Comparing text- and video-based interactions with virtual assistants in e-commerce

Human-computer interactions have been profoundly changed by the increasing use of Virtual Assistants (VAs) into a variety of digital ecosystems which improved user capability to perform several tasks such as receive product recommendations and complete online transactions. Although text-based chatbots have long been used in different contexts, little is known about the potential of video-based Vas, in particular in e-commerce settings. This study aims to examine how VA textual versus video communication modalities impacts user’s VA perceived social presence and its role in influencing trust, engagement, intention to use and their overall satisfaction with the assistant. The proposed conceptual model was tested using a structural equation modelling approach. According to the results, a video-based VA is perceived as more socially present as compared to the textual one. This positively influenced the dimension of trust, and engagement which lead to an increased satisfaction and intention to use the VA. This study adds to the expanding corpus of research on human-VA interactions. E-commerce services could benefit from these results by improving their customer satisfaction through more human-like interactions using video-based VAs.