Beyond functionality: how conversational elements and usefulness perceptions shape voice AI experience in healthcare interactions

This study investigates how conversational elements in AI healthcare chatbots influence user experience during chronic headache consultations. Using three controlled participant-based voice experiments (N = 1,200), the study examines how empathy type, linguistic softeners, and prosodic diversity shape users’ evaluations of voice-based healthcare AI. The results show that emotional empathy produces significantly more positive user experiences than cognitive empathy, challenging the assumption that factual accuracy alone should dominate healthcare AI communication. Linguistic softeners that convey psychological comfort increase user satisfaction by reducing perceived conversational harshness, while expressive prosodic patterns are preferred over monotonous delivery, indicating that vocal variation functions as an important cue signalling emotional awareness in healthcare interactions. Beyond these conversational features, the findings highlight the importance of cognitive–emotional harmony in shaping user experience. When cognitive usefulness and emotional usefulness are aligned, users report significantly higher satisfaction than when one dimension dominates the interaction. Overall, the results demonstrate that effective healthcare AI communication depends not only on informational accuracy but also on the manner in which information is communicated. These findings suggest that healthcare voice AI should integrate emotional empathy, linguistically sensitive communication, and contextually appropriate vocal expression to create balanced and supportive user experiences.