Persuasion-Building Fundamental Premises and Implications for Conversational Agents: A Conceptual Model in Captology

In the area of captology, computers as persuasive technology, conversational agent or dialogue system represent an interactive computing product as a key to changing people’s viewpoints. Our former original experimental research addressed that a virtual agent’s positive facial expressions and near-distance camera angle successfully influence participants’ thoughts during an experiment. However, fragmentation in this growing topic led to theoretical confusion. To move the field forward, this paper uncovers the evidence-based elements as the standard compliance for designing the persuasive conversational agent. First, the fundamental premises, implications, constructs, and indicators from our previous original research were generated. Second, a theory synthesis from related articles on the interaction between virtual agents to humans, with the aim of persuading, was analyzed. Thus, it derives twelve fundamental premises, two implications, three main constructs, and fifteen indicators. Third, a model elaborates on the relationships between these constructs and indicators was developed. Its foundation is based on the three related concepts of the Persuasion Knowledge Model, Stimulus-Organism-Response Theory, Fogg’s Captology Framework, and one measurement model of Partial Least Square Structural Equation Modeling. Hence, the theory synthesis provides the basis for inventing the proposed conceptual model, called the Agent-Subject Persuasion Model (ASPM), which emphasizes the originality of this research.