The Sound of Good Service: How Dominant and Submissive Language in AI Voice Assistants Shapes Consumer Responses Across Service Settings

Artificial intelligence (AI) voice assistants (VAs) function as service agents, delivering information and recommendations tailored to users’ specific needs. VA agents can engage in voice conversations, in which they may convey dominance or submissiveness in response to consumer requests, thereby shaping consumer experiences across service settings. Drawing on the notion of master–servant relationships from assemblage theory, this research examines how, and in which contexts, VAs should use dominant and submissive language to shape consumer interactions. Across six experimental studies, we find that VAs’ submissive language increases usage intentions and perceived interaction value by enhancing rapport and, in turn, comfort, which ultimately drives these consumer responses, even in authority-oriented service settings. However, under time pressure, VAs’ use of either dominant or submissive language equally predicts rapport, thereby disrupting the serial pathway through comfort that leads to usage intentions and perceived interaction value. Together, these findings help specify how service managers can best design VA language to enhance the value of consumer interactions with this technology in service contexts.