Using AI to Fight Increasingly Sophisticated Fraud: How Automation Levels and Process Redesign Drive Results in Financial Services and Payments
November 2026
Manager
Ryoichi Tanaka
With the diversification of payment methods and the expansion of digital transactions, the volume of transactions that risk management must handle continues to grow. At the same time, fraud is becoming more sophisticated in quality, as tactics grow more cunning, such as impersonation that abuses generative AI. In a manpower-centered operating model that keeps piling on new rules and manual review, rising alert volumes delay decision-making, and the limits are close at hand. This article examines “clarifying the level of automation” and “redesigning operations and controls” as the key points that determine whether AI adoption succeeds.
Introduction
Against the backdrop of diversifying payment services and expanding digital transactions, fraud tactics are growing more sophisticated at a rapid pace. Attack methods such as phishing, account takeover (ATO), and impersonation through SIM swapping and deepfakes have evolved to the point of exploiting machine learning and generative AI, and the limits of conventional rule-based detection models are becoming a reality. In this changing environment, risk management departments at payment providers are required to meet two imperatives at once: improving fraud detection accuracy and making operations more efficient. One promising means that is attracting attention is…
To read the full text, please download the PDF.
DownloadPDF(1.91 MB)