Developing an agentic RAG-powered chatbot for dynamic academic query management

Abstract

The growing reliance on Artificial Intelligence (AI) in education has paved the way for innovative solutions to automate and enhance academic support. Professors frequently face redundant and repetitive email queries from students, particularly about course details, which diverts attention from more meaningful engagements. This project develops an AI-powered chatbot using Retrieval-Augmented Generation (RAG) and LangChain for agentic workflows to address this challenge. The chatbot not only answers common course-related queries but also builds an evolving knowledge base by capturing unanswered questions, ensuring continuous improvement over time. Techniques such as multi-query generation, graph databases for knowledge structuring, and RAG-fusion workflows were employed to optimize information retrieval and query handling. By leveraging these advanced methodologies, the system reduces the workload for faculty while offering students a dynamic, responsive, and ever-growing source of course information. This agentic chatbot represents a significant step toward scalable and intelligent academic support systems.