Assessment of customer acquisition and closure in banking transactions with artificial intelligence

Customer acquisition and churn have become strategically critical in the banking sector, especially in the digital age where competition is intense and customer expectations are rapidly evolving. The increasing demand for personalized services and the acceleration of digital transformation have compelled financial institutions to manage customer relations more effectively. This thesis analyzes customer acquisition and churn processes in banking through artificial intelligence (AI)-based models. The analysis of the datasets was conducted using various machine learning techniques, including Artificial Neural Networks (ANN), Support Vector Machines (SVM), Decision Trees, and Natural Language Processing (NLP). Model performance was assessed through metrics such as accuracy, precision, recall, and F1-score. The outcomes revealed that AI-driven approaches are highly effective in forecasting customer behavior. Among the evaluated models, the ANN method delivered the best results, attaining an accuracy of 92.1%. Furthermore, applying sentiment analysis to social media content enabled early identification of negative customer sentiment, supporting risk evaluation based on customer segments. These insights suggest that artificial intelligence–based decision support systems offer substantial benefits in reducing customer attrition, designing personalized marketing efforts, and shaping digital banking strategies. This thesis highlights the potential of AI applications in banking customer analytics and aims to provide data-driven solutions that offer a competitive advantage to institutions.