<![CDATA[Students often face difficulties in finding relevant thesis references, which can hinder the completion of their final projects and delay graduation. This study aims to develop a chatbot using a Large Language Model (LLM) integrated with WhatsApp as an interactive and efficient solution for academic reference search. A total of 795 classified thesis documents were collected from the Faculty of Industrial Technology and Informatics, UHAMKA. The system was built using the LangChain framework, including Setting Table Schema, Semantic Search, Rank Result, and Natural Language Interface for Databases. Implementation results showed that the chatbot successfully responded to natural language queries with 100% accuracy. User Experience Questionnaire (UEQ) evaluations indicated strong positive responses, with Clarity (2.08) and Accuracy (2.00) achieving “Excellent” ratings indicate high levels of efficiency in conducting thesis searches. This research demonstrates the effective application of LLMs in conversational academic search systems and offers a foundation for the development of similar services in other higher education institutions.]]>
