Towards a theory of AI affordances actualizing: a case of an AI-powered CA in a hospital

Rapid advancements in generative artificial intelligence (AI) have ignited the discussion on its potential and prospects. It empowers conversational agents (CAs) to provide more fluent and contextually relevant responses. In healthcare, the use of AI-powered conversational agents (CAs) for conversational-based tasks, such as pre-consultation triage and post-consultation follow-ups, has been proposed as a promising way to supplement healthcare services. Despite the growing literature on the potential benefits of applying AI-powered CAs in hospitals, there is a paucity of studies discussing how their potential might be realized. Following an abductive research design, this research employs the theoretical lens of affordance and affordances actualization to investigate the process of realizing an AI-powered CA’s potential for triage in a Chinese tertiary hospital. It critically compares the existing theoretical framework with contextual empirical material, offering new insights into the process of actualizing AI affordances. Specifically, it identifies the feedback mechanisms between technology affordances and actualized outcomes and demonstrates the role of organizational factors in the process of affordances actualization. The newly constructed theoretical model in this paper lays the groundwork for future research into AI affordances actualization at a human-machine-institution interface.