Consumers naturally vary in how they speak to voice-enabled technologies. Some issue short commands (“Alexa, weather today”), while others make more polite requests (“Alexa, could you tell me the weather today?”). Despite this variation and the rapid growth of voice-enabled devices in the marketplace, our understanding of how these conversational styles shape the consumer experience is limited.
Using large-scale datasets and behavioral experiments, we demonstrate that while commands are the predominant conversational style, using more elaborate requests rather than short commands alters how consumers objectively speak and their subjective experience. We show that requests lead to enhanced prosodic fluency (i.e., the smoothness and continuity of speech production) which in turn increases consumers’ perceived conversational fluency during the interaction (i.e., how natural and intuitive the interaction feels).
These changes drive important downstream firm outcomes: Consumers using requests are more likely to accept product recommendations, develop positive firm attributions, give better assistant ratings, and envision a broader range of device tasks. These effects are attenuated when consumers feel time-constrained, suggesting situational factors influence the effectiveness of different conversational styles.
These findings provide novel insights for firms deploying voice technologies, demonstrating that altering how consumers talk to AI assistants significantly shapes consumer-firm relationships.
Christian Hildebrand is Full Professor of Marketing Analytics and Director of the Institute of Behavioral Science and Technology (IBT) at the University of St. Gallen. He had doctoral and post-doctoral visits at Stanford University, Duke University, and the University of Michigan.
His work is at the intersection of Human-Computer Interaction (HCI) and consumer decision making and explores how new technologies change how people think, act, and make decisions. His research focuses on understanding and optimizing consumer-firm interfaces with an emphasis on digital voice assistants, chatbots, and mobile devices.
His research combines classic and novel methodologies from large-scale field experiments to feature-extraction from voice, image, and text data. His research has been published in leading academic and practitioner-oriented journals at the intersection of marketing and information systems such as Information Systems Research, Journal of Management Information Systems, Journal of Marketing Research, Journal of the Academy of Marketing Science, Marketing Letters, and Harvard Business Review.
The seminar will be held in English.
Major information: Daniela Bolzani (daniela.bolzani@unibo.it).