About this formationAlmost every company here is attacking the same choke point from a different angle: the phone call and the payer portal, not the exam room. Parallel, CloudCruise, Health Harbor, Prosper, and Shasta Health are all building agents that click through legacy EHRs or call insurers exactly the way a human staffer would, because that mimicry, not clinical reasoning, is still the only reliable way into systems that were never built for API access. That reveals a real split in the group: a handful of companies like Vera Health and FutureClinic are pushing toward clinical judgment and patient-facing chat, while the large majority, from Dodo and Clarion to Cair Health and Careforce, stay firmly in scheduling, eligibility, and RCM, the unglamorous admin load that clinics will actually pay to remove first. This formation's 4.53x growth outpacing the 2.97x growth of Call Center and 2.37x of Telehealth suggests healthcare admin automation is compounding faster than the broader categories it draws from, even as Digital Health overall grows at a comparatively sluggish 0.41x, a sign that generic digital health tooling is cooling while narrow, workflow-specific agents accelerate. The live concern tracking this group, about reconstructing what an agent saw and decided when work passes between communication, records, and human review, is not academic given how many of these agents (Entropy, Fuse, Healthtech 1) hand off tasks across intake, billing, and staff approval in a single patient encounter. Widespread agent adoption in this space became plausible only once large language models got reliable enough at multi-step browser and voice interaction to survive contact with real payer portals and phone trees, a capability that only became commercially trustworthy in the 2023-2024 window these companies were founded in.
Cambrian Signal, “Healthcare Agents,” publication yc-2026-10-02-semantic-v8, snapshot October 2, 2026.
cambrian.jafar.com/formation/healthcare-agents