Purpose: The emergence of virtual influencers (VIs) as AI-driven digital entities has transformed the landscape of influencer marketing. However, the existing literature remains fragmented, lacking a unified conceptual framework to explain the complex dynamics of consumer–virtual influencer relationships. This study aims to synthesize fragmented qualitative findings and provide a comprehensive conceptual model regarding the mechanisms of the formation, maintenance, and consequences of these relationships.
Methodology: Utilizing a qualitative meta-synthesis approach based on the seven-stage framework of Sandelowski and Barroso, this study systematically analyzes 26 studies that contain qualitative empirical evidence or substantive qualitative components. for mixed-method and multi-method studies, only the relevant qualitative findings were extracted and coded, while conceptual, typological, and quantitative components were used only to contextualize or interpret the synthesized qualitative evidence. Quality appraisal was conducted using the CASP qualitative checklist, followed by iterative coding, constant comparison, and conceptual abstraction.
Findings: The analysis resulted in the identification of 81 basic codes organized into six overarching themes: consumer perceptions and motivations; virtual influencer design and characteristics; relational and emotional mechanisms; impacts on consumer well-being; marketing effectiveness and commercial outcomes; and ethical challenges. The model integrates these themes while recognizing consumer well-being as a distinct psychosocial outcome that may be experienced positively or negatively across different forms of consumer engagement with virtual influencers.
Conclusion: This conceptual model provides a roadmap for future research in digital marketing and offers a theoretical framework for understanding the nature of human-machine relationships. The findings also suggest practical strategies for brands to effectively manage virtual influencers and mitigate perceptual challenges, such as the uncanny valley paradox.