Putting Providers First: Paul Brient on Making Innovation Easy and Open at athenahealth

(upbeat music) - Welcome back to the Being Capital Healthcare AI Podcast. I'm Paul and I'm here with my co-host, Devin. - Thanks Paul, really excited to have Paul Bryant with us. Paul Bryant is the Chief Product Officer of Athena Health. Athena Health is actually an investment that we made at Being Capital in 2022. So one of our portfolio companies and one of the market leaders in the electronic medical record space and leader in developing technology for this intersection between patients and providers. In fact, Athena Health serves about 160,000 healthcare providers and touches about a quarter of the US patient population every year. - Totally agree, it was a fascinating discussion. I think what's so interesting about Paul is that he's really one of the pioneers of the healthcare technology industry. He's been in it for more than 30 years. Before he was in Athena Health, he was actually the CEO of PatientKeeper, which was really the first company to create software that actually helps physicians to interact directly with their patients to drive proactive care management. He also comes from a deep medical background and a family of physicians. And so has been thinking about how to improve this clinical experience for a long time. He's now bringing all that experience and insight to really bringing AI to the forefront of Athena Health to drive automation. Not only in the clinical workflows, but the financial workflows of the provider and being able to free up the time that physicians have to really reinvest back into patient care. - Totally agree. And I found it really refreshing Paul's take on the complexity and challenges that are faced with developing technology for this space. It brings a real reality to how that can be achieved. And I think he has a real commitment to making physicians and clinicians lives better and ultimately making care better for patients. So with that, let's jump in with Paul Bryan. - Paul, thanks so much for being with us today. Really excited to talk to you about Athena Health and everything that Athena is doing to transform the future of AI in healthcare and extend the power of AI to healthcare providers. It'd be great to start by giving our listeners a bit of background. So I'd love to hear just a little bit about yourself, your long history in the healthcare technology industry and how you came to Athena. - Great Paul, thanks. It's exciting to be here. I've pretty much spent my entire career in this healthcare IT ecosystem very early in my life. I wrote a PC-based practice management system on an Apple 2E computer, which I would not recommend trying, but was kind of what we did back then. And since then, spent time at the Boston Consulting Group working in their healthcare practice and then joined a company called HPR, which we sold to HPOC, which then sold to McKesson. And then up running all the payer businesses for McKesson, which was a ton of fun, although it was a time of crisis because of the HPOC acquisition, which I won't bore you with the details. And then from there, I joined a pre-revenue company called Patient Keeper focused on building technology that physicians would actually use as they care for their patients. We built that company, had a bit of a hiccup when ARA and the mandate for EMRs was passed, but got through that. And then ultimately sold the company to HCA. And I was the executive at HCA for five years. And then Athena was acquired and I joined the management team that helped turn Athena around and here I am today. - Yeah, Paul, I think you've had one of the most interesting backgrounds I've seen in the broader healthcare IT space at this intersection of patients and providers. I'd be curious to hear just from your pre-Athena experience, were there a couple moments or a couple of things that you learned or took away from all your prior experience that you helped kind of shape what you're doing today at Athena? - Yeah, I'll have to reflect on the fact that I probably use every single one of the things that I learned as part of my experiences to help with Athena. Athena is a very complicated and scale player servicing virtually every provider specialty and servicing the payer businesses as well. So yeah, I think the thing that I bring that I'm most proud of is focus on the physician workflow and kind of a passion for the provider and enabling them and freeing them to care for their patients in the most effective way possible. That's an understanding that you can really only get by being shoulder to shoulder with providers. Not a doctor, most of the people at Athena are not doctors, yet we're responsible for building software that physicians use. And that appreciation is something that I'm pretty proud of. It started when I was a kid, rounding with my dad. He was a general surgeon. This was pre-HIPAA, pre-whatever. I used to literally go to the hospital and round with him and I'd go into surgery with him sometimes. I joke that I can do just about everything in a hospital except for actually care for patients 'cause I just know the rhythm of the hospital. I spent a lot of time even at Athena going out to see a client this evening and tomorrow I will spend a couple hours with providers as they care for their patients. Shoulder to shoulder with them, seeing what they do, see how they use our software. And I think that perspective is really important given what we do. - Clearly, you've had a lot of experience. And I'm just curious, when you sort of sit here today and reflect on all the both opportunities and challenges that