CrediWise AI: AI-Powered CIBIL Score System for Macro Finance Businesses

This paper presents CrediWise AI, a novel artificial intelligence-based credit scoring system specifically designed to address the critical challenge of financial inclusion in the microfinance sector. Traditional credit scoring mechanisms systematically exclude Micro, Small, and Medium Enterprises (MSMEs) that lack formal credit histories, creating a significant barrier to economic growth. Our system introduces a unique hybrid approach that combines Gradient Boosting Regressor as the primary model with ensemble validation using Random Forest and XGBoost, achieving 89.5% prediction accuracy. The key innovation lies in our multi-dimensional alternative data integration framework that processes digital footprints, transaction behaviours, GST compliance records, and social media presence through a unified scoring mechanism. The technical implementation employs a microservices architecture using Node.js, Express, MongoDB, and React, enabling real-time credit assessments with response times under 500 milliseconds. Experimental validation on 10,000+ synthetic business profiles demonstrate significant improvements: loan approval rates increased by 35% for previously underserved businesses while maintaining default rates below 3%. The system incorporates comprehensive explainability features and regulatory compliance modules aligned with RBI guidelines and Basel III norms, making it a practical solution for real-world deployment.