Conversations Are the Source of Truth in Healthcare with Abridge CEO Shiv Rao
38:42 Share

Oh, no.

Hi, listeners, and welcome to No Pryors. This week, we're speaking to Shiv Rao, CEO and founder of Abridge, an AI company that processes medical conversations to unburden clinicians from clerical and financial work, allowing them to focus on patient care. A practicing cardiologist at UPMC, Dr. Rao has recently led Abridge to secure a $250 million Series D raise. Join us as we explore how AI is transforming healthcare delivery. Shiv, welcome to No Pryors.

So excited to be here. Thank you, Ilan. Thank you, Sarah. So ABRIDGE has been around for about seven years. Can you tell us a little bit about how the company has evolved over time, what your starting point was and what you're focused on now? Yeah, absolutely. So we started ABRIDGE in 2018. So it's been a minute. And everything that we've been building since then is really based on the same thesis. So that hasn't changed. And the thesis for us in healthcare delivery is

is that we don't think doctors or nurses are going to get fully automated over the next 10 years. And so what's the first signal in healthcare delivery? We think it's a conversation. It's a dialogue between a professional and a patient. And we believe that those dialogues are really upstream of so many workflows in healthcare. And that's where we focus. And so we focus on clerical work first.

But then that's a sort of wedge for us to expand into any number of like different value propositions over time. Could you tell us a little bit more about some of the products that you have currently and how people use them day to day and what sort of customers you work with? Just to give context to our listeners in terms of what business do you have and what do you focus on? I guess sort of starting at the top, what we do is we unburden clinicians from all the clerical work that crushes their souls at night.

And a little bit more color on that. So two out of five doctors don't want to be doctors in the next two to three years. And 27% of nurses per JAMA article that was published last year don't want to be nurses in the next 12 months. And so we have this crazy supply-demand mismatch. It's public health emergency. Patients are having to drive five, six hours from rural health settings to see a clinician in an inner city setting that could save their life.

And so we've got to do something about it. And I think that's where technology has a role that is finally sort of being recognized and acknowledged at the highest level. Like the entire healthcare industry understands now that they just need to find a way to assist, augment and automate any number of different workflows. And so where we come in is that we unburden clinicians from a lot of that clerical work that they hate to do. So they can walk in a room, they can hit a bridge,

have a normal conversation and talk about any number of different topics in whatever order. But when they hit stop and swivel their chair, there are notes there. But it's not the note that you might expect, you know, that my 14 year old daughter could sort of create using, you know, an off the shelf model. It's a note that checks off all the different boxes across not just who the clinician is, what their specialty is, what system they're a part of, who this patient is, what insurance plan do they have and what geography.

not just like the clinical note, but also what the billable note is, if that makes sense. Can you actually explain the difference between those two things, like a clinical versus a billable note? It's a great question. So in this country, we're not compensated as doctors for the care that we deliver. We're compensated for the care that we documented that we deliver.

So every single one of these notes is actually a bill. And that's why there's just like so these are really high stakes artifacts, not just from a clinical communication and patient outcome perspective, but also from a revenue cycle perspective. But I think another key insight for us that served us well for these last several years has been that health care is not homogenous.

And, you know, that healthcare industry umbrella underneath it, on one end of the market spectrum, there's a direct primary care doctor down the street who's taking cash payment out of pocket off the insurance grid. There's an independent PCP, a really small provider groups like mid-market, you know, that kind of stuff. But then on the other end of the spectrum, there are the large health systems. They're the integrated delivery networks, the academic medical centers.

