Subprime lender Prestige Financial Services is using artificial intelligence-powered credit scoring model ZestFinance to grow its origination volume without increasing losses or taking on additional risk, Steve Warnick, chief credit and analytics officer at Prestige Financial, told Auto Finance News.
“During the 2015-2016 timeframe, we were dealing with escalating losses and trying to figure out how to resolve that without totally cutting out volume,” Warnick said. Since integrating ZestFinance, the lender has achieved a 36% increase in new applicants, resulting in a 14% higher approval rate.
Within six months of deploying the AI-based credit underwriting model, Prestige doubled its lending volume. Warnick declined to disclose specific portfolio data since the new originations “haven’t been on the books” long enough to determine their current losses, he said. However, applying machine learning has enabled Prestige to rank-order its borrowers.
The lender then swaps out risky borrowers and replaces them with thin-file and new-to-credit consumers who are more creditworthy than traditional credit scores might suggest. While the credit metrics used remain the same as traditional scorecards, the difference lies with the vast number of variables that AI applies to make decisions on borrowers.
“[Prestige] was able to go from using 10 to 25 variables in a traditional model to 2,700 using machine learning,” Warnick said. To Prestige’s surprise, its dealer partners responded favorably to the new AI-based credit scoring. The challenge for them, though, has been adapting their practices to account for the machine learning.
To help, Prestige has external and internal representatives to provide training. “In the event we don’t have a rep on-site, we go out of our way to reach out to dealerships to see if they have questions,” Warnick said. Additionally, the lender has an underwriting group that can help dealers vet loan applications.
