Parlance Launches Version 12, Approaching 2 Billion Calls

Traditional IVR systems often fail when a caller uses terminology outside of a pre-programmed vocabulary. Parlance 12 addresses this by integrating a large language model to handle unresolved requests, moving beyond simple pattern-matching. When the system encounters an unfamiliar phrase, it now interprets the caller’s intent in real-time, routing them to the correct resource rather than hitting a dead end.

Beyond semantic routing, the upgrade refines first-pass recognition accuracy through a hybrid approach of statistical and language modeling. This ensures the system remains reliable for the 400-plus health systems currently using the platform. According to Sanjay Yadav, Head of Engineering Development, the goal is to infuse advanced AI into trusted environments, effectively reducing the reliance on live operators for routine inquiries. By turning the switchboard into an automated digital front door, Parlance aims to provide immediate answers for directions, hours, or common questions, ensuring the system evolves alongside changing patient expectations.