And what we decided to do, and I think what served us incredibly well, is we made the strategic decision years ago to actually run into the hardest part of the market, that large health system out of the market, as opposed to the small practice or the mid-market or the independent, you know, DPC doctor down the street. And the reason why we went there is that the barrier to entry and the barrier to good enough, I should say, is really, really high. And that's where we felt like we could flex.

a lot of our advantages, a lot of our differentiated muscles. We have a lot of science at the center of our company. Our chief technology science officer is this guy named Zach Lipton. He's a professor at Carnegie Mellon. He's full-time with us, but he's been able to recruit...

you know, pretty amazing like team of machine learning engineers and scientists who can really kind of reach their hands deeper down into this stack to be able to sort of meet that bar for all these large health systems where we need to be good enough for not just the individual doctor in whatever specialty, we have to be good enough for all the different doctors and all the different specialties and all the different settings, outpatient, inpatient, urgent care, emergency rooms,

and also in all the different spoken languages. And so the barrier to entry, the bar for good enough is a lot harder. But running into that end of the market allowed us to sort of compete with just pretty much one other company while a lot of the other startups were starting mid-market or down-market individual like primary care doctors with the hope probably over time that they could recruit the people and aggregate the data and do the post-training or whatever else to be able to swim upstream over time.

You're a practicing cardiologist yourself. I'm a little bit curious how that's informed with building this product, what customers to focus on first. I mean, you have kind of like a who's who of customers in terms of Kaiser, Sutter, you know, others. So I'm a little bit curious, like how this has impacted your strategy in terms of you yourself being an MD and physician. So a little bit of a story about the company and myself. Like we started in 2018. Prior to that, I was a corporate VC at a large health system called DPMC. Sorry to hear that.

So I played VC. I was a faux VC. I was a faux investor and put a lot of money into startups, but also a lot of capital into Carnegie Mellon. We started a machine learning and health program. And that's where actually I met Zach, our CTO. And we're not a spinoff. We didn't like spin out of like EPMC. I quit that job to start the company alongside some other folks from Carnegie Mellon. But a couple of lifetimes ago, I went to Carnegie Mellon as an undergrad. In the middle, I became a cardiologist.

And I still see patients. So this last weekend, I was on call in the hospital. I do about one weekend a month. And I'll do every Thursday night, I'm on phone call as well. So just for like emergencies, like heart attacks in the hospital that I need to come in and sort of help address. But it's like an incredible privilege. It's helped us

us not just have this sort of scientific center in our company with folks like Zach, but also have this sort of like for clinicians by clinicians ethos where I think we sort of get workflow and have that domain expertise to not just build that product in a better and more differentiated way, but also kind of understand go to market. Like, how are we going to sequence where we focus over time? And you mentioned some of our health systems like Kaiser and Sutter.

We're alive in over, I think it's like over 110 health systems right now. And the speed with which we've been able to like land these multi-year agreements, I think it's pretty historic. I don't think I've ever seen anything like this. When I was investing at UPMC, if a startup had like a handful of logos a year, it'd be like high fives all around the room, like amazing bottles of champagne. And so this is a really, really different moment right now, I think, AI and healthcare. How would you explain that experience?

I sit on the board of a healthcare technology company now. I believe in this. I'm in this boat. But for over a decade, looking at healthcare technology as another VC on the outside, I

It moves really slowly, right? In general, there are lots of reasons the market has been hard. Like, what do you think is different today? I mean, it's easy to say the abstract level AI, right? But like, how does that play out for your business? A few stars getting aligned at exactly the right time. And one star is like post pandemic, the amount of burnout that was in the that's been in the industry still. And we just sort of like stretched, I think, clinicians so far beyond their limits, that they're leaving the profession. And

And then health systems didn't know what to do. And all of a sudden, so many hospitals were just shutting down because they couldn't staff them anymore.

And so I think, and the cost is sort of like hire another clinician is like close to a million dollars and it takes a long time. And so I think that star is a really important one because people have talked about clinician burnout. People have talked about trying to, you know, create a better user experience in healthcare for, I don't know how many decades. Yeah. It's not new. Yeah. It's not new, you know, but I think it's, it's not, it's not lip service anymore. Now it really, really matters. And yeah,

If that was one star that aligned, I think the other one was generative AI and chat GPT coming out in early 23. So we started in 2018, three months after attention is all you need. And if Zach was here, he'd be very quick to say, well,

Everyone knew about Transformers before that paper came out, and certainly the research community was already interrogating it. But I think when we started the company, a part of what we wanted to do was interrogate all things related to these pre-trained models in healthcare, and specifically in relation to these sorts of workflows, these clerical workflows. And so we published any number of different papers. Zach and team won best paper at EMNLP, I think, in 2021. So we did a lot of really deep research on

But when we started with Bird or BioBird, a Longformer, a Pegasus, all these other pre-trained models, we got to the point where we had a product that worked. And I remember in 2021 and 2022, I think to your point, like we were demoing and it was just like, oh, cool story, bro. Like people would look at the demo and be like, put your hands up. Was that real? And then be like, OK, cool. I'll call you in like five years. It was like it was like, what are we doing here? And really felt like we were eating glass.

