Student Registration Chatbot: A Large Language Model Approach - A Case Study of Adekunle Ajasin University Akungba Akoko, Nigeria

Many educational institutions struggle with inefficient student registration processes, often involving manual intervention due to issues like long waiting time, burden of repetitive registration, forgotten passwords or login difficulties. The study aims to refine the registration processes at Adekunle Ajasin University by developing an automated student registration chatbot. Using large language model (LLM) and prompt engineering technique, the bot simplifies the registration processes and eliminates the over-reliance and direct dependent on administrative support. The student registration bot, built with JavaScript and Next.js framework, interacts with the OpenAI GPT-3.5 model and leverages Retrieval-Augmented Generation (RAG) with a Pinecone database to retrieve relevant information for responses. Also, the Postman facilitates the translation of students’ text queries into structured queries for database interactions with the backend. The waiting time and waiting costs of five major parameters were used as our performance metrics. These are login issues, password reset, fee payment status, result visibility and course registration errors. Comparative analysis is done with the existing model. Experimental results demonstrate that the bot's waiting time and costs have a better performance over the existing model on all the metrics thereby leading to improved administrative efficiency and better student experience. Keywords: Generative Pretrained Model, Retrieval-Augmented Generation, Pinecone Database, Postman and Prompt Engineering. Aliyu, E. O., Ajisola, C.O, Akingbesote, A.O., Ogunlana, S.O. & Adelola, M.A. (2025): Student Registration Chatbot: A Large Language Model Approach - A Case Study of Adekunle Ajasin University Akungba Akoko, Nigeria. Journal of Advances in Mathematical & Computational Science. Vol. 13, No. 4. Pp 15-24. Available online at isteams.net. dx.doi.org