Ellipsis Health buys Kouper Health: its AI care agent wants to live inside the medical record

By citybiz · November 5, 2026.

Ellipsis Health, a developer of AI-powered care management technology, has acquired Kouper Health, an AI care navigation company that specializes in integrating with electronic health records (EHR) and coordinating patient care. Financial terms were not disclosed.

The deal combines Sage, Ellipsis's conversational care manager, with Kouper's technology to embed AI agents in clinical workflows and maintain a longitudinal patient record: a continuous history that runs across appointments, treatment, discharge and follow-up. Ellipsis expects this to let Sage identify the right clinician and schedule appointments directly within the health record.

Kouper was founded in 2022, and its investors include CVS Health Ventures, General Catalyst and 25madison. Its team is joining Ellipsis. Both companies will keep serving customers under their current brands while they build a joint roadmap, and new products for health systems are expected in the coming months.

Sage works by phone and text message. It handles patient enrollment, clinical assessments, medication adherence support and post-discharge follow-up. It uses what Ellipsis calls the Empathy Engine, a proprietary system trained on more than 3 million labeled clinical conversations that analyzes vocal signals to estimate emotional state and adjust its responses.

The performance figures come from the company itself: enrollment in care programs six times faster, 60% fewer administrative tasks, task completion rates up to 66% higher than those of human agents, and returns on investment of up to seven times. The company acknowledges that results vary by deployment. The material includes no independent verification, and “up to” describes the best case, not the typical one. Ellipsis says it serves seven of the 20 largest health plans in the U.S.

Our take: what matters is not the acquisition itself but what Ellipsis is buying. Kouper brings the least glamorous and hardest part of healthcare AI: access to the systems where the data lives. As a rule, a highly empathetic agent that is disconnected from the health record can only talk. An integrated one can act, schedule and close the loop. That is the difference between a chatbot and a care manager. We see the same dynamic in other sectors: the winners control the “plumbing,” not just the model.

Some caveats are worth keeping in mind. Comparing task completion with human performance depends on which tasks are measured; enrollment and reminder calls are repetitive and favor automation. A system that reads emotions from a person's voice and writes to the health record also requires clear oversight, error auditing and privacy rules. The results should be checked against independent studies.

In the short term, this puts pressure on administrative and coordination jobs at insurers and hospitals, although the source itself presents adoption as a response to staff shortages. In the long term, the promise is powerful: if AI absorbs the paperwork and ongoing follow-up, human clinicians can spend their time on judgment and on their relationship with the patient, and care stops being fragmented across episodes. That continuity is also the foundation for catching problems earlier and, over time, moving toward more preventive medicine. It remains to be seen whether the combined product delivers on its promise.