Revolutionizing Primary Care with Agentic AI: Dr. David Carmouche of Lumeris

Hey there, welcome to Non-Fungible Human with Dr. O'Aisterani, where we chat with thought leaders, influencers and newbies at the intersection of Web 3.0, the blockchain, and FT's, and life. We're really excited today to be joined by Dr. David Carmouss, who is an internal medicine physician who's helped leadership roles across nearly every corner of healthcare from CMO at Blue Cross, Blue Shield, Louisiana to EVP of Value-Based Care at Oshner Health to SVP of Healthcare Delivery at Walmart Health. He's now the Chief Medical and Chief Commercial Officer at Lumeris, where he's leading the development of Tom, the company's Agentech AI platform designed to fundamentally rethink how primary care is delivered in this country. David, welcome to the show. It's awesome to be here. Thanks for having me. Yeah, I feel like we could have a super long conversation about all of the interesting perspectives and roles you have held. But I'll start with, kind of talking about this moment, right? We're at an interesting time. I think most people realize that we're going to kind of focus on AI. That AI is here. It's already creating meaningful change in a variety of aspects. There is, I think, still a hype aspect to it, so some things are still kind of maybe a little more far-fetched. And I think the theme of, you know, I've been to a few conferences. I'm sure similar to ones you've been to, where Agentech AI comes up everywhere. And so I guess my question for you, and it's probably a question you could spend a whole hour answering, but question is where are we at when it comes to healthcare and AI today in early 2026? Yeah, me. I mean, I'm a, if you like baseball, right? Nine in and game. It feels like it's in the second or third inning. It's pretty early. But there's been a lot that's happened in those two or three things. So, if you go back to, you know, just 2023 and 2023, 24, and just to where we are now, the pace of change is so dizzying that it almost is hard to really gate where we exactly are because every month it feels like it's moved. But I'm with you. I think there's, I don't know if I'll call it hype. It likely is to some degree, but it's also promise. I've seen enough now to know that this technology will be a factor in healthcare. It'll be a factor in all of our lives in many ways. Healthcare is just one example. And for me, it's actually optimistic because the problem I'm working on is literally not solvable with humans. I've been at it for 30 years. You mentioned an interesting degree. And it's the first 15, which is a practice in total medicine on the front line of care. And then the last 15 have been those kind of really amazing leadership opportunities across lots of sectors. And through all of it, right, I've been very interested in how do we extract value from primary care, patient value, outcome value, cost containment value. You know, primary care to me is a linchpin for a high functioning primary healthcare ecosystem. And the reality of it is that the people who've signed up to do that are going away. They're retiring. They're choosing other careers. They're not without making new ones. And yet, you know, there's a hundred million people who today in the United States alone don't seem to have access to a longitudinal first point of care that we would call primary care. And so I, you know, I've fought that for 30 years, largely adding different humans to the mix care team members nurse practitioners and PAs care coordinators, population health managers, those folks are expensive. They don't scale very well. And they largely haven't done much to kind of expand the human-constraighted position him or herself. And so I'm hopeful that this technology will mature to the point where it can be a useful adjunct in primary care so that we can make it available, make primary care available with more people. Yeah, and it's interesting that you say that obviously we've tried to tackle this issue every way that you can imagine. And it hasn't really bulged the lack of primary care physicians, the lack of access. I hate to kind of say that, hey, we've kind of, you know, given up on that side of the equation. But would you say that and give up is maybe the wrong word, but we've reached the extent of, you know, the most we can do from that perspective of trying to get clinicians to go into primary care to try to get people access. And probably not going to make much more meaningful headway from that side of the equation. And we do really need to use these tools and technologies that are now available to try to fix the rest of the gap. Or is there something else that we can do on that other side of the coin of trying to get more clinicians and PCPs and whatnot? I think there's more we could do. I don't think it's either where I think there's a little bit more we could do and we're going to still need technology. The more we could do, I mean, sure as a country, we could decide we want to pay, we want to pay primary care doctors more. We want it to be more valuable than going into subspecial medicine and we're going to invert the kind of payment pyramid and just appeal to economic motives, which I don't think frankly early on in one's career, really the drivers for why you choose fields in medicine. But, but you know, you could you could do more on the on the payment that would have some impact. You could expand who gets to do primary care. I come from Walmart where they were big advocates for