This research aims to explore how to enhance student engagement in higher education institutions (HEIs) using a novel conversational system (chatbots). The study applies a design science research (DSR) methodology and is executed in three iterations: persona elicitation, survey and student engagement factor models (SEFMs), and chatbots interactions analysis. In the first iteration, two k-means clustering analyses are applied to student data, including engagement on campus and student interaction with a virtual learning environment (VLE). The first analysis produces four different types of students based on their engagement and performance data, while the second analysis produces two clusters based on the students’ interactions with a VLE (in this case, Blackboard). The second iteration will produce SEFMs, which will include the factors that affect student engagement, confirmed using structural equation modelling (SEM). Finally, the third iteration will produce effective and usable chatbots that enhance student engagement. The pragmatic findings from this study will make three contributions to the current literature. Firstly, machine learning is used to build data-driven personas using k-means clustering. Secondly, a persona template is designed for university students, which supports the construction of data-driven personas. Thirdly, SEFMs will be built. Future iterations will build tailored interaction models for these personas and evaluate them using chatbots technology.