Human-Like, Real-Time AI Scheduler launched by ScienceSoft

ScienceSoft has introduced a HIPAA-compliant AI scheduling assistant that enables natural, human-like conversations with patients. Built on the first-ever integration of Amazon’s Nova Sonic speech-to-speech model and the LiveKit Media Server, the assistant supports real-time voice interaction and seamless data exchange with hospital systems.

What sets the assistant apart is how it communicates. Most voice agents process speech in separate steps, which slows down the conversation and makes it sound robotic. The model behind our agent is the first to support bidirectional streaming through a single API, enabling listening and responding in real time without delays or awkward pauses. So, the patients get a whole new level of experience. They speak normally and get a timely, helpful response, without the friction they’ve come to expect from automated systems.

The system handles both incoming and outgoing calls, verifies patient identity, checks provider availability, and updates hospital systems during the call. Integration with EHR, CRM, and practice management platforms ensures appointments are confirmed or changed on the spot, without follow-up.

The scheduler runs in a secure, HIPAA-compliant cloud environment and is monitored with tools like Amazon Bedrock Guardrails and AWS Macie.

Our estimations show the agent can potentially reduce appointment booking time by 40%, cut call abandonment rates by 30%, and lower operational costs by at least 50%. As it can handle multiple calls simultaneously, it is expected to process 70% more calls per hour than a patient service representative.

Looking for Insider IT Info?

ScienceSoft experts are ready to share practical insights into healthcare technology and patient-facing solutions. Learn how technology is reshaping the way providers interact with patients.

Get an Expert Opinion

View source version on businesswire.com: https://www.businesswire.com/news/home/20250702115247/en/

ScienceSoft’s AI assistant processes 70% more scheduling requests per hour than a human and reduces booking time by 40%.