Facts or Feelings? How Virtual Streamers Reduce Consumer Uncertainty Across Product Types

ABSTRACT Advances in artificial intelligence have accelerated the adoption of virtual streamers in live‐streaming e‐commerce. Yet the effectiveness of emotional language used by virtual streamers remains unclear in existing research. Drawing on signaling theory and a consumer‐uncertainty perspective, this study examines how virtual streamers' language styles (informational vs. emotional) interact with product types (search vs. experience) to influence consumers' purchase intention, and how this interaction effect is moderated by live‐streaming room activity. This study conducts three scenario‐based experiments. Three experiments were designed to progressively examine the proposed hypotheses. The results reveal a clear matching effect between virtual streamers' language styles and product types. When promoting search products, the use of informational language helps reduce consumers' perceived uncertainty about the product, which in turn increases purchase intention. In contrast, when promoting experience products, emotional language is more effective because it decreases consumers' uncertainty toward the virtual streamer, ultimately generating higher purchase intention. Moreover, live‐streaming room activity serves as an important boundary condition that moderates this interaction effect. Specifically, this interaction effect will further strengthen under high‐activity levels, whereas the advantage of emotional language for experience products will weaken under low‐activity levels. This study develops a dual‐uncertainty framework to explain when and why different language strategies adopted by virtual streamers become persuasive across product contexts.