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

PolyAI has launched Dialog-RSN-1, an audio-native model designed to process spoken caller input directly rather than relying on traditional speech-to-text transcripts. By integrating speech recognition, turn-taking, and function calling into a single architecture, the model aims to improve the responsiveness and accuracy of automated voice systems. While the text-to-speech output remains handled by a separate component to preserve voice customization, this development marks a move toward unifying complex voice processing tasks within a single neural framework.