Zero-Latency Patient Triage at Scale: How HealthNow Automated 50,000 Daily Calls

By: Caleb Bennett
February 15, 2026
Healthcare, Voice AI
Healthcare AI
95% Reduction in Triage Time
$12M Annualized Savings
0.0% Safety Incidents

In the high-stakes world of telehealth, latency isn't just an inconvenience. It's a clinical risk. HealthNow, one of North America's fastest-growing telemedicine providers, faced a critical bottleneck. As their patient base exploded by 400% in Q3 2025, their manual intake processes began to fracture.

Wait times for initial triage ballooned from 2 minutes to 45 minutes. The burnout rate for intake specialists hit 25% monthly. They needed an infrastructure shift, not just more headcount. They needed an autonomous, clinical-grade voice layer capable of handling infinite concurrency.

The Challenge: "Human Middleware" Limits

The math of human-led triage is brutal. An average intake call takes 6 minutes. To handle 50,000 calls per day with 24/7 coverage, HealthNow required a workforce of nearly 800 active agents. The logistics of hiring, training, and ensuring HIPAA compliance for a workforce that size created a drag coefficient that stalled growth.

"We didn't just need a chatbot. We needed an empathetic, clinically aware agent that could handle 50,000 concurrent streams without hallucinating. Qore delivered that infrastructure on Day 1."

— Sarah Jenkins, CTO at HealthNow

The Solution: Deep Clinical Voice Agents

HealthNow deployed Qore’s Voice AI layer to handle L1 patient triage. This wasn't simple IVR routing; it was a deployment of our proprietary Clinical Reasoning Engine (CRE). The system performs complex intake tasks in real-time:

  • Dynamic Symptom Analysis: Using non-linear questioning paths to gather history, ruling out red flags like "shortness of breath" or "radiating chest pain" instantly.
  • Identity Verification: 400ms voice biometrics and database matching to authenticate patients securely.
  • Insurance Validation: Real-time API calls to payer gateways (BlueCross, Aetna, UHF) to confirm active coverage before routing to a doctor.

Technical Deep Dive: Achieving Zero Latency

To achieve "human-level" conversational fluidity, we optimized the entire stack for speed. The average Voice-to-Voice (V2V) latency was clocked at 650ms, well within the threshold of natural conversation.

We utilized edge-deployed inference nodes closer to HealthNow's major user clusters (NY, LA, Chicago) to shave off network latency. The result is an experience where patients often don't realize they are speaking to an AI until the end of the call.

The Results: Infinite Elasticity

The impact was immediate and localized. Within 2 weeks of full deployment, HealthNow saw:

  • Zero Wait Times: System scales instantly to any call volume, eliminating queues entirely.
  • 72% Cost Reduction: Cost per intake dropped dramatically, freeing up budget for senior clinical staff.
  • 99.9% Data Accuracy: Automated transcription and EMR entry eliminated manual data errors.
LOGO