Tricolor Auto Acceptance has seen a lift in origination capabilities since integrating an artificial intelligence algorithm into its underwriting model last month. Chief Executive Daniel Chu expects the algorithm, which hones the company’s customer segmentation abilities, will contribute about 3% of the anticipated 25% origination volume increase expected for 2019. That 3% growth translates to about 900 loans, he added.
The algorithm, which was tested for eight months before it launched, effectively “sits on top” of the current risk model and provides additional granularity, Chu said. The AI model allows Tricolor to identify the upper and lower tiers of its six credit buckets and make incremental changes in the financing products it offers.
“For instance, a higher graded C customer might be offered better rates or a lower down payment requirement than a lower grade C customer, despite both customers being graded as a C,” Chu said. “That cleaner segmentation will ultimately result in incrementally better conversion. It will be more attractive to the customer, so more customers will take us up on that offer.”
Though Chu is “pretty confident” the AI-based originations will perform well, the company is treading carefully to ensure it doesn’t “turn a nob too much,” he said. Tricolor will continue to test and tweak the algorithm to make sure the loans are performing as expected, he said.
