Abstract: AI-driven virtual streamers can serve as a complement to human streamers to interact with consumers in live commerce. However, the uneven streaming performance of virtual streamers has hindered their widespread adoption in the e-commerce livestreaming industry. This study intends to investigate the influencing factors and underlying mechanism of consumers’ purchase decision-making behavior from the perspective of virtual streamers’ interaction characteristics. The specific research contents are as follows: (1) the effect of virtual streamers’ images on consumer pre-purchase behaviors with the moderating effect of product information presentation; (2) the effect of virtual streamers’ communication styles on consumers’ pre-purchase behaviors, as well as the moderating effect of consumer characteristics and product categories; (3) consumers’ post-purchase word-of-mouth evaluation of virtual streamers’ service quality. This research enriches the relevant research on virtual streamers, expand the theoretical system in the fields of live commerce and consumers’ online behavior, and provide practical guidance for merchants and e-commerce platforms. This study extends the theoretical system in the fields of e-commerce livestreaming and consumer online behavior, and provide a theoretical foundation and the construction of a theoretical framework for promoting consumer livestreaming engagement and purchase decision-making, as well as advancing the sustained and sound development of the digital economy and e-commerce.

Key words: livestreaming e-commerce, artificial intelligence, virtual streamers, purchase decision-making, electronic word-of-mouth