exist in healthcare and healthcare IT, what do you think are the biggest pain points or friction points that you see, particularly as between patients and providers, but more broadly as well? - Yeah, the US healthcare system is unbelievably complicated. And even as someone who's sent their whole career in it, when I go to access the healthcare system, I'm always stymied by something. Yeah, I think the requirement that the EMR, the physicians use EMRs, which is an interesting thing to have happen, made the EMR a barrier for a period of time for providing care. If you went and talked to physicians in 2015 or 2020 or even some cases today, they would say this EMR thing has really hurt their productivity and it's not helping them. We at Athena are working really hard to make that not the case. I think we've made a lot of great things for that. Now, I think on the patient side, I think the thing that continues to frustrate patients is access and financial. It's like, you go to your doctor, you have no idea how much it's gonna cost, right? Where else in any sort of market economy or any sort of anything where you go and get services and have literally no idea how much you can cost. It could be zero, it could be a thousand bucks. You know, despite lots of attempts and lots of technology, we still have not solved that particular problem. I think it's unfair to patients and I think it is obviously very frustrating for as a patient and we're working hard to solve that but it's a more difficult problem to solve than you might imagine. - One thing that we wanted to chat about is just, Gen AI, it's obviously getting a lot of press. I think there's a pretty widely held view that Gen AI may have the most significant impact on healthcare of all the sectors of the economy. One, because of the significant amount of administrative and efficiency that exists today across the ecosystem for all the different stakeholders but also too for the promise that it holds to deliver better quality, costs and access to care. And so I know you've been giving a lot of thought personally to how Athena can leverage the power of AI, not just internally but also leveraging the broader power of all the innovation in the ecosystem. I'd be curious, how do you think about the role that Athena can play in bringing AI to physicians and to patients and creating an open marketplace for AI innovation to take place the next couple of years? - Still a very interesting question, one that I spent good chunk of every day thinking about. Fundamentally speaking, if you look at healthcare, why is healthcare so ripe for AI impact? It's because in many ways we're still stuck in the 1990s. So much of healthcare is done via mail or faxes and telephone calls, unstructured data, not leveraging just current state of the art technology for lots of reasons. And AI does this really cool thing is it can read and it can interpret and it can fill in forms. So it can actually do a bunch of things that in many industries already went away and that we already have computer to computer interactions but in healthcare we don't. So it's particularly ripe I think for helping us get healthcare into the 2020s. The other thing that it's really helpful for is digesting all this information and summarizing it in a way that's relevant for the user. And in our case, the primary user being the provider. You know, as we went to EHRs and we did interoperability and we got all this information and we stuck it in a computer, we burdened the physician with it. If you think about the way the world was back before EMRs back in the actual 90s, now you had a nice paper chart in your office that had just the relevant information about the patient in that chart. And the patient comes to you and you look and there's the relevant information and you make a few notes about the next time and you'd move on. And today you have a computer full of all the information about that patient, which is much better potential care. But if that's stuck in a 75 page CCD in your EMR that you have to go read, you're not gonna go read it. Jenny, I will do, which is awesome, is take all that, it's adjusted, figure out what's relevant for you and give it to you. That's incredible. And the other thing it will do is it will listen to your actual visit. And you're not allowed to just jot a few notes to yourself anymore, you have to write a really nice note that is medical legal context, it's billing context, all these different reasons why you have a note. And the AI will write that note for you in a much better way than you do today. And that for many physicians is unbelievably liberating. Soon it will extract all the codes out of that encounter. So give you your diagnosis codes, give you your CPT codes, any orders that you suggested to the patient, and I'm gonna order you this drug, or I'm gonna order you a referral to physical therapy. It will place those orders for you, or at least tease them up so you can place them with a single click. And really transition the way you worked in the EMR to basically reviewing what the AI has done for you and having you make adjustments and move on. That is tremendous. I think I'm one of the few chief product officers in the world that spends most of their day trying to think about how we can get our users to use our system less. But that really is the goal, is to enable physicians to care for their patients and have the computer support them in the least invasive and least time consuming way. 