<![CDATA[The development of the digital economy requires Micro, Small, and Medium Enterprises (MSMEs) to utilize transaction data as a basis for business decision-making. However, limited digital literacy and technical capabilities have prevented many MSMEs from optimizing the use of data. This study aims to develop an Artificial Intelligence (AI)-based chatbot system capable of processing static transaction data sourced from the University of Malaya Malaysia Hackathon 2025 and providing simulated business insights and recommendations. The research method includes data preprocessing, rule-based intent detection implementation, and the integration of a Large Language Model (LLM) through the Gemini API to generate contextual responses. System evaluation was conducted using accuracy, precision, recall, and F1-score metrics, focusing on the chatbot’s technical performance and its ability to produce consistent responses. The results indicate that the chatbot is capable of presenting information such as best-selling products, busiest ordering hours, and sales summaries, with stable technical performance. Therefore, this system demonstrates potential as an AI-based solution to assist MSMEs in understanding transaction data and supporting simulated data-driven decision-making.]]>
Copyrights © 2026
