Case Study: Lending company boosts collections efficiency and cuts prediction time 220x with DataRobot Optimizer App

Optimizing Loan Predictions with DataRobot AI Apps

Lending Company, a US fintech that provides small consumer and point-of-sale loans, was struggling to collect on tens of thousands of delinquent accounts where low Right Party Contact (RPC) rates limited recovery. The team had built models on the DataRobot Enterprise AI platform but needed a scalable way to predict the best time of day to call each borrower; they partnered with DataRobot to apply a production-ready solution using the DataRobot Optimizer App.

Working with DataRobot’s AI Apps team, Lending Company implemented an Optimizer App that uses DataRobot models and batch scoring to predict optimal call times and push prioritized lists to the collections autodialer. DataRobot’s team rewrote the optimization algorithm for scale (a 220x throughput improvement), cutting run time from roughly three hours to about 15–20 minutes, freeing data-science resources and boosting connection rates by 0.1–0.2% on a 3–4% baseline — a measurable lift with meaningful revenue impact.