ABU Easy Go! Development and Optimization of a Machine Learning-Powered Chatbot for FAQ Assistance at Ahmadu Bello University, Zaria.

The advent of artificial intelligence (AI) and machine learning (ML) has significantly transformed how organizations interact with their stakeholders, with chatbots emerging as pivotal tools for enhancing user engagement and streamlining communication processes. This research introduces ABU Easy-Go!, a machine learning-powered chatbot developed to manage frequently asked questions (FAQs) at Ahmadu Bello University (ABU), Zaria, Nigeria. The chatbot leverages natural language processing (NLP) to provide instant, accurate responses to a broad spectrum of queries related to admission procedures, course registrations, campus facilities, and event schedules. The research outlines the design, implementation, and optimization of the chatbot, detailing the use of machine learning models, recurrent neural networks (RNNs) and transformer-based architectures BERT, to understand and generate responses in natural language. The technological framework incorporates Dialogflow platform and custom solution using TensorFlow and PyTorch to ensure a robust and scalable system. Integration with the Pantheon CMS and Tawk.to chat services was achieved to facilitate seamless user interaction. Testing and evaluation revealed high performance, with the chatbot achieving a training accuracy of 98% and a testing accuracy of 93%, reflecting strong effectiveness in handling queries. Post-optimization metrics demonstrated further improvements, with test accuracy increasing to 95%. This research demonstrates the potential of AI-powered chatbots in enhancing administrative efficiency in higher education institutions and provides insights into future directions for advancing chatbot capabilities through reinforcement learning and multimodal integration.