agent Pay Trust Framework, which spans identity, intent, controls, execution and intelligence. The goal is a shared view of AI-driven activity so issuers and merchants can make smarter authorization decisions with less unnecessary friction.
  • Mastercard says the score will deepen over time with signals covering agent behavior, merchant risk, transaction patterns, credential risk and consumer propensity, building a fuller picture across the transaction lifecycle.
  • The signals are designed to answer three questions for every AI-led purchase: is the activity consistent with expected patterns, does anything about the agent, merchant or credential look unusual, and should the transaction be approved or held for review.
  • Mastercard is working with Cloudflare on privacy-preserving environments that combine web and payment network signals, giving banks and merchants visibility into AI-powered transactions. It is also partnering with Skyfire, whose Know Your Agent technology handles identity and verification for AI agents, to help the ecosystem recognize trusted agents.
  • The company cites a projection that one in 10 consumers will routinely use agents to make purchases by 2030, framing trust infrastructure as the gating factor for that shift. Capital One's payments strategy chief endorsed the need for greater visibility into how agent transactions are initiated and executed.

Why it matters#

Agent payments only scale if the party authorizing the transaction can tell what is acting and whether it is trustworthy. A network-grade intelligence layer that scores agent-initiated activity and pairs it with verifiable agent identity is the card world's version of what escrow and reputation rails provide on open settlement networks.

Sources#