CRIF and UNIMORE Partner on AI Explainability for Credit Risk

CRIF has collaborated with the Data and Text Analytics Lab at the University of Modena and Reggio Emilia (UNIMORE) to improve the interpretability of AI-driven credit risk models. The research focuses on implementing counterfactual explanations, a technical approach that identifies specific changes in input data required to shift an automated outcome, such as turning a loan decline into an approval.

This functionality has been integrated into CRIF’s ORCHESTRA platform. By moving beyond traditional model outputs, the system provides actionable guidance that helps financial institutions understand model behavior and communicate decisions more clearly to applicants. The initiative is designed to address the growing need for accountability and transparency in automated decision-making processes.

As regulatory scrutiny regarding algorithmic fairness increases, the project aims to bridge the gap between complex analytical models and operational requirements. The collaboration enables lenders to manage model risk more effectively while maintaining the necessary human oversight and governance frameworks required for financial services.