But things really started to shift. And I don't think I recognized until 2023 that we were actually pre-selling, you know, the whole time. Like 2021 and 2022, we were like preparing the market. And then when TrackGPT came out, they called us back. You know, all these CIOs and CMIOs called us back and said, oh, I get it. You were talking about generative AI. You had a dinner about generative AI in 2022. I get it now. Like, let's try it. Let's do a pilot. Now, I think where I think we YOLOed it in 2023 is that...

We could have decided to go to the small mid-market or the independent PCP, but we were like, no, let's just let's go to the large academic. Knowing full well that the amount of virality on that end of the market is insane. Like all these CMIOs and CIOs are on WhatsApp groups every single day and talking to each other. And if you screw up with one of those health systems, maybe two of those health systems, you're kind of done for like a couple years probably.

You don't get another shot on goal for a really, really long time. So you got to hit it out of the park. And so we started with University of Kansas Health System and Emory and then Yale, and they were all home runs. And then all of a sudden we saw like we were starting to

Yeah, kind of go viral, if you will, at the enterprise, like executive level across the country. One of the other challenges in that end of the market is there's a lot of incumbency in the existing systems. And, you know, you were on the provider side. You understood this well. How did you think about navigating partnerships and like the systems people already had? Thinking about ecosystems is super important. And really, like the only currency that ends up mattering in health care is trust.

Like, can you somehow find a way to be trustworthy very, very quickly? Because especially on the provider facing side of technology, like this is the stakes are high. Two days ago, I'm just coming back from a red eye from like Vegas, where there was like a big health care conference called HIMSS. And while we were there, we met with an executive at a health system who was sort of asking us about like our stack, our infrastructure, how we're going to be able to scale and like redundancy.

And he was explaining to us that we are now part of his health system's infrastructure. Like we are core infrastructure. So if we go down, the entire health system goes down. They're not making money anymore because I sort of explained that these notes are essentially bills, at least the way that we generate them.

them, thinking really hard about that responsibility and then figuring out if we're going to market on that end of the market, that end of the spectrum, then how do we also sort of partner with the right players, earn their trust of like the right ecosystems so that we can sort of absorb some of that trust?

And it's easier said than done. But in 2022, as an example, like we had won that paper, that EMLP best paper. But in 2022, folks from like large healthcare technology companies that sort of started to take notice, not just because of that, but because of introductions and they had heard that we had something that worked. Now, the you know, who we're who we're competing with is Microsoft.

That's who we essentially almost like always have to do a head to head against. And it's usually like three to four weeks. And then we sort of move on from there. And so far, we've never lost a head to head in these last few years of doing this. But when we kind of like enter into a health system, I think that basically

Being able to demonstrate that you can kind of integrate with their stack is so important. And so we were able to forge relationships with players like Epic, you know, as an example. In 2022, we demoed up and down, I feel like the entire company, and we were able to build trust. And at that point in time, like the large competitor, they had a solution in the space, right?

but it was humans in the loop. So it was really like Indians in Bangalore who are listening to audio and writing the note and Wizard of Boz-ing it back into the medical record. And there was like, it would take time, you know, for all of that to, that workflow to go down. And so that's why people would always ask us like, put your hands up. Is that real? And again, like these weren't even LLMs yet in 2021, we were using like Burt and BioBird and like, you know, all those other pre-trained models and T5 and other sort of summarization techniques.

So when LLMs came out and when we started to really work with them in a serious way in late '22, '23, oh man, like game was totally on for us and we were able to really, you know, take it to the next level. Now that you've gotten to all of these systems, you obviously had to get to a certain quality bar to get deployed anyway. Like, what do you think about what is next in terms of being able to use that scale?

