This study aims to investigate how avatar attire and conversational style jointly influence consumers' perceived artificial intelligence (AI) trust and purchase intentions, considering the moderating role of relationship orientation.
Three experimental studies (N = 1,070) were conducted. Study 1 tested the effects of avatar attire (professional vs. casual) and conversational style (competence- vs. warmth-oriented) on perceived AI trust and purchase intentions. Study 2 examined whether consumers' relationship orientation (communal vs. exchange) moderates the impact of attire. Study 3 employed a 2 (relationship orientation) × 2 (avatar attire) × 2 (conversational style) design to test the full moderated mediation model.
Avatars in professional attire elicit higher perceived trust and purchase intentions than casual attire. The effect of avatar attire and conversational-style interaction on perceived AI trust and purchase intentions depends on consumers' relationship orientation, rather than being driven solely by congruence. Professional attire is more persuasive in exchange-oriented contexts, whereas casual attire is more effective in communal contexts.
This research integrates the computers are social actors framework and the stereotype content model to clarify how visual appearance and communication style jointly shape consumer responses to AI avatars. The findings provide actionable guidance for designing virtual brand agents, highlighting the importance of aligning avatar appearance with communication style and relational context to enhance trust and purchase intentions in AI-mediated interactions.
