Ellipsis Health’s voice AI agent Sage is now a core tool for health plans and care management companies, helping them interact directly with patients for medication reminders, program recruitment, and postoperative follow-ups—marking a shift from its original focus on detecting depression through vocal biomarkers.
According to company executives, Sage’s expansion into broader patient engagement reflects growing demand for AI-driven solutions in healthcare. The technology, which uses natural language processing and machine learning, is designed to handle routine but critical tasks that often fall through the cracks in traditional care models.
Yet as voice AI becomes more embedded in patient care, questions remain about its effectiveness, ethical use, and whether the healthcare sector needs dedicated specialists to manage these tools. We examined Sage’s capabilities, its real-world applications, and the challenges ahead for voice AI in medicine.
Ellipsis Health’s Sage is a voice AI agent used by health plans and care management companies to automate patient interactions, including medication reminders, program recruitment, and postoperative follow-ups. Originally developed to detect depression through vocal biomarkers, Sage now handles over 10 million interactions annually, according to company data. While the technology improves access to care, experts warn about privacy risks and the need for specialized oversight.
What Is Sage, and How Does It Work?
Sage, developed by Ellipsis Health, is a voice AI platform designed to simulate human-like conversations with patients. Unlike generic chatbots, Sage is trained on healthcare-specific language patterns, allowing it to:
- Schedule and confirm appointments via phone or messaging
- Remind patients about medication adherence with personalized follow-ups
- Recruit patients into care programs (e.g., chronic disease management)
- Conduct postoperative check-ins to monitor recovery progress
According to a company white paper, Sage processes over 10 million interactions annually, with an average response time of under 3 seconds. The system integrates with electronic health records (EHRs) to pull patient histories, ensuring personalized interactions.
From Depression Detection to Patient Engagement: Ellipsis Health’s Evolution
Ellipsis Health was founded in 2015 with a focus on vocal biomarker analysis for mental health, particularly depression. Early research showed that speech patterns—such as pitch variability and response time—could indicate depressive symptoms with up to 80% accuracy, according to a 2019 study in JAMA Psychiatry.
However, the company pivoted in 2020 toward broader patient engagement after health plans and insurers expressed demand for scalable, low-cost solutions to reduce no-show rates and improve adherence. Today, Sage is deployed by major players including UnitedHealth Group and CVS Health, handling tasks that would otherwise require human call center staff.
How Effective Is Sage Compared to Human Interactions?
Early adopters report mixed results. A 2023 McKinsey analysis found that AI-driven patient engagement tools reduce no-show rates by 15–25% when combined with traditional reminders. However, critics argue that voice AI lacks the empathy and nuance of human interactions, particularly for complex emotional or medical discussions.
Dr. Emily Chen, a digital health researcher at Harvard Medical School, notes that while Sage excels at logistical tasks, it may struggle with patients who require emotional support. “For someone in crisis, a robotic voice might feel dismissive,” she says. “The technology is a tool, not a replacement.”
Privacy and Ethical Concerns: Can Voice AI Be Trusted?
Voice AI raises significant privacy questions. Unlike text-based interactions, spoken conversations can inadvertently capture sensitive health data—such as symptoms, mental health struggles, or financial concerns—without explicit consent. The Health Insurance Portability and Accountability Act (HIPAA) requires strict safeguards for patient data, yet voice recordings may not always be encrypted or stored securely.
Ellipsis Health states that all Sage interactions are HIPAA-compliant, with data encrypted and stored only for the duration of the interaction. However, a 2023 FTC warning highlighted risks of unauthorized data sharing in healthcare AI, including potential breaches.
Does Healthcare Need Specialized Voice AI Experts?
As voice AI expands, some experts argue that the healthcare sector lacks the specialized workforce to manage these tools effectively. A Deloitte report estimates that by 2025, the U.S. will need 90,000 more AI professionals in healthcare—many with expertise in ethics, compliance, and patient interaction design.
Melissa McCool, COO of Ellipsis Health, acknowledges the skills gap but emphasizes that training programs are emerging. “We’re seeing partnerships with universities to create certificates in healthcare AI ethics,” she says. “The goal is to ensure these tools are deployed responsibly.”
What’s Next for Sage and Voice AI in Medicine?
Ellipsis Health is testing Sage’s capabilities in new areas, including:
- Mental health triage: Using vocal biomarkers to flag patients at risk of suicide or severe depression
- Multilingual support: Expanding Sage to handle non-English interactions (currently in English and Spanish)
- Integration with wearables: Syncing voice prompts with smartwatches or fitness trackers for real-time health monitoring
The company plans to roll out a new compliance framework by mid-2025 to address ethical concerns, including patient consent for voice recordings.
Key Takeaways
- Sage automates routine patient interactions, reducing costs for health plans while improving adherence and engagement.
- Effectiveness varies: Best for logistical tasks (reminders, scheduling) but may lack empathy for emotional or complex medical needs.
- Privacy risks remain: Voice data requires strict HIPAA compliance, though breaches are a growing concern.
- Specialized expertise is needed: Healthcare AI workforce shortages could hinder adoption without targeted training programs.
- Future expansions include mental health triage and multilingual support, but ethical guardrails are still developing.
For readers interested in exploring voice AI in healthcare further, Ellipsis Health’s resource hub offers case studies and compliance guidelines. The next major checkpoint for Sage’s development will be the release of its updated compliance framework in mid-2025.
We welcome your thoughts on the role of AI in patient care. Share your experiences or questions in the comments below.