Turn-Taking in VUIs: Design Strategies for Achieving Entrainment Using Audio Cues

Abstract

Spoken interactions with machines have become increasingly common in everyday contexts, from virtual assistants to voice-operated service kiosks. While full-duplex technology-allowing simultaneous speaking and listening-it has shown promise in research and select commercial deployments, it remains technically demanding, resource-intensive, and not always the most practical choice for many real-world scenarios. In contrast, half-duplex turn-taking, where the system alternates between listening and speaking, continues to be a highly effective model for enabling interactive conversation -a structured, multi-turn exchange in which the user and system take sequential turns to share information, clarify intent, and progress toward a task goal.