Autonomous Compliance Systems: AI, Event Streaming, and the Future of Financial Crime Prevention

Abstract

Autonomous compliance systems are systems that tackle (i) the ever-increasing volume of supervision data available but unsupervised within financial institutions and (ii) the operational burden of rule-based established systems, specifically those employed for transaction monitoring and associated event management processes. Such systems support ongoing monitoring and risk-scoring of transactions, entities, persons, and interactions and augment the functionality of conventional transaction monitoring systems. The incorporation of Artificial Intelligence (AI) offers automation and, if properly deployed, a continuous improvement loop by feeding back learnings into recommender systems. The use of high-throughput, low-latency data pipelines ensures the availability of decision-ready data for these systems while their design provides a mechanism for constant validation of the AI ML models. The autonomous compliance systems address important operational concerns, especially in the areas of data quality, privacy, and security, and establish additional forensic capabilities that support compliance liability. Research opportunities associated with the integration of AI ML into the compliance ecosystem are highlighted, and the discourse concludes with a succinct exposition of AI ML-enhanced autonomous compliance systems that advance the Financial Crime Prevention domain.