Tool gives lenders look into dealership loan approvals

Artificial intelligence and data provider Point Predictive Inc. has gone into the matchmaking business.

The company recently unveiled a product called DealerExplorer that helps lenders decide if they want to work with a certain dealership.

The new platform gives lenders insight into the potential fraud risk of a dealership's customers before they decide to fund a loan. It shows lenders the quality and quantity of loans a dealership originates by flagging defaulted loans and fraudulent applications.

Frank McKenna, co-founder and chief fraud strategist for the company, told Automotive News the current system of lenders' sales representatives signing up dealerships is "a blind process."

"They don't know who has a good history and who doesn't," McKenna said. "We give them a portal to see how [dealership customers'] apps are scoring, number of defaults, volume, risk insights so they can determine if they're going to sign up that dealer or not."

Point Predictive has more than 50 lender clients using its other fraud-detection products, but McKenna said this is the company's first offering giving lenders a view into dealerships not already in their portfolio. Five of its existing lender clients are piloting DealerExplorer, paying on a per-report basis.

McKenna and COO Tim Grace co-founded Point Predictive in 2017 after working for artificial intelligence firms that built technology detecting fraud at credit card companies in the late 1980s and 1990s, and at mortgage companies in the mid-2000s.

Point Predictive's data consortium has information on more than 126 million loan applications totaling $2.4 trillion from U.S. dealerships and lenders, McKenna said. Lenders in the consortium include automakers' captive finance companies, banks, subprime lenders and credit unions.

McKenna estimates annual fraud in the U.S. automotive market has increased about 260 percent to $8 billion since the pandemic, mainly because of more people going online and learning how to commit fraud. The fraud includes consumers lying about their income, employment or identity.

"We think probably 700,000 new fraudsters were radicalized during the stimulus," McKenna said. "That's why we're seeing this dramatic shift in what dealers and lenders are reporting. We're aiming to help dealers and lenders get in front of that dramatic rise. The only way to target this is a strong counterattack, using data and all the tools you have to really address the risk."

In 2024, Point Predictive plans to compile the data it's providing lenders and start selling it to dealerships on an aggregated basis so they can compare their ability to spot fraudulent borrowers against the broader market and benchmark their overall credit risk profile. This will let dealers evaluate their own performance in the consortium and see how lenders perceive their risk.

"A lot of dealers sometimes get a pushback on a loan because they have fraud, but they don't get a statistical view," McKenna said. "Dealers we're talking to are starved for this information. They really want to have some knowledge of their stats."