Development of an AI-Based sentiment analysis system for customer interest identification

A financing service company with a digital business unit plays a strategic role in supporting digital transformation, particularly within the telemarketing department. The process of identifying customer interest in product offerings is currently carried out subjectively by the sales team, resulting in inefficiencies when following up with potential customers. This issue is addressed through the development of a sentiment analysis information system based on artificial intelligence to analyze customer conversation transcripts. The system was developed using the Prototyping System Development Methodology and utilizing GPT-4o Mini model, as the core AI model due to its optimal balance of contextual understanding. The implementation and testing results show that the system is capable of identifying customer sentiment with an accuracy of 90% based on the confusion matrix evaluation. This study contributes by integrating a lightweight transformer-based model into a real-world telemarketing information system and assessing its effectiveness using operational data.