Are Self-Resembling Avatars Equally Effective Across Brands? Comparing Unknown and Prestigious Brands in Virtual Try-On

Grounded in privacy calculus theory and stimulus–organism–response framework, this research examines how three levels of avatar self‑resemblance (mannequin, digitally designed, photo‑realistic) influence perceived anthropomorphism, privacy risk, cognitive and emotional trust, self‑disclosure, and virtual try‑on (VTO) adoption across unknown and prestigious brands. Across both contexts, perceived anthropomorphism and privacy risk emerged as core antecedents of trust, and trust dimensions shaped behavior uniquely, with cognitive trust primarily driving outcomes for the unknown brand and emotional trust more influential for the prestigious brand. Unknown brands should leverage highly self-resembling avatars to build cognitive trust, whereas prestigious brands, less constrained by avatar type, could emphasize emotionally resonant and brand-aligned avatar. These findings highlight that privacy-related technology adoption in AI-mediated retail is shaped by both rational evaluations and affective trust processes. The study offers theoretical insights and managerial guidance for designing privacy-sensitive VTO and anthropomorphic AI systems aligned with brand positioning.