Using AI to Lower False-Positive Rates for Fraud Detection Video Transcription
Experience in 2008 and 2009 recession has taught us that fraud increases during recessions. AI based fraud detection software suffers from high false positive rates. It’s just the math. A 90 percent accurate model applied to a claims database with 1 percent fraud, generates 82 percent false positives. You wouldn’t eat food that had an 8 in 10 chance of being bad.
How Halo Based Fraud Detection Works
Halo based fraud detection uses sophisticated methods borrowed from the aerospace industry and it finds outlying claims, people, physical and virtual places, and social networks. Daisy’s Halo based methods significantly reduce the false positive rate and can be executed autonomously, minimizing investigative labor, requiring no data scientists. AI done right delivers verifiable financial outcomes.