Cadence Reduces High-Risk Patient Contact Time by 97% with ElevenAgents

Chronic disease management company Cadence deployed ElevenAgents-based voice AI agents to significantly reduce the time from anomaly detection to patient contact. Previously, handling over 40,000 monthly alerts via manual IVR and clinicians took a median of 1 hour and 48 minutes, but after AI implementation, this dropped to 3.5 minutes, achieving a 97% reduction in time.

Implementation Background and Selection Criteria

As patient volume doubled annually, making it impossible to cope with hiring alone, Cadence introduced an automated voice response system. The key reasons for choosing ElevenAgents were the application of Zero Retention Mode for HIPAA compliance, low latency and natural voice to maintain conversation rates with anxious or elderly patients, and deployment within 1 month through native integration with existing infrastructure (such as Twilio).

Operational Metrics and Clinical Impact

After expanding through a human-in-the-loop pilot of approximately 2,000 cases, the system handled 250,000 calls in its first seven months. Key achievements are as follows:

• Auto-resolution rate: 60% of alerts are automatically resolved according to protocols without clinician intervention.
• Response rate: 47% increase compared to the 41% baseline for human nurses (RNs).
• Clinical workforce efficiency: Eliminating screening tasks allows nurses to focus on high-risk tasks.

ElevenLabs' transcription accuracy is equivalent to or better than alternative models, with particularly high recognition accuracy for drug names. Cadence plans to expand the scope of agent applications to reservations, device troubleshooting, and appointment preparation to extend the reach of clinical staff.