Fraud and Disputes Rank as a Top Cost for 42% of Issuers

That question is becoming more pressing as artificial intelligence (AI) systems begin to recommend products, select payment methods and potentially complete transactions for consumers. It also helps explain an important finding in “The Issuer Risk Playbook,” a PYMNTS Intelligence report produced in collaboration with Visa Digital Processing Solutions. The research surveyed 500 executives in head-of-payment roles at U.S. bank and non-bank card issuers.

Fraud is already expensive enough without that added complexity. Forty-two percent of issuers rank fraud and disputes as either their largest or second-largest platform-related operating cost, excluding employees.

At the same time, shortcomings in fraud systems are tangible. The share of issuers citing insufficient fraud systems and high false declines as a processor pain point increased to 21% from 14% a year earlier, while cybersecurity and network concerns rose to 36% from 30%.

Real-Time Detection Gets a Bigger Job

Real-time fraud detection itself is no longer much of a dividing line between issuers. Roughly six in 10 issuers across customer lifetime value (CLTV) tiers are deploying or enhancing AI-powered card fraud detection and prevention, suggesting that the capability is becoming a basic requirement of modern issuing.

Sixty-eight percent of high-CLTV issuers say enhanced security and fraud prevention are necessary for agentic commerce, compared with 47% of medium-CLTV issuers and 43% of low-CLTV issuers. In other words, high-CLTV issuers are roughly 1.5 times as likely as the other groups collectively to make the connection between security and commerce conducted by AI agents.

Fraud controls therefore have to evaluate more than whether a transaction resembles the cardholder’s previous behavior. Issuers will need sufficient information to determine whether an automated purchase falls within the authority the customer granted, while still making the decision quickly enough to avoid turning additional security into additional checkout friction.

Real-time detection offers an advantage precisely because the decision can be made when authorization occurs. Better decisioning can help issuers stop suspicious activity without unnecessarily rejecting legitimate transactions, an increasingly important balance when false declines themselves carry operating and relationship costs. The report notes that high-CLTV issuers are concentrating near-term AI investment on real-time fraud detection and operational automation, with 41% identifying each as an implementation priority during the next 12 months.

More precise authorization can preserve legitimate payment volume, reduce the expense associated with investigations and disputes, and limit the customer frustration caused by false positives. It can also give issuers a stronger foundation for transactions in which software acts under a consumer’s delegated authority.

That makes real-time fraud detection a capability that issuers increasingly need whether agentic commerce develops rapidly or slowly. If AI agents become meaningful payment initiators, issuers will need controls capable of evaluating those transactions at authorization. If they do not, issuers still have the existing problem of stopping more sophisticated fraud without stopping good customers along with it.

At PYMNTS Intelligence, we work with businesses to uncover insights that fuel intelligent, data-driven discussions on changing customer expectations, a more connected economy and the strategic shifts necessary to achieve outcomes. With rigorous research methodologies and unwavering commitment to objective quality, we offer trusted data to grow your business. As our partner, you’ll have access to our diverse team of PhDs, researchers, data analysts, number crunchers, subject matter veterans and editorial experts.