AI Now Calls Patients When Their Vitals Are Dangerous

When a patient's remote monitoring device shows a dangerous reading, every minute counts. But until recently, Cadence's human clinicians took an average of 1 hour and 48 minutes to reach that patient. That's a long time when someone's blood pressure is dangerously high or their heart rhythm is off. Now, Cadence uses AI agents from ElevenLabs to make those critical calls. The AI calls the patient, asks about symptoms, and either escalates to a nurse or closes the alert. The result: time to patient contact dropped from nearly two hours to just 3.5 minutes.

That speed isn't just impressive—it's potentially life-saving. Chronic conditions affect 95% of Medicare patients, and with a projected shortage of up to 86,000 doctors, clinical teams can't keep up alone. AI agents help fill that gap. And patients actually respond better to them: answer rates on critical calls rose from 41% to 60%, and patients answer the AI 47% more often than human callers. People pick up when the call is timely and relevant.

But building a reliable healthcare AI isn't just about good prompts. Cadence's engineering team learned that clinical agents need identity verification and automated escalation to keep patients safe. They also need testing and analytics to ensure the AI behaves correctly as call volume grows. The webinar shared best practices for expanding to new workflows, emphasizing that mistakes in this field carry real clinical consequences.

For everyday people, this means faster help when health emergencies arise. It also means more efficient use of doctors' time, potentially reducing costs and improving care for millions with chronic conditions. As AI takes on more patient outreach, the focus must stay on safety and reliability—because in healthcare, speed saves lives, but trust makes it work.