Absolutely. So maybe useful to sort of break down the stack a little bit, and then we can kind of talk about where we're going and where our research team is focused.

So at a really high level, like the core part of this stack is speech recognition. And so that's where we have like an in-house model. It's really a set of models that create best in class output for health care conversation. Can you help us understand that? Because like, you know, an outsider looking at AI and trying all of these, you know, voice based experiences might say, like, it looks like a solved problem. There's an API for that.

Well, there are APIs, but I think if you're really trying to differentiate where like 3%, 5% error rates can make a huge difference. Our ability, for example, to lean into the way a doctor pronounces a new oral oncology drug, an oral oncolytic. And, you know, I'm convinced no doctor knows how to pronounce any of these medications and they all have their own way of saying these drugs. But we have to lean in and actually recognize the way they say them.

And we have to recognize all the different symptoms, medications, diagnoses and procedures across all the different specialties. And we also have to be multilingual because, you know, sort of like a bit of history of the voice game in health care is that before this world of generative AI and conversations and dialogues, there were dictations.

And that's where I would go into a clinic, I'd see a patient, and then afterwards, I'd pick up a dictaphone or maybe my phone, and I'd start to just rattle things off as fast as I could. I'd say like 25-year-old female with a past medical history of diabetes and hypertension who presents with shortness of breath. Next line, next line, capital B, past medical history, colon, next line. You're just kind of go as fast as you possibly can. You're going through like 20, 30 dictations in the course of like 30 minutes.

And it's lossy because what you're dictating off of is chicken scratch, like stuff that you wrote on a piece of paper while you were in the room. And, you know, later that day or maybe that night, and doctors call this pajama time, you're hoping that you'll remember the details. Sometimes I would write on a piece of paper, tall guy in the meds hat.

And that was supposed to trigger all my memories around who that tall guy was and what his like symptoms were. And then it would like start to mesh with another patient who had the same symptoms. It's not encouraging. Yeah. Not encouraging. Not good for doctors. Terrible for patients. Not good for like revenue cycle or billing. So lossy, you know, and...

So I think in this new world, what we have to do is recognize all those words, those medicalese, like all those medical terms. We also have to recognize all the different languages because it's not a dictation. It's not a monologue. You have to lean into whatever the patient speaks. And so today in California, we'll probably do 50,000 conversations at least in Vietnamese and Haitian Creole today in French.

Boston will do thousands of conversations in Brazilian Portuguese, in Spanish. Today in Indiana, there's a doctor who's speaking in Punjabi to her truck driver patient population at Reed Health. But regardless of what language anyone speaks, our job is to create the note in English within seconds and put it right into the medical record and all the different discrete fields for them to trust and verify.

So part of what we do on the speech recognition side is we're like sampling the audio so that you can have these polyglot conversations where you're speaking in like 10 languages in the same conversation. Not that that has ever happened, but we'll still do a good job because we've been able to bias the model towards whatever the language is that that we're hearing at any given time.

But then obviously we're on this treadmill of always improving, always recognizing the latest FDA approved drug or the latest pronunciation. That's just speech recognition. I think as we move past speech recognition in like the core part of our stack, you start to get into all the text and language.

work that we do. So there are models that in a sense sort of abridge the conversation where we're trying to like distill what would the doctor need to communicate to other doctors and nurses? What would the doctor need to communicate with the patient? Because that's also like an artifact that's created. It's called an after visit summary. And then what would the doctor need to create for revenue cycle? Because these are bills.

And I think a part of the reason why clinicians have burned out or burning out is that they're serving multiple stakeholders all the time. And so it's really hard for them to sort of focus on the one person they went to medical school or nursing school to actually serve the patient. Instead, they're always thinking in the back of their head, what would a RevCycle person think of this note?

oh, I'm going to get a bunch of emails for how crappy this thing is, or I didn't like elucidate exactly, you know, where the symptom was or what the differential diagnosis was. And so that's part of the challenge. And that's what we're doing in the background. And obviously these are like agentic systems in the background that are listening for all the right things.

