ARTIFICIAL INTELLIGENCE IN DIGITAL BANKING: TRANSFORMING CUSTOMER EXPERIENCE, RISK MANAGEMENT, AND FINANCIAL DECISION-MAKING

Abstract

Artificial intelligence (AI) is now transforming digital banking in ways that enhance customer engagement and operational efficiency, guide risk-based decision making, and drive predictive financial analytics. But there is not much empirical evidence comparing machine learning models to predict customer subscription. This study examined the role of artificial intelligence in digital banking by evaluating how machine learning models predict customer subscription behavior and support customer-focused financial decision-making. A quantitative secondary-data approach was applied using 45,211 customer observations and 17 variables. The dataset was preprocessed through data quality checks, categorical encoding, and stratified train-test splitting. Three supervised machine learning models—Logistic Regression, Decision Tree, and Random Forest—were evaluated using accuracy, precision, recall, F1 score, confusion matrices, ROC-AUC, and feature importance analysis. The results showed that Random Forest achieved the strongest overall performance, with 90.51% accuracy, 69.01% precision, and an AUC of 0.921. Decision Tree achieved the highest recall at 39.60%, indicating better identification of actual subscribers. Feature importance analysis revealed that duration, previous campaign outcome, month, age, and previous contact timing were the most influential predictors of subscription behavior. The findings indicate that AI-driven predictive analytics can strengthen digital banking by improving customer targeting, supporting personalized engagement, and enhancing data-based financial decision-making. Customer interaction variables were more influential than basic liability indicators, highlighting the importance of behavioral data in intelligent banking systems.