Enhancing the usability of a university student support services FAQ Chatbot

Abstract

Chatbots play a vital role in customer facing interaction. They offer real-time text or voice responses via intuitive human interaction systems and are often driven by AI technologies. Despite widespread adoption, their optimisation for university environments remains elusive. With a lens on Universities of Sub-Saharan Africa, this dissertation positions usability as essential in a chatbot’s ability to provide effective support for student support services. This dissertation identifies with the dire need for more rigorous design and development in line with the needs of a modern, inclusive university sensitive and responsive to its students’ varying degrees of multiculturalism, multilingualism, socio-economic standing and technology and digital literacy baseline skills. The topic of chatbot integration in University systems has received significant attention in recent years but few have focused on the interplay between usability factors such as, anthropomorphism, NLP, or UX. This has limited our understanding of how best to enhance chatbots, specifically in University student support services. This study aimed to identify the key design factors for an enhanced usability FAQ chatbot, tailored for University student support services. In pursuit of this aim, a usability design framework as well as a FAQ chatbot was developed and tested in a popular University in South Africa. The base functional requirements were inferred from extant literature and then fused with data collected from students and administrative members of staff. The design framework was also influenced by well-known usability principles and standards from ISO, Nielsen and Shneiderman and others. Google Dialogflow was used to develop the chatbot, architected by the design framework. Based on the DSR paradigm, the research followed a systematic approach encompassing usability design, framework development, tool evaluation, and FAQ chatbot development and testing. First-year students and administrative staff members were active participants and served as change agents during the iterative DSR process. Thematic analysis was used to carefully analyse the feedback from participants during the development stages and seed this into the ongoing design process. This iterative process of development and refinement allowed for a richer understanding of how users perceive and interact with the chatbot. During analysis of the final evaluation feedback, PLS-SEM illuminated relationships, dependencies and interactions among various usability design factors which influence the chatbot's overall usability. The major contribution is a blueprint for the design and development of an effective University student support services FAQ chatbot. Theoretical contributions include a usability design framework, iterative DSR development process and evaluation and feedback instruments using robust analysis techniques. There is a need for further research and refinement at the confluence of NLP, anthropomorphism and FAQ chatbot design frameworks.