Visa Says Traditional Login Checks Fall Short as AI Enters Checkout

Although check usage continues to decline in the United States, check fraud remains a major area of financial crime. Visa said check fraud accounted for 30% of U.S. fraud losses in 2024. More significantly, the methods used in such crimes are crossing from paper to digital channels, threatening faster-payment systems and banks’ risk controls.

Visa recently warned that anomalies involving accounts, identities and transactions linked to check fraud have spread to faster digital payments. The company is deploying AI-powered image forensics and behavioral analysis to help banks identify altered checks and suspicious cross-channel transactions. The report did not disclose the actual amount lost or the date when the system became operational.

Visa is expanding beyond its role as a credit and debit card transaction network into value-added merchant services, focusing on agentic AI, stablecoins and digital assets. The shift could determine whether Visa can reduce its reliance on card-processing fees as payment methods rapidly evolve, while generating revenue from fraud prevention, data analytics and new settlement tools.

Visa executives recently said the company is advancing agentic commerce, in which AI agents can independently search, shop and make payments. It is also using AI to detect and block payment fraud such as romance scams while developing stablecoin applications. Reports did not provide investment figures, a formal launch date or projected new revenue, and Visa has yet to disclose a specific timetable.

Financial scammers are shifting their focus from breaching systems to exploiting victims' trust, fear and sense of urgency. Even as banks and payment providers deploy AI to detect unusual transactions, criminals can still use impersonation and social engineering to persuade victims to authorize payments themselves. Visa says the trend matters because traditional models struggle to determine in real time when seemingly legitimate activity is fraudulent.

Visa recently said AI can serve only as one layer of defense against the new wave of scams and must be combined with transaction-data analysis, intelligence sharing across institutions and expert human investigations. As of July 20, 2026, the reports provided did not disclose the date of a specific case, the amount involved or a prevention rate. The key development is Visa's explicit position that fraud-prevention systems cannot rely solely on automated models.

Payment fraud has evolved from isolated offenses into an organized business with specialized roles and the ability to operate at scale, posing an ongoing threat to global payment networks such as Visa. Blocking suspicious transactions without wrongly declining legitimate payments is critical to consumer trust, merchant operations and financial institutions’ risk costs.

James Mirfin, Visa’s global head of risk and identity solutions, said the company is embedding artificial intelligence throughout the payment process to raise the cost of committing fraud and reduce false positives without adding friction to transactions. The latest report did not disclose a deployment date, investment amount or performance figures. For now, the focus remains on Visa’s broader fraud-prevention strategy.

Visa said in its Spring 2026 threat report that generative AI has sharply reduced the cost and technical barriers involved in creating phishing messages, impersonating identities and automating attacks. This allows criminal groups to target more consumers simultaneously. The focus of payment defenses is also shifting from preventing system breaches to identifying transactions made by real people who have been manipulated, creating new challenges for banks, merchants and card networks.

The latest report shows that both attackers and defenders are expanding their use of AI. Fraudsters use it to rapidly generate convincing content and test attacks at scale, while financial institutions deploy AI to strengthen real-time monitoring and anomaly detection. Visa warned that the key shift in Spring 2026 was not an increase in the value of individual fraud cases, but a simultaneous rise in the speed, volume and personalization of attacks, making them harder for traditional rules-based fraud controls to stop in time.

Generative AI is reshaping how consumers discover products, compare prices and narrow their choices, with large language models increasingly serving as shopping assistants. Visa Chief Executive Ryan McInerney said that shift shows AI-assisted commerce has arrived. But allowing autonomous agents to spend on a customer’s behalf presents a tougher challenge, requiring trusted authorization and safeguards that satisfy consumers, merchants and payment networks.

McInerney said AI can already guide purchase decisions, but consumers generally have not handed agents authority to complete transactions. Shoppers still return to a merchant’s website to confirm the order and pay manually. Visa disclosed no launch date, transaction target or spending amount for broad agentic-payment adoption, underscoring that payment infrastructure, accountability and risk controls must be developed further before autonomous checkout becomes mainstream.