'Cause every minute you're spending in our application is a minute you're not face to face with your patient. - I think a lot of what you just said would be music to a lot of physicians' ears. I think for the last 15 years, EMRs have been mostly viewed as a statutory requirement, an important repository of data, but not necessarily something that overly joys their users. They really wanna be providing patient care and going back to providing medicine the way that they did 20 years ago, as you mentioned. I'm curious as you think about the future role of a doctor in a health system, in a physician practice, with all the time that all these benefits will unlock, how do you think that their role changes over time? - That's a very, very, very cool question with a lot of different answers to it. I think there's certainly a problem right now of a physician burnout where physicians are spending a whole bunch of their time after hours trying to do all the administrative work they couldn't get done 'cause they're seeing lots and lots of patients. And I think at first order, what you'll see is basically they'll get that time back. They'll go to the soccer game. They'll be able to go spend time with their family or do whatever they like to do in their spare time. So I think first order, you won't see a massive impact on care delivery necessarily. You'll see a happier physicians, which is a really good thing. I think though the interesting thing becomes, hey, as AI starts to do some of the work or really automate some of the work, what does that mean to provide care? Even today, if you go into chat GBT and you describe a set of symptoms, it does a remarkably good job of diagnosis and in some cases, a prescription. If you need physical therapy or you want a really good personal trainer, chat GBT is quite good at that right now. And I think what we'll see is physicians kind of getting back to the way it used to be. I mean, it used to be you didn't go to your doctor and see them for five minutes and move on, which is kind of what we got to here. You might go to your doctor and actually engage with them and they might actually have a relationship with you and they might help you make the changes that you need to make in order to be healthy. And that's kind of the way it used to be. If you got to go back way back to the old family medicine game, right? You had a primary care doctor who might have delivered you as a, you know, when you were born and has taken care of you for your life. And if you're in the hospital, they would go around on you before we had hospitals. And I don't think we'll go back to that world, but we might go back to a world that's actually more personalized, even though perhaps is more AI involved in the care, but it allows the people that are involved in it to have more time to be able to engage with their patients in different ways. I don't know what that'll completely look like, but I'm, when I get excited about it, I kind of think of that becoming a much more personalized healthcare system. - That's a really exciting vision. You know, Paul, thinking about both AI, but even more broadly at Athena, and I think you personally in the company, I've gotten a lot of accolades from the industry around being an open architecture type of system. You have a lot of partners. You have a marketplace. You work with a lot of other technology providers. What has been your guiding principle around how you decide what Athena's gonna build on its own, how you decide when Athena's gonna partner, and also where you might be an acquirer of technology and partners. You're great to hear your thoughts on that. - Yeah, if you look at our philosophy, is we are rabidly an open ecosystem. You know, I think that kind of comes from the core mission of Athena, which is really to help improve the healthcare system. You know, if you believe that it takes a village, and it definitely does, and that we don't have the monopoly on innovation, the right way to do it is to say, hey, let's get as many people innovating as possible. Give Jonathan Bush, our founder, a lot of credit, created this thing called More Disruption Please. That was the name for the marketplace. And the idea was, let's just drop healthcare together. And if I were a startup today, you know, the first place I would go in terms of building my product will be to build it against Athena, both because of the technological support that's available. We have over 800 APIs. You can read, write, data all throughout our system so you can pick where we wanna innovate. But also because of the reach of the system. We have 160,000 providers on the network. Not only do you get the best, most open technology, but you also get access to a whole lot of potential customers. So it really is an incredible ecosystem for innovation. And something that we're very proud of and that our customers benefit from every day. - The more difficult question comes, hey, what new things do you Athena go build? Or what things would you bring out of that marketplace into Athena in some way, either through acquisition or more formal partnerships? - You know, when I first came to Athena almost six years ago, if you'd asked me, I would have said the marketplace is a net good. There's nothing bad about it. There's no pros and cons. It's just both. And then I started meeting with some of our customers. You know, what they said to me was, hey, you know, we love the fact that you're an open ecosystem. We love that there's all this choice in the marketplace, but we don't wanna have 10 marketplace partners. And we kind of looked to you Athena to bring us the system. Like we bought this from you. We'd like ideally to have, you know, very few. You know, we get it where there's some things that are kind of out