期刊
INTERNATIONAL JOURNAL OF HUMAN-COMPUTER INTERACTION
卷 36, 期 12, 页码 1136-1149出版社
TAYLOR & FRANCIS INC
DOI: 10.1080/10447318.2020.1722399
关键词
-
Imagine walking into a department store to shop for various products. Based on the social heuristics related to expertise, you would likely favor and trust the advice conveyed by a product specialist dedicated to his/her product base than by a generalist advisor who opines on all product categories. As per the computers are social actors theory, this effect should also apply to people's interaction with embodied conversational agents simulating product advisors in a multi-product category e-commerce. This study evaluated the effects of specialist agent design using surfaces cues that were 1) agents' self-introduction as product specialists, and 2) agents' assignment to dedicated product categories, in a multi-product category e-commerce website. An experiment (n = 122) was conducted to compare the effects of the specialist agent design against the conventional generalist agent design where one agent advised on all product categories. Consistent with source credibility theory and multiple source effect theory, the results demonstrated that specialist agent design increased the perceived agent's expertise, message trustworthiness, social presence, website trust, and purchase intention. Further, mediation analyses revealed that perceived agent's expertise and message trustworthiness mediated the effects of specialist agent design on purchase intention, thus affirming the source credibility model. As predicted by multiple source theory, the implementation of multiple agents in the specialist agent design prompted a higher social presence, which was found to be a mediating factor that led to higher perceived website trust ability and benevolence. Finally, the effects of specialist agent design on purchase intentions were also mediated by perceived website trust ability and benevolence. Theoretical and practical implications are discussed in this paper.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据