Designing an AI Agent for Live Football Viewing: A Game Design Approach to Social Presence and Engagement

Abstract

Live football is one of the most widely shared viewing experiences in the world, yet the analytical tools that make the sport richer to understand remain largely inaccessible to everyday fans during the match itself. This study explores how an AI companion, designed using game design principles, can enhance the live football viewing experience in terms of engagement and social presence. A functional prototype was developed in collaboration with Twelve Football and evaluated through user testing with five participants. Data was collected through the System Usability Scale (SUS), Net Promoter Score, and semi-structured interviews. The results show that the prototype achieved a SUS score of 74.5, placing it in the excellent range, and was consistently perceived as easy to use and non-intrusive. Engagement was primarily driven by the gamification mechanics rather than the conversational AI itself. Social presence was only partially achieved, with most participants perceiving the system as a useful tool rather than a social companion.