Company profile · 4 min read
Novoflow: healthcare operations agents across legacy EHRs
Computer-using agents for healthcare operations.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
Novoflow is selling the part of healthcare operations most vendors prefer to hide: the old systems.
What it does
Novoflow describes itself as computer-using agents for healthcare operations (YC profile). The company says its agents operate inside a medical practice’s EHR and automate operational bottlenecks without requiring staff to change how they work (current homepage).
The public workflow is specific. An administrator provisions EHR access. Novoflow’s forward-deployed engineers map the practice’s SOPs and configure the agent. The agent then operates like trained staff, including a live demo for a new-patient referral scheduler (homepage).
| Fact | What the public source says |
|---|---|
| Buyer | Healthcare providers, medical practices, and hospitals |
| Workflow | EHR-based operations, calls, scheduling, and interoperability |
| Deployment | The current homepage says deployment takes weeks |
| Public claim | The launch post reports multilingual calls, EHR scheduling, and projected savings; these are company claims |
Why it is interesting
The product decision is not really “AI receptionist or no AI receptionist.” It is whether a practice wants a configured operations layer that can work through its existing EHR and legacy interfaces.
That makes Novoflow different from a clean API-first product. The value is partly in the integration work. The company says its agents can work across Epic, Athena, Medent, and older HL7 feeds. The buyer should treat that as a claim to verify in the specific practice, not a universal compatibility promise.
The founder story lines up with the problem. The YC profile describes Mathieu Rihet as an Emory University scholar and former medical interpreter, and Georges Casassovici as a voice-AI engineer focused on healthcare interoperability. The background is relevant because the product sits between patient communication and messy practice systems. It is not evidence that the system improves care.
What I would resolve
There is a useful source conflict to clear up before a buyer gets excited. The current homepage says deployment takes weeks. The YC launch copy says the team can point a SIP trunk at the clinic’s number and go live in an hour. Those might describe different scopes: a demo, a narrow workflow, or a fully configured deployment. That difference should remain visible.
The company also reports 25-plus languages, large missed-call losses, and projected monthly savings. Keep the numbers labeled as company-reported. Ask for the clinic type, call volume, baseline, time window, and what “savings” means.
For a practice, the useful fit question is narrower: which workflow is painful enough to automate, structured enough to configure, and safe enough to escalate? If the answer is new-patient scheduling, Novoflow has a clear story. If the practice needs clinical judgment or unusual exception handling, the handoff matters more than the demo.
Short version: Novoflow is a fit for a provider willing to let a team map its existing systems. It is not a plug-in shortcut until the deployment claim and EHR boundary are clear.
Sources checked — 2026-09-19
Cohort context
Novoflow is listed in Spring 2025. In our 2026-09-18 directory snapshot, 10 of 143 listed companies in that cohort have YC’s primary industry label Healthcare (7.0%). This is a current-directory comparison, not an original intake count or a performance ranking. Nine-cohort dataset.
Public website snapshot
Observed 2026-09-19T16:15:46.638Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.
| Signal | Homepage observation |
|---|---|
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Observed |
| Typed structured data | Observed |
| Docs/developer link | Not observed in this response |
| Pricing link | Not observed in this response |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |
Public observations · Collection method. Missing links here do not establish that a capability or file is absent elsewhere.
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.
First1000 ↗ · X ↗