AI PERSUASION: Interaction Cues, Emotions, and Purchase Intention
conversational artificial intelligence; persuasion; emotions; purchase intention; consumer involvement.
This thesis project investigates how conversational artificial intelligence operates as a persuasive agent in consumer decision-making processes, with the objective of examining how central and peripheral cues in conversational AI shape consumer emotions and purchase intentions, considering consumer involvement. Grounded in an integrative framework that combines the Stimulus–Organism–Response model and the Elaboration Likelihood Model, the study analyzes how different interaction cues embedded in conversational artificial intelligence systems influence consumers’ affective responses and subsequent behavioral intentions. Central cues are operationalized through recommendation reliability and recommendation accuracy, whereas peripheral cues are represented by human-like empathy and recommendation choice. Emotions are conceptualized as positive and negative affective states that function as mediating mechanisms linking AI interaction cues to purchase intention, while consumer involvement is proposed as a moderating variable shaping the strength of these relationships. The research adopts a quantitative, descriptive approach based on a survey that will be conducted with consumers who have recently used conversational artificial intelligence systems in purchase-related activities, such as information search, comparison of alternatives, or evaluation of buying options. Data will be analyzed using structural equation modeling, allowing for the simultaneous testing of direct, mediating, and moderating effects within the proposed conceptual model. By conceptualizing conversational artificial intelligence not merely as a functional tool but as a communicative and persuasive actor, this thesis aims to contribute to the consumer behavior literature by proposing a process-oriented explanation of the affective-cognitive mechanisms underlying AI-mediated persuasion, as well as by discussing theoretical and managerial implications for the design of AI–consumer interactions.