TRANSFORMING PRODUCT INNOVATION TO MEET CUSTOMER NEEDS THROUGH AI MARKETING, A CUSTOMER FEEDBACK ANALYSIS WITH GPT-4O MINI

Abstract

This research uses GPT to conduct sentiment analysis on customer reviews for biodegradable products. Sentiment analysis uses 4 categories, namely positive, negative, neutral and mixed, then for product improvement this study focuses on negative sentiment by adding negative sentiments from the mixed sentiments. Data collected from Amazon reviews regarding one brand of biodegradable trash bag in several stores. The GPT-4o Mini model was then used to categorize sentiment.The results of sentiment analysis show that most reviews are positive, but there are also many negative sentiments regarding product durability, leakage and price. The model used is able to accurately identify and extract negative sentiment even from a mixed sentiment, thereby providing a more complete understanding of customer dissatisfaction. This research emphasizes the importance of integrating AI-driver sentiment analysis into the marketing process.