Westlake Financial has enlisted the services of PointPredictive, an artificial intelligence-enabled risk management and anti-fraud solution. Westlake will use the scoring solution – called Auto Fraud Manager – to identify misrepresentation and prevent delinquency on high-risk applications.
The risk engine is based on an algorithm that is trained with each new lender’s data set. More than 100,000 dealers are represented in the company’s consortium of dealer data.
“Lenders send us two years of their historical loan and application data,” said Tim Grace, PointPredictive CEO. “With that data, they know how many of those loans performed and how many defaulted. We score those applications with our scoring engine – the higher the score, the higher the risk. We can show them where all their losses are going to come from.”
PointPredictive’s tool is also helping lenders fight fraud.
“Analyzing all of the losses from default, typically the only things that are marked are factors that can be associated with identity theft,” Grace said. “Identify theft is only 10% to 15% of that loss bucket. The rest – income, employment, collateral misrepresentation, all the misrepresentations and fraud types – those aren’t usually labeled as fraud. They end up being classified as first-payment default or early-payment default. That’s where we look to stop fraud.”
For Westlake, the solution was selected based on results of a retrospective test and live pilot. “In our test with PointPredictive, we saw an added lift in identifying defaults tied to misrepresentation on top of our existing fraud scores and tools,” said Westlake Group President Ian Anderson in a prepared statement.
So far, about 20 auto lenders are using or testing the solution, with another 40 slated by yearend, Grace said.