Cleafy, an Italy-based platform for fraud detection and response in banking, has introduced Nyx, an AI-driven investigation tool tailored for financial institutions. This new solution tackles a critical operational challenge: the inability to promptly investigate all fraud signals as they arise.

Reimagining Fraud Investigation

Despite advancements in fraud detection, the investigative process remains largely manual and time-consuming. As a result, many financial institutions only manage a fraction of the alerts generated by their systems, leading to delays that can inadvertently allow coordinated attack campaigns to grow unchecked.

Nyx automates the investigation of every fraud signal, reconstructing complete attack scenarios, and producing detailed forensic reports. According to Cleafy, this process is completed in less than four minutes, significantly faster than manual processing which can take over four hours.

Moreover, Nyx is proactive; it identifies potential coordinated attacks before fraudulent transactions occur, offering a predictive analysis capability that has already detected attack patterns up to 15 days ahead of the first confirmed case at some institutions.

Automation and Regulatory Compliance

For each pattern it detects, Nyx creates optimized detection and response rules, refining an institution’s existing rule set. This includes removing unnecessary rules, resolving conflicts, and enhancing coverage. Analysts monitor the system and provide feedback that is integrated into continuous learning to improve AI agents’ performance.

Nyx supports regulated environments by incorporating full logging and traceability in line with DORA and NIS2 directives, ensuring compliance with EU cybersecurity requirements.

Currently, the platform is operational at three European banks serving over 30 million digital banking customers. Cleafy’s CEO emphasized that as fraud increasingly becomes an industrialized operation driven by AI, Nyx aims to bridge the gap in investigative capabilities.