The dataset of 200,000 customer records with 200 anonymized numerical features is applied to the Santander Customer Transaction Dataset, and its objective is to predict customer transaction behavior and future transaction likelihoods. Five classification models—Logistic Regression, Decision Trees, Random Forest, Gradient Boosting, and XGBoost—are evaluated for their ability to classify who are likely to do specific transactions. In addition, two regression models are trained to predict the transaction values using Linear Regression and Random Forest Regression. For classification tasks, it uses standard metrics such as accuracy, precision, recall, and F1 score, while for regression tasks, it examines MAE, MSE, and R². Logistic Regression proved to be the most accurate classification model; it achieved an excellent accuracy of 90.00, and it was found to provide actionable insights for financial organizations in terms of customer risk assessment and revenue prediction models.