Confiding in a maven or a charmer? Impression management behavior of artificially intelligent conversational agents and consumer self-disclosure

Purpose Data-driven hyper-personalization promises stronger consumer relationships, yet growing privacy concerns and regulations limit marketers’ access to consumer data. In this context, voluntarily self-disclosed consumer information becomes critical for creating mutual value and sustaining consumer–brand relationships. This study aims to investigate whether artificially intelligent conversational agents (CAs) designed with impression management behavior can effectively elicit consumer self-disclosure. Design/methodology/approach Synthesizing research across human-human, computer-human, and brand-human communication, the study examines the effects of impression management tactics – specifically ingratiation and self-promotion –through a series of experiments employing simulated conversations with custom-built CAs. Behavioral and intentional self-disclosure are assessed using both observable and latent measures. Mediation analyses test the roles of perceived responsiveness and trust in the CA. Findings The findings establish that CAs displaying impression management behavior elicit greater consumer self-disclosure than those without such behaviors. Results further indicate that perceived responsiveness and trust in the CA mediate this effect. Moreover, in hedonic contexts, ingratiating CAs (the “charmers”) prompt more disclosure, whereas in utilitarian contexts, self-promoting CAs (the “mavens”) are more effective. Research limitations/implications The study relies on simulated conversations with custom-built CAs, which may not fully replicate real-world AI-mediated interactions or the dynamic evolution of long-term consumer–brand relationships. Practical implications The results provide actionable guidance for designing AI-mediated consumer interactions. Incorporating impression management tactics into CA communication can enhance perceived responsiveness, build trust, and increase voluntary consumer self-disclosure – particularly valuable in privacy-restrictive environments. Originality/value This research advances understanding of trust-building and interpersonal communication in AI-mediated consumer–brand interactions by demonstrating that impression management behaviors in CAs can strategically influence consumer self-disclosure. The study bridges theoretical insights across communication domains and identifies context-specific effects of ingratiation vs self-promotion in AI-driven exchanges.