LAS VEGAS — Unit21, a global leader in risk and compliance technology, announced at Money20/20 the launch of its AI Rule Recommendation Agent, a new capability designed to combine machine learning precision with the transparency of rule-based systems to improve fraud and anti-money laundering (AML) detection.
The AI Rule Recommendation Agent aims to address a long-standing challenge for fraud and AML professionals, who have traditionally chosen between adaptable but opaque machine learning models and explainable but static rule-based systems. Unit21’s new offering seeks to eliminate this trade-off by integrating the learning capacity of AI with the clarity and control of human-defined logic.
According to Unit21, the system’s AI Rules Engine analyzes patterns derived from alerts, outcomes, and contextual data. This analysis allows it to automatically suggest optimized rule logic, which can help teams reduce false positives, identify previously missed signals, and decrease the need for manual tuning. The recommendations can be tested using Unit21’s shadow mode feature against historical and simulated future data. Once validated, new rules can be deployed immediately with measurable confidence.
Kunal Datta, Head of Product at Unit21, commented on the development, stating, “Traditionally, teams had to choose between speed and safety. With AI Rule Recommendations, they have the best of both. The AI finds a better rule, tests it silently in the background, and proves it works before deploying it to the analyst. It’s a no-brainer.” He added, “This bridges the long-standing gap between ML models and rules. You can see why the AI made a recommendation, measure its impact, and push it live, all with human control intact.”
The AI Rule Recommendation Agent is the latest addition to Unit21’s AI suite, which also includes AI Agents capable of autonomously reviewing, summarizing, and investigating alerts. Together, these tools form a self-improving feedback loop: one AI component identifies, writes, and validates optimized rules, while another processes the resulting alerts, providing outcomes that continuously strengthen the overall system. This integration is intended to enhance both the quality of detection and the efficiency of investigations, enabling risk and compliance teams to operate with increased precision.
Datta further elaborated on the multi-faceted impact of the technology: “It’s AI working at two levels. At the system level, it keeps your detection logic sharp. At the operational level, it helps your analysts make faster, better decisions. That’s how you truly turn fraud and AML from reactive to proactive.” By combining automation with explainability, Unit21 aims to help financial institutions transition from reactive monitoring to proactive intelligence, identifying risks earlier, continuously refining rules, and scaling operations securely.
Unit21, which has raised nearly $100 million from investors including Google and Tiger Global, focuses on developing solutions that identify and mitigate risks associated with money laundering, fraud, and other illicit financial activities.
