INFERSITY V1: A RETRIEVAL-AUGMENTED GENERATION (RAG) BASED CHATBOT FOR INTELLIGENT ACCESS TO UNIVERSITY RESOURCES

Abstract

University websites and documents like UG Rules and Prospectus can be hard to navigate, especially for students searching for specific info like admission or scholarships. Student services are not on 24-hour operation, which may slow down the process of information access and frustrate students. The proposed research paper develops the Infersity v1 chatbot that would improve the user experience regarding academic material requests. It’s based on a large language model (LLM) working with Retrieval Augmented Generation (RAG). This chatbot's goal is to solve a common issue that students encounter, the inability to get accurate and prompt responses from university administration about academic departments, programs, campus amenities, and student services. To provide precise and contextually aware responses, the suggested method integrates a domain-specific knowledge base sourced from university data into a RAG pipeline using Gemini 2.0-flash. The Retrieval Augmented Generation Assessment Score (RAGAS) was used in a case study conducted at Muhammad Nawaz Sharif University of Engineering and Technology in Multan (the First Public Sector Engineering University of South Punjab) to evaluate the system's performance. With a high RAGAS score of 0.96 for questions about prospectuses, 0.95 for questions about undergraduate regulations, and 0.95 for questions about websites, the chatbot showed excellent performance. These findings show that it is a dependable and easy to use academic assistant that has the potential to be widely used in higher education institutions.