It is 6:47 p.m. on a Tuesday in Cleveland. The last patient of the day has just pulled out of the parking garage. The physician who saw her is not still at her desk typing. She is in her car. The note is already in Epic. The codes are already attached. The referral letter is drafted. Somewhere in San Francisco, a piece of software named AutoScribe did the paperwork while she was doing the medicine.
That is Ambience Healthcare's pitch in one sentence. It is also, more or less, the reason the company is now worth $1.25 billion.
Who they are, right now
Ambience Healthcare is a San Francisco AI company that sells ambient documentation, coding, and clinical workflow software to large American health systems. Cleveland Clinic. UCSF Health. Houston Methodist. Memorial Hermann. The kinds of institutions that do not buy software on a whim, and certainly not AI software.
In July 2025 it closed a $243 million Series C led by Oak HC/FT and Andreessen Horowitz, with OpenAI's startup fund and Kleiner Perkins continuing in. Total raised: $343 million. Headcount: roughly 200. Specialties supported: 100-plus, which is, depending on who you ask, either ambitious or slightly absurd. So far it appears to be the former.
The product, in plain English
A microphone in the exam room. A model listening. A structured clinical note appearing in the EHR roughly twenty seconds after the visit ends - with the diagnosis codes, the assessment, the plan, the patient summary, and the referral letter already drafted.
The problem they saw
Doctors in the United States spend roughly two hours on documentation for every one hour they spend with patients. They call the after-hours typing "pajama time." It is the largest single contributor to clinician burnout, which is itself the largest single contributor to clinicians leaving the field. The ledger is well known. Nobody had been able to balance it.
EHRs - Epic, Cerner, athenahealth - solved billing. They did not solve typing. Voice dictation made transcription faster but still left the doctor at the keyboard at 9 p.m. Human scribes worked, sort of, until the math of paying a human to follow a doctor around stopped working.
The founders' bet
Mike Ng and Nikhil Buduma met at MIT in 2013. They had a thing in common that nobody puts on a pitch deck - both had lost someone they loved to cancer. Ng went to business school intending to work in healthcare. Buduma went into deep learning research and, before he was 25, co-authored an O'Reilly textbook titled Fundamentals of Deep Learning. They co-founded Ambience in 2020.
The bet was unfashionable at the time. Most AI in healthcare in 2020 was being aimed at radiology, drug discovery, or chatbots. Documentation was considered boring. Boring, it turned out, was where the money was - and where the doctors were drowning.
Why the early bet aged well
By the time transformers got good enough to write clinically usable prose, Ambience had spent three years inside hospitals figuring out which prose hospitals would actually accept. Most competitors built the model first and looked for the workflow later. Ambience did it in the other order.
The product
The flagship is AutoScribe - the ambient scribe that turns the visit into a structured note. Around it Ambience has stacked an entire post-visit suite. AutoCDI checks the note against ICD-10 and CPT requirements in real time, the rare AI product that pays for itself by closing audit risk. AutoAVS writes the patient-facing summary. AutoRefer drafts the specialist letter. AutoPrep reads the chart before the doctor walks in. And the newest module - Chart Awareness, launched February 2026 - grounds every note in the patient's longitudinal record so the AI stops re-introducing the patient to herself.
Pricing is per provider per year, in the low thousands for the scribe and the mid thousands for the full stack. Health systems do the math against a single avoided burnout-driven physician departure and the conversation usually ends quickly.
The Ambience Timeline
The proof
Talking to a health-system CIO about AI in 2024 was an exercise in deferred enthusiasm. Talking to one in 2026 is different. The pilots have ended. The contracts are signed. Ambience is now embedded across outpatient clinics, emergency departments, and inpatient units at health systems that, together, treat tens of millions of Americans a year.
Funding rounds, in millions
The mission
The official mission is short - eliminate administrative burden in healthcare so clinicians can focus on patients. Read it twice and it stops sounding generic. American medicine has spent two decades adding administrative load and one decade noticing the cost. Ambience is, at its core, a company built around the idea that you can subtract instead of add.
Internally the team talks about building an "AI operating system for clinicians" - not a single product but a layer that sits between the doctor and the EHR, doing the work that nobody went to medical school to do. That framing matters because it explains why Ambience keeps shipping modules instead of resting on AutoScribe.
Why it matters tomorrow
Two things will determine whether Ambience becomes generational or merely successful. The first is whether ambient AI moves from documentation into diagnosis-adjacent territory - decision support, longitudinal patient summarization, automated triage. The Chart Awareness launch suggests they are walking, carefully, toward that line. The second is regulatory. Hospitals are conservative buyers. Regulators are conservative readers. Ambience has so far been disciplined about staying on the documentation side of the wall. The competitive question is whether discipline scales.
It is worth noting the competitors. Abridge raised even more. Suki has been at this longer. Nuance has Microsoft's distribution. DeepScribe, Nabla, Augmedix all want the same shelf in the same hospital IT closet. The market will not have one winner. But it will probably have a top three, and Ambience appears to be in it.
Back to that Tuesday in Cleveland
The physician in our opening scene did not used to drive home at 6:47. She used to log back into Epic at 9:30. She used to send a draft note around midnight. She used to wonder, occasionally and quietly, whether she was still doing the job she trained for. Now she is. Not because medicine got simpler. Because something else took the typing.
That is what Ambience built. It is also what they are still building. Whether a billion-dollar valuation overshoots the moment or just barely catches it depends on the next twelve months. Either way, the keyboard is no longer where American medicine ends its day. That is already worth filing.