Visa says AI payment fraud defense now hinges on third parties
Visa says stronger AI payment fraud defense is now urgent as third-party incidents jumped 62% in the first half. Visa Scam Disruption flagged more than $1.4 billion in potential fraud, up 40%, Visa said.
AI is not inventing new fraud, only making old fraud faster.
Global payment account enumeration rose 35% year over year, Visa said. Its Risk Operations Center blocked 200% more enumeration transactions, helping cut losses from enumerated credentials by 21.4%.
Third-party agents were linked to about 56% of compromised payment credentials identified in the period, Visa said. Recovered credentials from breaches rose more than 1,200%, a sign one intrusion can expose many downstream sellers.
Paul Fabara, Visa chief risk and client services officer, framed the shift as economics, saying Visa is fighting bad AI with good AI across network intelligence and human review. Ransomware cases rose 19% in six months and 312% since 2021, with the 10 most active groups accounting for more than 80% of incidents Visa observed.
Agent checkout widens this surface. OpenAI put agent checkout behind Face ID after pushing bots closer to payment credentials. StartupHub.ai data tracks OpenAI's checkout push as agentic-commerce infrastructure that now sits inside Visa's fraud perimeter, not outside it. Risk advisers at MyInsurect said AI-enabled fraud has changed how firms must protect payments, employees, vendors and insurance programs.
Visa shows volume blocked, but not false declines or merchant friction when controls tighten.
Michael Jabbara, Visa global head of payment ecosystem risk and control, said the fight will be won upstream before transactions emit a clear fraud signal. That leaves travel sellers, event ticketing and holiday storefronts carrying the test, with Visa warning spoofed booking sites now use real reservation details to win extra payments.





