Smart Academic Enquiry System using Retrieved Augmented Generation

Dr.A.S.C. Tejaswini Kone ,B.L.V.Nagalakshmi , B.Ramana , G.Vijayendra Naidu , L.Jeswanthprasad , S.Charan Madhav
Department of Computer science and Engineering, Visakha Institute of Engineering @ Technology(A),Narava ,Visakhapatnam , India

Abstract:With institutions require intelligent solutions to handle large volumes of student enquiries. This paper presents a Smart College Enquiry Chatbot based on Retrieval-Augmented Generation (RAG), designed to provide accurate, context- aware, and real-time responses to queries related to admissions, courses, fees, schedules, and campus facilities. The proposed system integrates a semantic document retrieval module with a transformer-based generative model. A structured institutional knowledge bse—including prospectuses, FAQs, and policy documents—is indexed using dense vector embeddings. Upon receiving a user query, the system retrieves the most relevant information andaugments the generative model to produce precise and natural language responses. Unlike traditional chatbot systems that rely solely on pre-trained knowledge, the RAG-based approach enhances factual accuracy and domain adaptability by grounding responses in dynamically retrieved data. The architectureconsists of a document indexing module, a dense vector retriever, and a generative model that synthesizes user specific answers. Experimental evaluation demonstrates improved response relevance, reduced hallucination, and higher user satisfaction compared to baseline conversational models.