expanding the capabilities of pharmacists. You know, we could we could take the reins off of nurse practitioners and others and and allow them to to to be maybe play different roles. And every time we do any of those things, we bump up against either budgetary constraints or challenges or professional societies that are going to get, you know, their their feathers up. And so all of those things that we could do, we could figure out how to make more primary care doctors, maybe shorten medical school to three years for people who want a primary care, it's that or those things we could do. But ultimately, I don't think it's going to be a no. And I think we're going to need these technologies to help. Absolutely. And so speaking of the technologies before we kind of get deeper into your role at Lumeris and the technologies, they are creating their agentic as a big part of that. So to someone who may be listening and there's, you know, medical students and a lot of trainees that listen to these conversations, what is agentic AI and within the realm of healthcare, what is agentic AI going to achieve or will try to achieve? Yeah, I mean, agentic AI is basically, you know, application of artificial intelligence in an autonomous way to achieve a task and to be able to actually reason to accomplish that task. So the way that plays out in our world and in primary care might be that if we have a real time or near real time data driven kind of summary of you as a person, your medical history, your maybe some predictive models that show you're projecting, your trajectory of your history, real time access to maybe sensor data or remote monitoring data. So we actually have as complete a picture as we can from data, maybe whether you filled your last prescriptions, et cetera, all of the data that we could compile and could run that through a model to use AI to discern. Just of all the things we could do to proactively engage you in moving you one step closer to better health and then decide what that is and then delegate that to AI to go do it. And that could be that I'm at home and you're my, I'm your patient and I get a phone call from a name and number that I've saved in my phone. I know it's your AI care team member and it's basically either through a voice conversation or text, whatever my preference is, is basically going to say, hey, it, you know, it looks like you didn't fill your last medication for your diabetes. You know, any, you know, is that, is that true? Have you been able to fill it? No, I, you know, I haven't been able to do it. I, I, I, I, I expect to build this month and, you know, I just haven't been able to fill my medication. Well, you know, have you thought about, do, you know, have you thought about talking to the pharmacist and seeing if you could maybe fill a partial prescription or, or is there anything we could do to maybe double your existing medication to make up for that until you're able to do it. Whatever AI might suggest. And then the patient kind of say, oh, that, I didn't know I could do that. I'm happy to do that. And then AI say, well, okay, well, it's really important. Dr., you know, harm issues is very worried about your, your diabetes. And so stopping your medicine can create problems. I'd like to check back with you in a couple days to make sure you were able to do what we just talked about. And I'd also like to see what your blood sugars are doing at that time. Is that great? What time do you let me call you? AI store that in memory. And then in two days, reach back out to you, pick up the conversation. Were you able to do what we talked about? How are your blood sugars run, etc. That's an example of how, you know, I think of it. Like if I had an unlimited number of like residents or nurses that could just look at every one of my patients every day or every other day or every third day and just decide, what do we want to do to reach out and kind of help them? That's kind of the power of what this technology can do and the skill of this technology. And that's kind of what we talk about deploying a gentigai in a healthcare environment, like primary care. Yeah. And it's really, I really cool. And I think about it from like an ER perspective, right? And I think that's really what I think about the patients that come in for like an issue and then I refer them as specialists and then they unfortunately come back like a week later. And they're more decompantated because the specialists didn't take their insurance or there was no appointment. And I think of like, man, it'd be really cool like where I can say, hey, this patient needs a GI specialist and it automatically like looks at what insurance they have, which GI specialist take that insurance, which ones actually have appointments within like say the three day period, they need that in and it kind of does all of that. And then sends them a reminder to show up to that and do all of those things. And so it excites me because it's going to fill in all of those gaps that exist. It's super exciting. I think the way I like to think about it is even if you are lucky enough to have an awesome primary care doctor, you know, a good friend of mine who's not in medicine, who uses me as kind of his like advisor, his counselor, personal customer. So he'll have his physical and he'll share his labs with me, you know, say, you know, these were my labs and this is what the doctor told me and he said that my last physical doctor can gradually send me some great. He goes, can you tell me what a hemoglobin