AI-Driven Credit Risk and Decision Intelligence in Digital Fintech Ecosystems A Machine Learning & Financial Analytics using ROC & Shap Algorithm

The rapid expansion of digital fintech ecosystems has transformed credit provision, risk evaluation, and financial decision-making. While artificial intelligence (AI) and machine learning (ML) techniques significantly enhance predictive accuracy in credit risk assessment, their integration into decision intelligence frameworks remains fragmented and insufficiently aligned with regulatory requirements. This study proposes an AI-driven credit risk and decision intelligence framework tailored for digital fintech environments, integrating predictive modeling, explainability mechanisms, and decision-support analytics. Using a policy-realistic synthetic dataset, multiple supervised learning models are evaluated for estimating Probability of Default (PD), Loss Given Default (LGD), and Expected Loss (EL). Experimental results demonstrate that Gradient Boosting Machines achieve superior performance (Accuracy: 95%, AUC: 0.9892). Explainable AI (XAI) techniques enhance interpretability and support governance compliance. The proposed framework bridges the gap between predictive performance and actionable decision intelligence, enabling transparent, auditable, and regulator-aligned fintech operations.