The art and artificial intelligence behind reducing KYC/AML false positives, increasing effectiveness

November 8, 2018

Ninety to 99 percent of alerts generated by traditional rules-based KYC/AML systems will be false positives. ThetaRay’s Jim Heinzman talks to Thomson Reuters about using AI to improve effectiveness of legacy AML systems.  Heinzman said that, in a recent client engagement to reduce false positives, his team at ThetaRay took the firm’s false positives data and, using AI techniques, identified 4,200 anomalies as opposed to the legacy system’s 120,000 with a 99.5 percent false positive rate. There was almost no overlap between the alerts the ThetaRay system and the legacy system generated. The ThetaRay system identified all the true positives the firms’ legacy system had more quickly: on average 70 days earlier, according to Heinzman.