The model helps lenders increase revenue by identifying viable loan applications that have been wrongly rejected by the banks’ automated credit risk assessments.

The problem
Credit scoring models are used during the loan approval process to predict the likelihood of borrower delinquency. Online lending banks receive thousands of loan applications per day, and use automated assessments to approve or reject them in real time. Because these models are based on limited applicant data and subject to strict discretionary measures to keep risk levels manageable, a significant number of applications are usually rejected.

ThetaRay’s solution
When integrated into a bank’s automated assessment, ThetaRay’s multi-source hyper-dimensional detection system automatically identifies the potential customers with relatively low credit risk while collecting and analyzing additional data on each rejected loan applicant, including credit score, historical loan performance and personal information from government databases. These insights empower lenders to convert many rejected loan requests into approved loans while maintaining acceptable risk levels.

“This system is based on the same sophisticated algorithms that enable our fraud detection solution to so effectively catch the bad guys; they are now proving equally adept at identifying the good guys,” said ThetaRay CEO Mark Gazit. “We have built our reputation on our ability to produce solutions detecting risks and threats that companies wish to avoid. But with our new solution, we can help them discover untapped business opportunities as well.”

ThetaRay recently tested the model with a large online lending bank. After examining massive amounts of rejected loan applications over the course of a week and a half, ThetaRay determined that more than 30% had been excluded due to the limitation of the existing tools. The lender is now implementing the ThetaRay model on a long-term basis in order to increase its approved loan rates, grow its client base and boost revenues.