AML Analytics Qorus-Infosys Finacle Banking Innovation Awards 2025 - Winner

Concept and Objectives The AML Analytics & Investigation project was designed to revolutionize how our bank detects and investigates potential money laundering activities. The core objective was to transition from rule-based alert systems to a data-driven, AI-powered framework that improves precision, reduces manual workload, and accelerates case resolution. By integrating machine learning, graph analytics, and Generative AI (GenAI) with existing transactional and customer datasets, the platform aims to increase the efficiency and accuracy of AML processes while ensuring regulatory compliance and supporting proactive risk management.

Reasons Behind the Project Traditional AML systems generate a high volume of false positives, straining investigative resources and delaying the identification of genuinely suspicious activity. We identified the need for a smarter system that could detect evolving laundering behaviors, explain anomalies, and prioritize alerts effectively. Additionally, growing regulatory scrutiny and an increase in complex financial crime patterns made it imperative to modernize our investigative capabilities.

State of Competition While many financial institutions are exploring AI in AML, few have achieved true integration of supervised and unsupervised models, graph-based analysis, and GenAI for case explanation within a single framework. Our ability to connect these layers—alongside customer 360° integration—has positioned us ahead of the curve. Early benchmarking shows a 20x uplift in alert accuracy compared to traditional systems, setting a new internal and industry standard for AML effectiveness.

Sources of Inspiration The project was inspired by global innovation in fraud detection, cybersecurity threat modeling, and social network analysis. Externally, we examined how tech-forward banks and fintechs use graph databases and anomaly detection for fraud prevention. Internally, we drew from prior success in marketing analytics and customer segmentation projects that demonstrated the power of integrated data and advanced modeling.

Departments Involved This was a cross-functional initiative involving: • Compliance and AML teams, who defined key indicators and oversaw regulatory alignment. • Data Science & Analytics, who developed the ML models and anomaly detection algorithms and for building the data pipelines, graph databases, and Power BI layers. • Internal Audit, who reviewed the model governance and assurance processes.

Main Results So Far The platform has already delivered strong results: • 30% reduction in investigation effort, due to automation and unified customer 360° views. • 20x increase in alert accuracy, with significantly more alerts resulting in actual SAR filings. • 13% reduction in manual work, thanks to semi-automated alert closure and workflow redesign. • Real-time, explainable insights, powered by GenAI, supporting better investigator decision-making. • Improved case prioritization, driven by decile scoring and anomaly confidence levels. This project has become a cornerstone of our financial crime strategy and a model for future AI adoption across the bank.