And so distilling and then structuring data. So those are information extraction models where we pull out those symptoms, medications, diagnoses, procedures. We map them to data dictionaries. And then, of course, there's summarization. And the way you summarize for anyone looks different. So if I wrote a note as a cardiologist and in my note I wrote transcatheter aortic valvuloplasty as a recommendation for my patients.

And then my patient sees that term and I never said that to them. Understandably, I'm going to get like blown up, right? I'm going to get emails and phone calls asking what was that term? You never said it. I looked it up. It sounds scary. And so what we can do is do that sort of style transfer across all the stakeholders that clinicians serve and sort of meet all of their different needs.

And that, I think, you know, has allowed us to serve the executives, like the buyer personas in large health systems. Think about what's next and like, you know, greater ambition for a bridge. Do you have to choose to go down one of those paths first in terms of that translation or you just choose like totally different clerical workflows? I think it comes back to that thesis. And so if you really, you know, believe, as we do, that healthcare is about conversations, that it's like one of the first things

original signals in healthcare, then you start to see that any number of different workflows are beyond it. It's not just clinical notes, it's also orders. After I see a patient, I might say to my patient, like, let's start you on metoprolol or let's get a CT scan.

And so we talked about an order so we can distill, we can extract those orders, we can structure them and we can place them in the medical record. What's after orders is a claim, is a code, is a bill that goes to the insurance company. There's all things revenue cycle. There are clinical trials that come up in a conversation as well. Whether I know it or not, maybe this patient in front of me has inclusion and exclusion criteria for some trial that could save their life.

And so what if some, I had the superhero power and in the moment at the point of care, I was being told by a technology at the right time, like, hey, Shiv, like this patient in front of you has inclusion and exclusion criteria for something that could save their life. Do you want to bring it up? Here's the information. So that's another sort of aspect of where we're going already. But then there's clinical decision support. And so in many ways, I'd say, oh,

Clinicians, I think, you know, they see that as the real holy grail, where what if we could not just sort of level out or like raise the bar on the quality of documentation and billing and revenue cycle, but what if we could raise the bar on the quality of decision making?

What if at the point of care, a bridge could say, hey, shit, like this patient in front of you, like Sarah, actually, she looks like 10,000 other patients in California that have been seen in the last few weeks. And for them, people have decided that this is amyloidosis and not sarcoidosis. And you should skip to the cardiac MRI and not screw around with the CT scan and also maybe consider this therapy and look into this New England Journal of Medicine study to get

you know, more inspiring or, you know, get more insights into what the differential diagnosis could be like. That's, I think, a big part of what we're pushing in the infrastructure that we're building is all going to, you know, amount to. I think you have a really unique perspective as a clinician, a cardiologist, an AI entrepreneur, you know, somebody who's actually operating at scale in terms of the application of technology to healthcare. I'm a little bit curious how you think about the impact

of AI more generally to healthcare? Is it anybody can log into a website and access the equivalent of the world's best doctor? Is it tooling for physicians in really rich ways? Is it, to your point, sort of mining the porpoise of everything that's happened

to people seeking out healthcare and then providing recommendations. I'm sort of a little bit curious, like, what is the big picture view of where all this is heading and on what timeframe? I think it's all of the above, but like the timeframe piece is the key thing. And obviously like all of our time machines are broken right now. And it's hard to predict where we're going to be in, you know, even a year or six months, you know, with how fast things are moving. But,

So much of what I was describing earlier around what conversations are upstream of, I used to think that was like a three-year roadmap. And we're building all of that right now at the same time and deploying it across all of our health system customers and learning already. And being at scale, by the way, and I think maybe we were getting at this earlier, is really magical now. We're live. We're doing millions of conversations online.

like every couple of days, like it's real scale. And with every single one of these notes that are generated, we're getting edits. And so it's fascinating. Like we have this contextual reasoning engine, we call it, that's sort of pulling in information, not just from the conversation,

like the core stack that I was describing earlier, but we pull information from other sources, disparate sources, not just like the clinical system, like the electronic medical system, like the past medical history or the problem list that the patient has. We're pulling information from insurance systems. We're pulling them from clinical textbooks. And so all of that information is sort of orchestrated together in the right way, in the right order, so that we can generate like the best possible artifact.

