The rapid adoption of real-time payment systems has introduced a critical challenge for financial institutions: the irreversibility of transactions. Once funds clear over instant rails, they cannot be recalled, effectively eliminating the window for recovery. This fundamental shift has led to a compounding problem, with PYMNTS Intelligence reporting that 40% of financial institutions experienced greater financial losses to fraud last year, while 38% contended with higher fraud volumes.
A significant driver of this escalating trend is authorized push payment (APP) fraud, where victims unwittingly authorize payments to fraudsters. PYMNTS Intelligence data reveals that scams now constitute 23% of all fraudulent transactions reported by financial institutions, marking a substantial 56% year-over-year increase. The financial impact is even more stark, with the share of dollars lost to scams surging by 121%. In the U.K. alone, APP fraud losses climbed 19% last year to 576.4 million pounds (approximately $774 million), with a concerning 66% of these cases originating on online platforms.
Agentic AI Reimagines Post-Clearance Fraud Investigation
Historically, fraud detection AI primarily focused on a single question: should this transaction be approved? The new wave of AI deployment shifts this paradigm, asking a different, more complex question: given that a transaction has cleared, what does it connect to, and where did the money go? This strategic pivot is exemplified by Nasdaq Verafin’s expansion of its Agentic AI Workforce.
Nasdaq Verafin has introduced two new role-based agents: an Agentic Fraud Analyst and an Agentic AML Analyst. These agents are engineered to automate the intricate investigative work traditionally performed manually by fraud and compliance teams. The Agentic Fraud Analyst will initially concentrate on triaging unusual ACH activity, while the Agentic AML Analyst will tackle cash structuring alerts—cases where criminals fragment large sums into smaller deposits to evade regulatory reporting thresholds. The AML agent’s capabilities are slated to expand to include flow-of-funds analysis and the detection of unusual international transactions. Both new agents are expected to reach general availability in the third quarter of 2026.
The efficacy of Nasdaq Verafin’s platform is bolstered by its extensive reach; more than 650 financial institutions have already adopted it, and it operates on a consortium data network encompassing over 2,800 institutions. This vast network enables the system to identify counterparty fraud risk across multiple institutions, rather than being confined to a single bank’s ecosystem. The company has reported significant workload reductions from its existing agents, with the Agentic Sanctions Analyst cutting alert review time by up to 90% and the Agentic EDD Analyst reducing enhanced due diligence review time by up to 50%.
Global Infrastructure to Combat Cross-Border Financial Crime
The challenge of real-time payment irreversibility is driving the development of sophisticated AI fraud infrastructure globally. In India, the Reserve Bank Innovation Hub, an arm of the Reserve Bank of India, launched MuleHunter.AI. This AI system is now operational across 26 banks and is designed to detect approximately 20,000 mule accounts each month. Mule accounts are critical to criminal operations, serving as intermediary conduits through which stolen funds are routed across multiple banks before being withdrawn.
Data from the Indian Cyber Crime Coordination Centre, reported by The420, underscores the immense scale of this problem. As of December 31, the agency had identified 2.65 million first-layer mule accounts utilized by cybercriminals to move stolen funds. Authorities estimate that these networks facilitated the theft of nearly 200 billion rupees (about $2.4 billion), of which approximately 81.9 billion rupees (roughly $980 million) has been successfully recovered and returned to victims.
While the focus is increasingly on post-clearance recovery, preventing fraud before it occurs remains a primary defense. JPMorgan Chase and ACI Worldwide have announced a partnership to integrate JPMorgan’s Kinexys Liink account verification directly into ACI Worldwide’s enterprise fraud platform. This collaboration aims to apply consistent controls across various payment rails before funds are disbursed. PYMNTS has previously reported that the speed of faster payment rails has rendered post-settlement recovery largely impractical, challenging the traditional assumption that finance teams would have ample time to rectify errors after money had moved.
The current advancements in AI represent a crucial evolution in financial crime fighting. While blocking fraudulent transactions before they clear remains the first line of defense, banks are now building parallel AI systems designed to reconstruct complex transaction patterns, connect related activities across diverse institutions, and meticulously trace stolen funds before criminals can withdraw them. This proactive, data-driven approach is essential in an era where the speed of money movement demands an equally rapid and intelligent response to illicit activity.