there. And that was a bit eye-opening for me. Definitely changed my thinking a little bit. To the point where, you know, I think it is our responsibility to define what it is Athena is as a product and what works for kind of the majority of our customers. And to build those products or partner and bring them together into one packet. And then the things I get most excited about on the marketplace side are things that are either unique to a subset of our customers or unique to a specialty or a site of care or a particular workflow that isn't the majority or really cutting edge innovation like AI. There's this, you know, Cambrian explosion of AI companies. There's a gazillions of companies that can't keep track even all the AME no companies can't keep track of all of them. There's so many of them. And that's just innovation, right? It's just, hey, this new thing, lots of innovation. You know, most of those companies will not make it, but some of them will. But all of them are cutting cool new graph. And for us, it's the same kind of deal. It's like the good news is our customers can access any of those companies right away in the marketplace. And we will thoughtfully craft a path for the majority of our customers where we will incorporate AI into our software built in in a thoughtful and lovable way. And I think they'll still be obviously ruined for lots of innovation, but as those things become more mainstream, we will incorporate those into the product in the best way possible. And, you know, our ambient node product is a great example. Three or four years ago, you know, Nuance was basically the only game in town. Technology wasn't really ready. They had humans in the loop building the notes. You know, today there are 50 similar companies that have an ambient node product. And we just announced our own product partnered with three firms. So we were using three firms, but built into Athena, into the workflows sold by Athena. And you can imagine that continuing to evolve as the capabilities continue to evolve. - It's been very clear that Athena has been a market leading company for a long time now. And, you know, part of that, I think at the essence of what Athena does is it's an innovation engine. And when we think about all this evolution of AI and new technologies and rapid adoption, you know, I think that, you know, this innovation in healthcare and healthcare IT should accelerate. What's your model, Paul, for, you know, how you as the center and the leader of these efforts within Athena will continue to innovate, you know, in a world that probably, you know, the pace will kind of accelerate from here. - You know, this is one of the things that's, it's very challenging actually for a large company with a whole lot of customers to do this disruptive innovation. You know, if you're a big fan of the innovator's dilemma, right, it's like, you know, we have a lot of what I'll call incremental innovation that we're asked to do every day. But when the big disruptive things come along, we have to be very disciplined about figuring out the right time to go after them and then the right way to go after them. Because, you know, in general, our customers do not go say, hey, can you please disrupt my life? I like the way it is. I'd like it to be a little bit tweaked. And so it is not easy and there's not a recipe for great success. But the marketplace does help a lot because that's where the disruptive innovation will show up first. Frankly, you know, it's like, hey, new cool idea. It's going to go in the marketplace a lot and we can be observers. That worked. That isn't working. People are excited about that. And then we can elect to either, you know, buy a company, build it ourselves, partner in some way. So we, you know, well, it is maybe very noble that we have an open ecosystem. It is a great source of the disruptive type of innovation 'cause that's where it will likely start. And then we have to make the decision as to, you know, how do we, how do we reflect that or do we reflect that in core Athena for our customers? We rewrote our scheduling system after 25 years of it. The old scheduling system looks like it was written in the early days of the internet because it was. And it's been an incredible journey to get that both rewritten and then adopted by our customers. And if I showed you as a new customer, which one do you pick? You, of course, pick the new one and they do and they like it fine. But there's a lot of change. A lot of, you know, a lot of workflows are built in. And so it is, it is challenging for us to innovate and a lot of things we do is built too deep in the weeds. But, you know, we are a single instance SaaS system and the cool thing that allows us to do is that we can put disruptive change in right next to the old stuff and we can give customers the ability to go back and forth, which is what we did in schedule. Okay, you don't like the new thing today? Great, stay on the old thing. It's fine. Our new customers look at the new thing, they don't even know the old thing exists. Over time, you can choose when to go and in fact in schedule it, you can actually just go back and forth. And so it allows us to bring innovation to customers in a way that is more acceptable to them from a change curve perspective. Cause look, at the end of the day, our customers are trying to care for patients. You know, the technology is a nice tool. You know, when it's all just working, it's almost like, you know, change even for the better. Like please don't move my cheese. Like I'm trying to care for patients over here, don't change that. And so we've got a pretty good