
PointPredictive, a developer of anti-fraud software, has launched a program that tells lenders not only which loan applications are risky, but why they’re risky.
Auto Fraud Alert, which launched Tuesday, exposes potential fraud attributes directly to the lender, Chief Strategist Frank McKenna said. “A lot of times lenders were in the dark about what kind of information they were getting,” McKenna said. “They know some borrowers are actively trying to work the system, so if there’s something on your application that’s significantly different than what another lender got, we can fill in those gaps and say, ‘Hey, there’s been prior fraud here, and it looks like there’s a huge discrepancy in the information you’re getting.’ That’s the gap this product is filling.”
With the new product, PointPredictive is aiming to expose, at a granular detail, information about what makes a loan risky by providing a complete view of the application and history. That information “can be a discrepancy in what’s provided to you by the dealer and another lender, a change in income or employment, or if the Social Security Number has been used by multiple borrowers,” McKenna said.
Read more: Veros Credit Turns to Machine-Learning Fraud Solution
Auto Fraud Alert has more than 100 new alerts and red flag indicators that are pulled from PointPredictive’s auto fraud database of 65 million applications submitted by 70,000 dealerships. The company takes that information, analyzes the trends, and provides that information back to the lender in the form of an online report and PDF file.
So far, there are “a couple” of lenders who are piloting the product, and PointPredictive aims to have 10 to 15 lenders using the solution by the end of September. “We need to get more people using this shared intelligence approach, because we really believe it’s one of the predominant ways to address fraud,” he said.
For more content like this, check out our upcoming event Auto Finance Accelerate, May 13-16 at the Omni San Diego. Visit www.AutoFinanceAccelerate.com to register.