And I think where we are now is that

Those best possible artifacts still get edited, you know, and we're seeing in the metrics that we use these validated instruments that we're reducing cognitive burden by like 60% within six weeks of a clinician using this. And clinician burnout per one survey that Stanford came up with, we reduced that by like 50% sometimes in the first couple of months. And like no technology has ever done this in healthcare, like had that kind of impact.

So it's a pretty awesome moment. But now that we're getting these edits, because like nothing's perfect, you know, and we don't claim to be perfect at all. Like we absolutely, we're creating drafts that people can kind of leverage and take from there. But we save them hours a day with these drafts. But now that we have these edits, we really like we're going to town on all things related to

you know, post-training. And for us, it's like preference tuning, like DPO and reward modeling and reinforcement learning. And I think having this incredible amount of feedback coming in on a daily basis means that we're always, at least, you know, in our estimation, like getting less imperfect. Even if we're never going to be absolutely perfect, we're getting less imperfect and it's worth it. Like it matters at that end of the spectrum in healthcare.

So that's sort of like the big game for us. Yeah, part of the basis of my question was, you know, I started a digital health company 10, 11, 12 years ago. Yeah. So a long time ago. And what I've observed is the technology cycles are really slow in health care. So they're always a decade behind at least. And this is an odd example where actually certain health systems are ahead by using a bridge.

And relatedly, if you go back and you look at some of the early research, you know, MedPalm 2, for example, came out, I don't know, what, two, three years ago now? Yeah. On the older Palm models. And even then, it provided output that outperformed physicians in terms of predictability of a disease state or other aspects of care, but it never really got adopted. And so I'm a little bit curious about the adoption curve versus the technology curve, because the technology curve is clearly there. The adoption curve is starting through things like a bridge. Yeah.

But there do seem to be these almost like systemic obstacles to adoption of new technology in healthcare. I totally agree. And I think it's like finding the right wedge is so important. There's some kind of two by two that's always in my head. Like when you have high stakes and like high frequency workflows, like that's probably not going to get

absorbed into like the healthcare system proper very, very quickly. But when it's lower stakes, higher frequency, like our, you know, workflow, because there is that clinician in the loop who's making those edits, making sure that things look right. I think that there's an incredible moment right now. Like the window is open, especially if you can demonstrate increase in productivity, improvement in user experiences for doctors and for patients.

And, you know, the biggest deal for us increasingly is like we're also talking to the CFO at these health systems and demonstrating that if you used some other technology that didn't put all that work into this orchestration of different models, that you'd actually be losing money.

And with us, actually, you're getting full credit for the care that you delivered. In relation to your point, so this last weekend while I was on call, I used GBT for a lot of my different patients and I played with Claude too. And what I would do is sort of try to distill the call that I got, the patient that I was about to see. And I would put in, like prompt all these different models and ask it, like, what do you think I should do next? Or what's the differential diagnosis? Or do you agree with this treatment plan? Yeah.

And oftentimes I'd say it was like 100 percent correct off the bat. But maybe just as often, though, it was like a dialectical experience where like two hands to clap, like it would be me and it going back and forth like three or four times before we got to something that really was the right thing.

And the art was like getting it there or getting there together with it. And so I think that clinicians, I think medical trainees, residents, medical students, you know, like they're figuring this out faster than maybe the older generation of attending doctors and consultants out there. So I'm super optimistic that as those clinicians sort of like mature in their careers, it's going to be like game on and they're going to all be leveraging this technology to be even better.

And there's also no question in time, like I think, you know, in any of our minds that like this technology is going to get to the point where it's going to be able to like take on some aspect of care. But I think when most of us get sick, but you can disagree if you don't agree with me, but I think when most of us get really sick, we're probably still going to want to see, you know, a real life doctor to sort of parse through information and use tools like this to figure out what the care plan is. Can I ask one last question on just how...

how like product and engineering and research work at a bridge just because you're deep into the journey, you're at scale in a way few people are with these AI applications now. You mentioned you run headlong into the really tough piece of the market where the scope is large and the quality bar is high. And yet, like today and forever, the product will be imperfect. Yeah.