paradigm for introducing change into the system without, you know, in some systems, it's like, you know, you take a massive upgrade and your whole world is disrupted for two months. We're not like that because of our Cortex. That's a real advantage that we have. - That's a huge advantage. And, you know, certainly the change management requirements within healthcare are, you know, really important to consider. And I think that you've kind of had a great balance of, you know, disruptive innovation, but also innovation that can be adopted. So congrats and thanks for everything you've done on that dimension. - One thing I think that is really important as we enter this world of very rapid innovation led by Gen AI, you know, really the fuel stock for Gen AI to flourish, ubiquitous access to structured data and ubiquitous access to federated data that exists not in silos, but is broad spread across all the different pieces of the healthcare ecosystem because you have to put that whole mosaic together in order to see the full picture of a patient. One of the biggest hindrances to federating all of a patient's data historically has been interoperability, which I know is a topic that you also think deeply about. I'm curious your perspective on how the government, as well as large players like Athena, can also start to advance the ball forward on interoperability between sites of care to make sure that we actually have the data in place to unlock all these use cases. What are you seeing out in the market and what do you think needs to change? - So it's really interesting. I've been around this industry long enough to have seen all four acronyms in place for interoperability. So we had CHINs, REOs, HIEs, and now we have QHINs. And all of those things are focused around getting the data together and providing access to the data. And you said something in your question, which is really interesting. You said, "Hey, the notion of structured data." One of the really cool things that Gen AI does for us is it does no longer needs to be structured. And a big problem with all of the interoperability to date is that just a bunch of the data isn't structured. And so yeah, you get it, but the only way historically process that data in a thoughtful way is to have the doctor read it isn't realistic. So the cool thing about Gen AI on the data aggregation side is that it can go read all the stuff and it can give you the discrete data out of it. And that's a big step forward. And I think in many ways, I believe that that will solve the, get the full picture of the patient problem. But when you really think about interoperability, there's another element which is even more important, which is, hey, I'm actually caring for the patient and now I need to move that patient around the healthcare ecosystem. And how do I do that and not use a fax machine? Which unfortunately is the way that tends to occur now. I'm gonna send you a referral. I will fax you the referral, like that's crazy. And I think that there's a bunch of work being done now. I think some of the government regulations on the last round, the HDI-1 and 2, where we're starting to get some of these systems that were previously closed, get them to be open is the right answer. And just if you take referral as a great example, right? If you're referring people between two practices that have different EMRs, that's a very difficult thing right now to make electronic. And we're doing some partnerships right now with several of the big acute care EMR systems around a protocol called 360X. It's been actually kind of created within the industry to help solve that problem. And it's not AI enabled, I don't think it needs to be, but it's those kind of workflows that I think are the real data silo and physician frustration or patient frustration type situation where you get referred to someone, you show up, the doctor doesn't have the right stuff. You get referred to someone who's out of network. So you show up and they're like, oh, it's gonna be $5,000. But if you go to some other doctor, it'll be $200. Like just a bunch of things in the day-to-day mechanics of the workflows that are right up. You get someone admitted to the hospital and you wanna have your doctor have to do another HNP. That happens all the time. The doctor prints out the HNP or looks at it on the screen in their ambulatory EMR and dictates it into the hospital system, right? These are crazy workflows that could reasonably easily be solved through interoperability. It definitely seems like there's a great opportunity with better interoperability. I think we're all hopeful for that. We've been talking a lot about if provider efficiency and enhancing the overall patient experience, where's Athena headed on the billing or the revenue cycle, part of your solution? - So Athena is a fairly unique company in the EMR space and that we do a whole bunch of the actual work that if you're a client, you buy an EMR from most vendors, you, the client, have to do a bunch of the work, right? File the claims, post the payments when they come in, follow up on the things that don't get paid, bill the patients, all that. Or you hire someone to go do that as a separate. But Athena, it's all bundled together. So we do a whole bunch of that work. And we do it through a whole lot of automation. So about 80 to 85% of that work is fully automated by technology today, including using some, I'll call it older AI technologies. And then we have an offshore BPO and some onshore BPO staff that do the work that the automation doesn't do. And a lot of that involves picking up the phone and calling payers. A