Mastercard Deploys New Risk Insights to Support Agentic Payments
Mastercard has added new trust and intelligence services to its agentic payments program, Agent Pay, the company said in a Wednesday (Sept. 30) press release emailed to PYMNTS. The new services bring together identity, intent, behavioral and fraud insights to help financial institutions and merchants make smarter authorization decisions, reduce unnecessary friction and provide more […] The post…
Mastercard has enhanced its Agent Pay program with new trust and intelligence services, the company announced on Wednesday (Sept. 30). These services integrate identity, intent, behavioral and fraud insights to assist financial institutions and merchants in making more informed authorization decisions, reducing unnecessary friction and providing a smoother experience for consumers.
The new features include a probability score to gauge the likelihood of a transaction being initiated by an AI agent, as well as additional agentic trust and intelligence signals to determine if the activity is unusual and may need further review, according to the release. Ann Johnson, Mastercard's executive vice president of security solutions, emphasized the importance of trust in AI agents, stating that they can make commerce more intuitive, efficient and personal, provided people and businesses can trust the systems acting on their behalf.
Mastercard launched Agent Pay in April 2025, intending to unlock agentic commerce for consumers and businesses. The program offers an early example of how Mastercard is addressing agentic commerce. The company has developed agentic tokens and standards to support agent-initiated transactions, while also working with merchants to keep the consumer in control of the decision.
According to a PYMNTS Intelligence report, 56% of consumers are willing to let AI shopping agents compare products, and 35% would allow the agents to access their saved payment methods.
Written by urgent.news from PYMNTS's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.