How did you think about what was good enough, what the minimum viable quality is? And you continue delivering more and how to communicate that or negotiate that with users. So for us, on that hard end of the market, we're always threading a needle through a few different buyer personas and then the end users.

And so on the buyer persona side, there's the CMIO, the chief medical information officer. That's the person who sort of represents all the doctors and the nurses inside the system. Then there's the CIO, the chief information officer. And so that person is representing sort of like the long-term technology investments for the system. They're worried about sunk costs. They're worried about integrating with existing stacks. They don't want to have too many apps inside their ecosystem. Microsoft is probably like something that will never get

them fired. And so there are a certain set of challenges there. And then there's a CFO and the CFO just wants to make sure that there's actual real tangible ROI. And so for us, we knew like in early 2023, we couldn't check off, we couldn't run the table on all three, but we could do two out of three. And that was enough for us. We were like CMIO, CIO, awesome, let's go. And then the CMIO, the big challenge was, could we serve all the different specialties?

And I'm a cardiologist. My note, my output looks so different than an oncologist. We just announced Sloan Kettering yesterday. And their notes at Sloan Kettering look so different than a primary care doctor's note or like a surgeon's note. And so all these different specialties, there's different stylistic sort of preferences. There's different structures to the note.

there's different content that actually gets pulled into the note and there's different workflows. Like in the emergency department, you go into one room and I don't know if you're watching like pit, but it's actually like a pretty real, I think like you go into one room and then you're like,

patient to another, and then you go back into room one, and then you order an x-ray, and then you go to like room three, and you come back to one. And so what we had to do was figure out a workflow for the emergency department where we could stitch together these discontinuous conversations. And now we're doing that, like we're working to do that for the broader care team, stitch together all their conversations around one patient to create like one set of artifacts for that whole encounter.

So I think like that was the barrier to entry for us. It sounds like a big barrier. Did you have that? Do you bring that expertise in-house? Are you just working really closely with customers? Like you are not every version of that doctor. We have some people we call mutants in our company who are doctors who are also engineers.

Like we had this one, for example, we have an engineer who was like a principal engineer at like Meta, who's also a clinician. We have doctors who are like in the weeds of like just prompt engineering on a daily basis.

But then we have others that can go like even more scientific. We have others that also work on other aspects of like partner success or go to market as well. And so I think we try to find those interesting combinations of people because it helps us go faster. They're having like interdisciplinary and multidisciplinary meetings in their own mind. And we just don't have to, we can just like skip steps, I think, with those folks sometimes. But in general, I'd say like where we've like,

really invested. We've raised over $500 million now. And so where is that capital going? I think so much of it, 80% of it should continue to go into R&D. And so it's just figuring out what's next on this roadmap? What else can we build? And our ability to reach down lower into the stack and also get into new workflows and user experiences at the top, I think is

served us really well. You were having this very successful career in corporate venture. Prior to that, you know, you continued to practice medicine throughout. And then you decided to take this giant leap and start this company. What prompted that? And how did how did that bridge come together? Like late 2017, like it was clear already, like deep learning was starting to take off, at least on the research side of things.

with computer vision and there were a lot of companies out there doing interesting things in like the CT scan world, for example, detecting pneumothoraces or being able to predict benign versus malignant like nodules on a scan. And I think now actually that stuff is starting to take off in like a more real way. It's going to be exciting to see where those technologies go, those products go.

But at the time it was clear there was like something out there that we could do. And I think that this idea, once we saw it and once I saw it, it was like hard to unsee. And it was really easily like easy to get like super obsessed about it. Interestingly, like we knew when we started the company that we wanted to serve both sides of the story.

And there's a professional side of this, but like always keeping that patient in mind and like thinking about that bigger system was also a big deal for us. And, you know, in terms of thinking about not just the professional, the doctor, because that's my professional pain point, and also thinking about patients, I saw this one patient in clinic in March of 2018.