lot of that involves faxing things back and forth. A lot of that involves filling out forms, which we are not, are not automatable today. With AI, a whole chunk of that can be automated and it can be done much more efficiently, much more effectively without having to go through a standards body to go say, "Hey, here's this random form from X, Y, Z, L scale pair "that we have to fill out every time." Okay, well, now the AI can just fill it out and send it. Even by fax. AI can fax too. So we have the opportunity to take the automation level from say, 85% to 95 to 99%. And then what we're planning to do is to take on more of the work, to have people that are more expert humans at the given specialties to be able to do work that we currently ask our practices to do. And I think it's a huge unlock for everybody. I mean, who wants to do this administrative work? And I think the AI is gonna allow us to automate the road work so we can focus on the more complex work that we currently push back to practices. - That's definitely an exciting opportunity. What do you think are the biggest barriers from here to there that you're facing? Is it the right talented engineers, the right data scientists, cooperation with, you know, customers and partners? Like what are the biggest barriers to realizing that vision? - This AI has changed so much in the past 18 months that part of it is just keeping up. Like when we have a project to receive two billion pages of faxes a year, and we process them, we have humans that actually process the stuff. AI does some of it, old AI does. When we look at new AI applied to that, it was gonna cost us about the same money as the humans. And we're like, fine, all in, I'll do a better job. 18 months later, it's about 50 times less expensive. And so the whole thing that we planned just 18 months ago is now very different. Both the capabilities that we can apply and the cost. And so, you know, with the cost being that cheap, it's like, well, we can do a lot more stuff now. We were very constrained before. So, but that involves, you know, it's really annoying when you're, you know, having an engineering project and, you know, the world changes underneath you like every three weeks. How do you keep up? And then also, how do you not get into some weird change loop where, no, no, I can't do this today 'cause the next good thing is coming out. Like at some point you gotta pick it and go. So that's really been, I think the biggest challenge is just the rate of change. You know, you mentioned data scientists, and we had to completely restructure our data science team. We had a centralized, best in class data science organization that frankly was no longer necessary in that form because all the model building that they used to do is now available in the general purpose LLMs. We had a project, we spent a year and a half building a model that would extract lab analytes of a paper fax. And it was a very nice model. But then we took the same fax when we put it into chat GPT and asked it to please summarize the lab analytes in there. And guess what it did? It summarized the lab analytes. I mean, there's like no training. It's like, wow, that's incredible. I'm super excited about that 'cause now we're gonna go use that. But, you know, the world just changed really, really rapidly. And I think we haven't yet fully understood that. I think it's still changing. I think that actually the biggest challenge is just, how do you actually build solid GA quality products in a world that is changing so fast? And, you know, what you built six months ago is now obsolete. So it's an unprecedented time in my career. I've never seen anyone like this before. - Now the rapid change is really amazing. And here, I have to laugh for a second as we're here, we are talking about AI and healthcare and all the exciting things we're doing. And yet it's amazing how often we still mentioned the word facts, you know, what's your prediction, Paul, for, you know, when are we not gonna be talking about faxes in healthcare and healthcare IT? - I don't think, I will see that in the rest of my career, honestly. There are a couple of things that are just so unbelievably appealing about faxes in healthcare. One is they're just not standards for so much information that needs to get exchanged. And so the faxes, the universal standard, all it's gotta do is throw in a piece of paper. The second thing is that faxes are in places and sending documents, at least the way we send them, we send them to people. And if you want to, for example, send something to the fifth floor at the hospital, 'cause that's where the patient is, the fax is actually more convenient than sending it to nurse Smith or whomever who may not be on shift. And that's one of those kind of things you kind of go, ooh, right, paradigm needs to change, right? But the fax is this thing. Everyone goes to the fax machine at the nurse's unit and they're the fax. So there's some kind of built-in things. And then of course there's human behavior, just like, hey, that's how we do it. Now, the cool thing again is with gen AI is that it may not be a problem. Like, right now it's a problem. You know, you receive faxes, it's a pain. If gen AI can take them, OCR them, summarize them, put them where they need to go, take the screw dead out of them, it's like, fine, I don't care, I don't care how it came, it's gonna look like it came in electronically in a discrete form. So it might be that in fact, there's this weird fax infrastructure, but we don't have to worry about it anymore. - That's interesting, that's interesting. I could see that, I could