Beyond engagement rates: how opinion leadership and parasocial interactions drive virtual influencer effectiveness

This study examines the psychological and behavioral mechanisms driving the success of AI-powered virtual influencers. Drawing on the Stimulus-Organism-Response framework, the study investigates how specific digital stimuli, including posting frequency, credibility, visual attractiveness, and expertise, trigger internal cognitive (opinion leadership) and emotional (parasocial interaction) evaluations, which subsequently drive the consumer’s intention to follow the influencer’s advice. Utilizing a near-census dataset of 116 virtual influencers analyzed through Partial Least Squares Structural Equation Modeling (PLS-SEM), the findings reveal that expertise, credibility, and visual attractiveness are robust predictors of both perceived authority and parasocial closeness. Notably, and contrary to established models of human influencer marketing, frequent posting does not significantly impact these internal psychological states for virtual personas. The findings further suggest that perceptions associated with authority and parasocial closeness may emerge even without human lived experience. Overall, the effectiveness of virtual influencers depends more on perceived expertise and a consistent, believable persona than on simply posting frequently or engagement metrics alone.