PolyAI Releases Dialog-RSN-1: An Audio-Native Dialog Model That Fuses Turn-Taking, Speech Recognition, Function Calling, And Response

MarkTechPost · Michal Sutter · 31 July 2026, 05:06

PolyAI released Dialog-RSN-1, an audio-native dialog model that processes caller audio directly and integrates turn-taking, speech recognition, function calling, and response generation into a single system. The model achieves sub-300ms response times, 11% relative improvement in containment at a restaurant group, and 37% latency reduction at an insurer. Deployment is available exclusively through PolyAI's platform for enterprise customers in industries including restaurants, insurance, and hospitality, with no open-source weights or public API planned.

Why it matters

PolyAI has introduced Dialog-RSN-1, a dialog model that perceives caller audio directly instead of reading an ASR transcript. It fuses turn-taking, speech recognition, function calling, and response generation into a single audio-native model, keeps TTS separate so the output voice stays controllable, and runs as a request-based LLM rather than an always-on stream. PolyAI reports sub-300ms responses in live deployments. The post PolyAI Releases Dialog-RSN-1: An Audio-Native Dialog Model That Fuses Turn-Taking, Speech Recognition, Function Calling, And Response appeared first on MarkTechPost.