see that. Fax is here to stay, but will be simplified by gen AI. - Yeah, and they'll be all like, hopefully we'll get away from the actual fax machine. - Yeah, that would be good. - As one of the original pioneers of the faxing industry, I'm sure that's really comforting for Devin to hear. Maybe just to close out, I guess, you know, one thing that I think is a bit of a controversy right now in the world of healthcare, IT and also the world of gen AI is that there seems to be both an extraordinary amount of optimism around the application of AI to healthcare. I think we've talked about some of the really amazing use cases, but also an extraordinary amount of skepticism and a belief that there is a lot of overhyping happening, in particular in Silicon Valley, around companies that have raised exceptionally large amounts of money, but haven't really generated as much revenue to show for it. I'm curious how you think about how to parse the hype versus reality as someone who is charged with determining the buy-build partner strategy for Athena, but equally how participants in healthcare should think about parsing the wheat for the chaff? - I'll separate my answer into two pieces. One is kind of on the tech itself, and then maybe the kind of company dynamics and innovation. You know, I've gone from when the LLMs first came out, I was AI skeptic. I was like, "Hey, yep, it's got a lot of hope and promise." It was not right. I am on the opposite of that spectrum from a rock technology impact perspective. It is ready. It is really, really good. I use Gemini and ChatGBT not to try to plug in too, but I use those too, and it's amazing. I use it in personal life, I use it in my professional life. The technology is really good. So the technology is gonna have a very, very big impact. It's gonna transform the way the EMR looks, for sure. The question though is on hype. Like, you know, when the internet came out, we knew the internet was gonna be cool. We didn't know exactly how it was gonna be cool completely. And there were gazillions of companies that have ridiculous valuations that are no more. And I think we have, unfortunately, that going on in spades and company wall. Like, we do not need 50 different companies doing ambient notes. And in fact, the problem is, is that the underlying LLMs are getting good enough, I met a doctor the other day, that does their own ambient note thing with a prompt against Gemini. 'Cause they like their notes a certain way. So they just have a really big, long, nice prompt. They feed in all information and Gemini writes the note for them. So it doesn't take the amount of work it did when Nuance was trying to build it from scratch. Now we have a whole of these companies that early on it was really hard 'cause the LLMs weren't that good. Now the LLMs are really good. The value add part is getting smaller and smaller. And I think that does not bode well for some of these companies. We'll see, but I would hate to be a VC. I'm so happy that I'm the Chief Product Officer at Athena, charged with taking all this cool technology and transforming our product and not having to invest in these other companies and trying to pick winners. 'Cause I think it's really, really hard. - Yeah, I can see that perspective from your vantage point as well for sure. We've been talking a lot about really innovative things that Athena is doing. And there's a lot that's happened in the last year and there's a lot that's been going to happen in the next 12, 18 months. If you kind of zoomed out and said, what's gonna be the biggest change 10 years from now in terms of how a patient experiences healthcare? What do you think that would be? - I think the whole notion of how you receive wealthier delivery will be different. I think it will start with AI. I think that there'll be a lot more availability of things to help you change. So much of being healthy or responding appropriately to an illness or a situation is changing your behavior in some way, taking this medication, eating better, exercising more, getting the interventions you need, getting like, and a lot of that doesn't happen. About 20% of the prescriptions that are written aren't ever filled even, much less people actually taking the meds. And I think we'll see an infrastructure where AI will help us both figure out what's wrong, but also help us do the right thing once we have something wrong to help improve our health through helping us change. You know, AI is getting very heavily on EQ now, people. I predict that people are gonna prefer to talk to AI than other people, which is kind of scary, but it's infinitely patient. It is infinite memory of all the things you've ever told it. And I think it could be pretty powerful in terms of helping people change. And I think we might see a world where we're healthier and happier as a result. I don't know quite what that does to overall society, but we'll see, but certainly on the healthcare delivery side, I think it is, we're gonna go from kind of a very, where experts are scarce to a world where there's a lot more expertise available with quasi-infinite type manpower. And I think that's gonna be a neat unlock for healthcare. - This has been a great conversation, Paul. We really appreciate your time and everything you're doing at Athena Health. Thanks for being with us. - Agreed, thanks so much, Paul. We really appreciate it and good luck with everything. - That was a great discussion. That's where we'll leave it for today. As always, we'll keep exploring what real innovation looks like at the ground level and practice, not just in theory. Thanks again for listening.