And she had a 10-year history of breast cancer. And she was starting to see me because she was just prescribed doxorubicin, which is a chemotherapy that can affect your heart muscles. So she needed the clearance from somebody in cardiology to move forward with that chemo regimen.

And she was super nervous and anxious, like crawling out of her skin the whole time I was with her in the exam room. And so at the end, I asked her why and if there was something I did or something I said. And she told me that for the last 10 years since she was diagnosed with breast cancer, her husband would come to every single visit with a new type of doctor. And he couldn't come this time for whatever reason. And so I asked her, what does he do that's not obvious? And she told me that he sits in the corner. He's quiet. He just takes notes.

And she's an English professor at the University of Pittsburgh and allows us to tell this story. But she told me that him taking notes for her meant that she could feel more present with me and she could make eye contact and she could build a relationship. And then they could go home and unpack all of his notes and rewrite them in words they understood and then go to the next doctor.

and feel like the main characters as opposed to someone looking in from the outside. And so much of her story on that patient side of the room, that story is about agency. It's about ownership. It's about control. And I think so much of the story on the clinician side, on the doctor side, is about agency. There's an American Journal of General Internal Medicine article from last year that suggests that doctors need 30 hours a day to get all of their work done. And they broke down where all that time needs to go. And so you're always paying debt

on work, you're never able to get ahead. And so you don't have any control over your time. And that's why they call this pajama time, this affliction where doctors are writing notes after dinner or after their kids are in bed or whatever it is.

And so finding a way to thread that needle as contrived as it might sound, but like build that bridge between like the two people, the doctor and the patient or the nurse and the patient, people who matter most in healthcare is really like what we're aspiring to do. And now we're doing it. Like maybe one last thing I'll leave you with. So we use Slack as a company and inside of Slack, we have a channel called Love Stories.

And so every day we're getting feedback from our doctors across the country, like feedback in droves. And I think it's pretty heroic in general for a doctor to give you feedback like, hey, this sucked and you got to do better. Like, you didn't recognize the way I said this medication or I'm a gastroenterologist and I would never, you know, sequence my problems in my assessment and plan section of my note this way. It doesn't serve me well and makes me look like terrible as a doctor or whatever. We get that feedback. We love it. It's oxygen. Right.

But then we also get the feedback that's like, hey, this is amazing and I'm not going to retire anymore. And I've got like years, decades left in my career now, thanks to this technology. But in this channel, Love Stories, all of that feedback, that positive feedback, we just get it like programmatically funneled. So any one of our people inside of the company can always go into that channel. And it's like purpose, you know, it's like fulfillment immediately. Like you immediately understand why we're all working so hard and why it makes sense. Because like being on this

very telephone pole like journey these last couple of years is obviously like it's news for so many of us and we're all kind of building new muscles, but it's a lot of pressure.

But this is my favorite bit of feedback. So this love story comes from a doctor at Tanner Health, which is a rural health system. And she wrote to us. She wrote, I was sitting at dinner last week and my son asked me, Mommy, why aren't you working right now? I literally took my phone out and explained to him that a bridge is a new tool that lets mommy come home early and eat dinner with her family. I started to tear up and looked over at my husband who then said, Mommy's going to be able to eat dinner with us every night now.

And we get feedback like that, like every day, you know? And so like there's dopamine hits, you know, in hypergrowth. And like those are awesome. But I think that they get us through like sprints.

But I think it's the oxytocin hits like this. It's the purpose. It's the fulfillment. It's like that's, I think, what I think we're really after in this company. And so like everybody's mission driven out there. But I think this mission, like it hits me at least a little bit different. Me too. You know, congratulations on all the amazing progress with Abridged, Shiv, and Keep Climbing. Awesome. Thanks so much, Sarah. Thank you a lot. Thank you.

Find us on Twitter at NoPriorsPod. Subscribe to our YouTube channel if you want to see our faces. Follow the show on Apple Podcasts, Spotify, or wherever you listen. That way you get a new episode every week. And sign up for emails or find transcripts for every episode at no-priors.com.