The State of Voice AI
in Healthcare
247 companies. Three layers of infrastructure. One map to make sense of it all.
The first comprehensive mapping of the entire voice AI ecosystem in healthcare, from foundational models to patient-facing applications. Your reference for evaluating vendors, mapping the space, or choosing your stack.
The Ecosystem Map
Core Voice Infrastructure
76The organs. STT, TTS, audio intelligence, and foundation models that power every voice AI system.
Voice Agent Platforms
40The nervous system. Orchestration, telephony, and tooling that connect infrastructure to real conversations.
Engagement Applications
130The interface. Where voice AI meets patients, clinicians, and care workflows.
Key Observations
What the map reveals
The engagement layer is crowded and undifferentiated
107 companies in healthcare engagement and patient access alone. Most are thin application wrappers around the same infrastructure. The question for clinics isn't 'which one?' but 'which architecture?'
Infrastructure is consolidating
The TTS and STT layers are maturing around a handful of well-funded players. ElevenLabs, Deepgram, AssemblyAI are becoming the default primitives. The competitive advantage has moved up the stack.
The platform layer is the strategic bottleneck
Only 44 companies in the orchestration and telephony layer, yet this is where the real complexity lives. How you route, fallback, and maintain memory across conversations is what separates a demo from a deployment.
Healthcare native voice AI barely exists
Most 'healthcare voice AI' companies are horizontal platforms with a healthcare landing page. Very few have built from the ground up for clinical workflows, safety, and compliance.