‘Once you know what I’m like, nothing else will feel right’: Unpacking drivers of AI-enabled influencers’ mystical persuasion

Influencer marketing has become a compelling communication strategy; however, human influencers (HIs) pose challenges due to high costs, limited control, and growing consumer concerns about authenticity. These limitations have led brands to explore virtual influencers (VIs), synthetic agents created using computer-generated imagery or deepfake technologies. Despite their growing prominence, critical questions remain about how VIs perform as influencers, particularly the mechanisms that drive or hinder audience receptivity, relational engagement, and persuasion, which are essential factors for generating transactional value for brands and creators. This research addresses these gaps through a two-stage study. Stage One involves a systematic literature review of HI marketing to identify core determinants of influencer effectiveness. While VIs aim to deliver similar outcomes, the review highlights distinct mechanisms underpinning artificial agents, justifying an interdisciplinary approach. Stage Two develops and tests a novel theoretical model integrating task-technology fit, social penetration, and congruity theories to examine how VIs’ technological attributes and narrative strategies influence perceived influencer-role fit, persuasion, and engagement. Theoretically, this study contributes by shifting the analytical focus beyond human-centric frameworks, demonstrating that VI effectiveness is shaped by the interplay between technological embodiment (e.g. visual design and interactivity), narrative strategies (e.g. self-disclosure), and endorsement-context congruence. The findings highlight that traditional source-based models are inadequate for explaining the unique dynamics of artificial influencers. Practically, this research offers actionable guidance for brands and creators on persona design, strategic storytelling, and VI-product alignment in sectors such as technology, beauty, and fashion. It also provides considerations for policymakers concerning responsible deployment and the potential impact of hyper-realistic VIs on vulnerable audiences. By unpacking the technological, relational, and contextual dimensions of VIs’ influence, this study lays a foundation for future research on AI-enabled influencers and into how their evolving interactivity and realism may shape influencer marketing dynamics.