Academic Mentoring Enquiry Chatbot: Design, Evaluation, and Educational Integration

Abstract

Academic mentoring in higher education is frequently constrained by limited staff time, scheduling conflicts, and the repetitive nature of routine student enquiries. This study designed and evaluated the Academic Mentoring Enquiry Chatbot (MSU ChatBot) as a lightweight conversational interface for academic programme enquiries, subject add/drop workflows, and mentor information access. The system was developed using a prototype-driven Agile approach and the open-source Rasa framework, with natural language understanding used for intent recognition and dialogue management used to coordinate multi-turn interactions. Requirements were gathered from 20 undergraduate students at Management and Science University, Malaysia, through a questionnaire and were analysed using thematic analysis. The findings showed unanimous demand for programme and major information (100%, n = 20) and for an integrated subject add/drop function (100%, n = 20). Mentor-detail access was supported by 90% of respondents (n = 18), while a minority raised privacy-related concerns. Functional verification further confirmed the feasibility of student-facing programme search, transactional course-adjustment flows, authenticated mentor retrieval, and administrator monitoring functions. The study demonstrates how an intent-driven, open-source chatbot can support routine academic advising while preserving a role for human mentors in complex or sensitive cases. The findings provide a practical basis for larger-scale usability and outcome evaluation in future deployments.