The voice in the machine: designing embodied conversational agents for rapport in complex online service delivery

PurposeWhile emerging research explores the effects of communication cues of warmth and competence in conversational artificial intelligence separately, less is known about their combined impact. This study aims to investigate the interplay between an enthusiastic versus calm tone of voice and a socio-emotional versus task-focused interaction style in service encounters involving embodied conversational agents (ECAs) and customers.Design/methodology/approachThrough a conceptual framework and empirical analysis involving two interactive experiments (n = 298, n = 390) and the Gabor-Granger pricing method (n = 212), this research examines how warmth and competence are conveyed through ECA voice prosody and how interaction styles impact perceptions of believability, rapport and customer service usage and advice compliance intentions.FindingsThis study demonstrates that a complementary configuration of warmth and competence (e.g. enthusiastic voice with task-focused style or calm voice with socio-emotional style) increases ECA believability and influences higher price-point preferences. This study also evidenced a moderated serial mediation process where believability enhances feelings of rapport, amplified by a willingness to suspend disbelief.Research limitations/implicationsThis research advances the understanding of ECAs by highlighting the synergistic effects of voice prosody and interaction style on customer perceptions. It emphasizes the critical role of complementing warmth and competence cues to increase believability and rapport in ECAs that deliver complex services.Practical implicationsThe findings underscore the importance for service providers to strategically align voice prosody and interaction style to optimize customer responses to ECAs. This suggests significant benefits from investing in advanced voice technologies and adaptable interaction styles in ECA design and deployment.Originality/valueThis research contributes to the field by demonstrating that interaction style and voice prosody are not only vital in determining the believability of ECAs as providers of complex services but also that their complementary relationship optimizes ECA dialogue capabilities and has implications for the pricing of such services.