R-G Federal Credit Union cuts indirect auto loan losses 33% with Scienaptic AI credit decisioning

R-G Federal Credit Union, chartered in 1957 to serve military and civilian personnel at Richards-Gebaur Air Force Base, has since expanded its membership across twelve Missouri counties—including Jackson, Cass, Johnson, and Bates. As membership grew and consumer lending volumes increased, the credit union's manual underwriting processes struggled to keep pace. Indirect auto lending in particular carried elevated loss exposure, and inconsistent decision-making across applications created both risk management gaps and compliance concerns. The absence of scalable, data-driven decisioning was limiting R-G FCU's ability to serve a diverse membership equitably and efficiently while containing portfolio losses.

R-G FCU deployed Scienaptic AI's credit decisioning platform—founded in 2014 and purpose-built for financial inclusion—to automate underwriting across its consumer lending portfolios. The platform applies advanced machine learning algorithms trained on broader data sets than traditional credit bureau pulls, enabling it to evaluate creditworthiness for underbanked and underserved applicants who may be invisible to conventional scoring models. Scienaptic's system integrates directly into the credit union's loan origination workflow, delivering real-time automated decisions while embedding rigorous fair lending and regulatory compliance monitoring at each decisioning layer. The deployment went live in March 2026, covering consumer lending including the indirect auto channel. Scienaptic's iCUE framework—which layers large language model capabilities onto predictive credit scoring—provides loan officers with explainable outputs, keeping human oversight in the loop for edge cases.

Following go-live, R-G FCU achieved measurable improvements across both loss performance and operational throughput:

• 33% reduction in losses on indirect auto loans—the portfolio segment with the highest prior loss exposure
• Up to 68% of consumer lending decisions now handled automatically, reducing manual review queues and accelerating time-to-approval
• Faster loan approvals across the membership footprint, strengthening the credit union's competitive position against regional banks and online lenders
• Compliance and fair lending standards maintained throughout, with no degradation in regulatory posture

President and CEO Steve Deere noted the platform enabled greater speed and consistency without compromising prudent risk management—a balance that had previously required substantial manual effort.

• AI credit decisioning can reduce indirect auto loan losses meaningfully without tightening approval rates—automation and risk discipline are complementary when the model is properly trained.
• Credit unions with community or military heritage can adopt AI underwriting while preserving member-first values; fairness monitoring built into the platform is essential to maintaining that trust.
• Automation rates above 60% are achievable for smaller credit unions, but require platforms that integrate into existing origination workflows rather than replacing them.
• Explainability features matter for credit union loan officers—staff adoption depends on understanding why a decision was made, not just what it was.
• Vendor CUSO structures (as with Scienaptic's credit union service organization) offer community lenders strategic alignment and shared governance that pure SaaS relationships do not.