A physician who kept losing patients to a busy signal, and an engineer who taught Alexa to talk, decided the clinic front desk deserved better. This is the machine they built to answer the calls no one else could.
Here is an uncomfortable fact about American medicine: a large share of the money and frustration in a clinic doesn't come from the MRI or the malpractice premium. It comes from the phone that nobody has time to answer.
Clinics, by Clarion's own accounting, miss somewhere between 30 and 40 percent of their inbound calls. Not because the staff don't care - because there is a finite number of humans at the front desk and an effectively infinite number of people who need to reschedule, refill a prescription, ask about a bill, or figure out whether their referral ever went through. Every one of those missed calls is a small financial leak: a no-show that could have been prevented, an appointment that gets booked at the practice across town, a refill request that turns into an angry voicemail. Multiply it across a day, a week, a health system, and the telephone quietly becomes one of the most expensive pieces of equipment a clinic owns.
Clarion, a New York company founded in 2024, is a bet that this particular problem is now solvable by software - specifically by conversational AI agents that can answer a call, understand what the patient actually wants, and then do the unglamorous thing: book the appointment, process the refill, log the referral, and write it all back into the clinic's system of record. The company calls itself "the AI communication layer for healthcare," which is a tidy way of saying it wants to be the thing that picks up when the humans can't.
Clinics miss 30-40% of inbound calls due to staffing shortages.
The headline numbers are modest and the ambition is not. Clarion says it already serves tens of thousands of patients a month across virtual-care companies, health systems, and a health insurer it describes as roughly $5 billion in size. It does this with a team that has hovered somewhere around a dozen and a half people - the kind of ratio of patients-touched to employees that only works if a machine is doing most of the talking.
The clever part of Clarion's pitch is not that an AI can talk. Plenty of things can talk now. The clever part is that the same agent meets patients wherever they are - on the phone, over text, in a web chat - and then reaches into the messy back-end of a clinic to finish the job. Here is the work it takes on.
Answers inbound calls and books, moves, or cancels appointments - and captures after-hours demand the front desk never sees.
Handles refill requests around the clock, including the 2am request no clinic is staffed to take.
Every new referral gets a follow-up: Clarion reaches out, answers questions, and gets the patient scheduled before they slip away.
Fields routine billing and account questions so staff aren't tied up on calls that don't need a human.
Checks in after visits and during recovery, closing the loop that usually goes silent once the patient leaves.
Connects to 80+ EHRs including Epic and Cerner, so actions land in the system of record - not a spreadsheet.
That last card is the one that matters most, and the one most voice-AI startups quietly avoid. Talking is easy; integrating with Epic is not. By connecting to more than 80 electronic health record systems, Clarion is doing the plumbing that turns a nice demo into something a clinic can actually run. The demo is the marketing. The integration is the product.
Clarion's agents are tuned for the specialties where phone volume and scheduling complexity pile up. Each has its own vocabulary and its own version of the refill-and-referral shuffle.
The inclusion of value-based care is not decorative. Under that reimbursement model, clinics are paid for outcomes rather than visit counts - which makes every missed follow-up and every no-show cost more than it used to. In other words, the healthcare system is slowly rearranging itself so that the exact thing Clarion automates becomes the exact thing clinics can least afford to drop.
Healthcare software has a familiar failure mode: engineers who don't understand clinics, or clinicians who can't ship. Clarion's founding pair is an attempt to avoid both.
A Stanford/Harvard-trained physician who was on the founding teams of Two Chairs and Ophelia - two healthcare companies he helped scale past $100M valuations. He knows what a patient needs on the other end of the line, which turns out to be the hard part.
Built voice AI on Amazon Alexa and led AI/ML teams at Salesforce. He knows how to make a machine hold a conversation and connect to enterprise systems, which turns out to be the other hard part.
The rest of the team pulls from Amazon, McKinsey, Airtable, Regal, Ophelia, and Tandem, with degrees scattered across Harvard, Stanford, Penn, Columbia, Waterloo, and UC Berkeley. It is a small, deliberately technical group - the kind that can only touch tens of thousands of patients because the software is carrying the load.
| Round | Amount | Year | Notable Backers |
|---|---|---|---|
| Early venture (reported) | $5.4M total | 2024 | Accel · Y Combinator · Airtable Ventures · Sequoia (scout) |
Beyond the institutional names, Clarion took money from healthcare founders who have built in the same trenches - people from Ophelia, Medallion, and Counsel Health. That kind of cap table is its own form of distribution: it means the people who know exactly how painful clinic operations are decided this was worth funding. As with any early-stage company, the precise figures reported across sources vary; treat the total as approximate.
Clarion joins Y Combinator's W24 batch and begins building its healthcare communication platform.
Recorded early seed-stage funding activity as the company gets off the ground.
CEO Ryan Gallagher publicly introduces Clarion, framing it as front-office AI for healthcare clinics.
Reports serving tens of thousands of patients monthly across virtual-care companies, health systems, and a large insurer.
Most healthcare AI wants to do something dramatic - read the scan, suggest the diagnosis, replace the specialist. Clarion is doing something quieter, and arguably harder to make work at scale: it wants to reschedule you. The appeal of the boring problem is that it is enormous, universal, and measurable. A voice agent doesn't get tired at 4pm, doesn't quit, and doesn't take a call, get interrupted, and forget to follow up.
Whether patients ultimately prefer a competent machine to an overwhelmed human is the open question, and Clarion is careful to wrap its agents in safety features - bias assessments, risk monitoring - because a talking machine in healthcare is a trust problem before it is a technology problem. But the early traction suggests a straightforward truth: when the alternative is hold music and a callback that never comes, a machine that actually books the appointment starts to look pretty good.
Clarion is building the most advanced AI communication layer for healthcare.
Product walkthroughs, interviews, and demo content live on Clarion's own channels and its Y Combinator profile. Start here.
Figures are drawn from public sources and company statements; funding totals and patient counts are self-reported and approximate.
Clarion builds an AI communication layer for healthcare that answers the phones, texts, and web chats clinics can't keep up with. Its voice, SMS, and web agents handle scheduling, prescription refills, billing questions, referral intake, and after-hours follow-up, integrating with 80-plus EHR systems including Epic and Cerner. Founded in 2024 by a Stanford/Harvard-trained physician and an ex-Amazon Alexa voice-AI engineer, the New York company is a Y Combinator W24 startup backed by Accel and Airtable Ventures, and says it serves tens of thousands of patients a month across virtual-care companies, health systems, and a large health insurer.
Last updated: