Point Predictive tool automates car loan pricing, fraud protection

The company says its new Point Predictive tool allows autolenders to automatically reject applicants for possible fraud and set interest rates to reflect that risk. Combined with the new ability to process rejected customer inquiries, the new system is federally compliant, according to Point Predictive.

The AutoPass system, announced last summer, increases the pool of loan applications that can be “auto-decided.” This is instantly approved or rejected by the lender’s software without human review. This allows dealers and customers to get faster answers.

Point Predictive’s early Auto Fraud Manager screening system had software that alerted lenders to potential fraudsters, but it was only able to deal with suspicious applicants in a “semi-automated manner,” he said. said Frank McKenna, Chief Fraud Strategist at Point Predictive. Banks must request documents from customers and approve or deny them based on that, he said.

“They can’t completely deny consumers,” McKenna said.

Making decisions based on software modeling would prevent lenders from complying with the law, he said. But the additional work required for compliance means that lenders can waste time and resources on applicants they deem likely to be committing fraud.

McKenna said the new AutoPass tool allows lenders to skip this step, save “considerable” costs and automate the application.

The new system expands Point Predictive’s role in the process when a loan is automatically rejected for fraud risk.

Lenders must tell borrowers why their applications were rejected and customers must be given the opportunity to appeal the decision, McKenna said. In this case, they contacted his Point Predictive, and McKenna said his fraud analysis firm had to invest in infrastructure to receive these customer inquiries.

“We have to be prepared to accept these controversies from our customers,” McKenna said.

McKenna said the new product was not prompted by changes in federal law. Such fraud-related automated decisions were already allowed. Point Predictive decided to pursue this feature after observing growing interest from lenders.

AutoPass may enable automatic determination of up to 80% of applications. That’s double what he’s at some current lenders, McKenna said.

McKenna said Point Predictive did a lot of modeling exercises to ensure compliance and put a lot of effort into developing, training, and validating models for the system. He says the accuracy of the end result is “comparable” to his traditional Auto Fraud Manager fraud scoring model, which was expected to retain a sizeable customer base. Some lenders prefer to continue to check for fraud, but want to retain practices such as policies.

The AutoPass system can also be used by lenders to price loans based on perceived fraud risk and maintain compliance, McKenna said. This is a new service from Point Predictive.

Customers flagged by the system as being at high risk of fraud are likely to be automatically rejected, McKenna said. But lenders may approve edge cases simply by raising interest rates to compensate for potential fraud threats, he said.

McKenna gave an example of how lenders think when misrepresenting small amounts of income is suspected.

“‘I don’t want to go through the whole process of seeing everything. [stipulation] Paperwork. I’m just going to price it,” he said.

This cuts both ways. McKenna expected lenders to cut interest rates for borrowers flagged as lower fraud risk, and said this was a likely pricing scenario as a result of his AutoPass.

“I think most of the profit is actually at the lowest point. [fraud] Score and perhaps more aggressive pricing compared to punitive pricing,” he said.