Transforming KKU Library's Inquiry Services: From a Basic Chatbot to an Intelligent AI on LINE

Abstract

             This study aims to develop a LINE Chatbot utilizing Generative AI technology for the question-answering service of the Khon Kaen University Library, with the objective of addressing the limitations of the previous-generation chatbot system. The system was designed and developed by integrating Generative AI technology through the Llama 3.3 large language model in conjunction with the Retrieval-Augmented Generation (RAG) technique to retrieve information from the library's specialized knowledge base, alongside a translation API to enhance processing efficiency. An evaluation conducted among 56 users revealed an overall satisfaction level of 80.40 percent, particularly with regard to system stability and domain-specific query responses. Nevertheless, response accuracy remains a primary challenge. The findings of the system development indicate that the implementation of Generative AI has successfully elevated the user experience to a more natural level; however, continuous improvement of the knowledge base and ongoing model fine-tuning are still required in order to achieve a balance between conversational flexibility and information accuracy—both of which are fundamental to the provision of quality library services.