A1C of 5.8% is? I said, actually, that is abnormal. That is in a range that is approaching what we would call pre-divis. And he's like, well, you know, my mom had diabetes and my sister has diabetes and he didn't even mention it. He said it was in the normal range or he said it was fine. And he goes, and I'm not seeing him again for a year. What am I supposed to do between now and then? Like there's no, like he told me, he checked me, he examined me, he looked at labs, it was a snapshot in time. He says, he'll see me in a year. But I have nothing, like I have no plan. I have no plan to stay healthy between now and then. And I think, so, so even when we see primary care, like it kind of stops outside the walls. And when people are outside the wall, they're kind of left their own devices. And I think if we can figure out how to harness technology and this technology to stay connected to people to help guide them based on an understanding of their health conditions. And by the way, if they start getting off track, alert the clinician. Because as a physician, I only know the information that's put in front of me. And if you're out of sight and you're not calling my office because you're feeling bad, then I'm just going to assume you're okay. And to the next time I happen to see you, I think there's a lot of mismatch of capacity of primary care. There's a lot of folks who are coming in our office on regular intervals who are just fine. And there's people who should be coming in our office who can't get in. And if we could use AI to monitor people to know they're doing okay. So that we would have time to see people who aren't doing okay. I think we would even make more use of the human capital that we have today. And I think that's another real value of this technology that excites me. Yeah, 100%. Because the next year when they see you in a year, they may be in full blown diabetes. And then you're talking about starting them on a medication. And that's like all these missed opportunities where they could have stopped that. And that happens all over health care. And the whole point of primary care literally is to get upstream of all of those problems in complications. And you know, you're working in the university. And you know better than me, you see what happens when there is no primary care, right? When people go forever, they show off when it's a crisis. And you know, in most cases I would imagine, especially as it relates to medical complications, there were lots of opportunities to intervening upstream. And unfortunately you catch it because you're the safety net. And frankly, when all of those systems fail, they show up in the emergency department. And that's a failure of our system. And I think technology's got to be a part of the solution. Yeah, 100%. So speaking of technology, you're at Lou Maris. So tell me, you know, you've obviously had these very interesting roles and starting like you mentioned as a PCP. What stood out to you with Lou Maris and kind of what was happening there? And when it comes to technology where you were like, I want to work here. I want to lead some change here and join this team. Yeah, so it's only been 18 months. And so it's pretty fresh in my mind. You know, I'd come off Walmart, which was, I mean, that's a whole segment. We could do a separate conversation about that. There was three years there. Amazing people, amazing organization, lots of learnings. And for me, because I really thought they had the opportunity to do something really meaningful in primary care, in communities that were underserved. And it was a shame that that ended. But when I was looking about what to do at the next point, the first thing I thought is, I do want to help primary care. And if you look at this country, whether it's Vitterbad, 55% of all primary care physicians in this country work within health systems. They're employed today. And so from a scale and from a efficiency of accessing primary care, you know, going after health systems is a good thing, except they're really hard. They're not many organizations that are born and they set the segment, the customer segment is health systems, big bureaucratic organizations, lots of decision makers, long sales cycles, etc. But Lou Maras had been in the history, you know, it's over its 15 year history, was partnering with health systems and was partnering largely in primary care to help manage value-based care contracts, risk contracts, Medicare, Medicare, Advantage, largely, and then quality performance for Medicaid and commercial. And so, so first of all, Lou Maras had the DNA of knowing how to build relationships and build partnerships with health systems. That was important. Number two is, Lou Maras had become very good at aggregating data. Healthcare data is very messy. It's disparate. It's siloed. at times and putting that data together and normalizing it and cleansing it is really hard, but it's like the foundation for doing AI, right? Clean data, it makes AI plausible and health systems with Lumarice's help have pretty clean data. For me it was like that's a great starting point to think about how we would deploy AI. So that was kind of really the genesis of it. And then, of course, I have to mention John Doerr, who of Clienter Perkins, Silicon Valley fame, who is intimately involved with Lumarice because his brother Tom Doerr, who was an internist in St. Louis, was co-founder of the company and John is an investor and a board member in the company. John, and I had a conversation fairly early on, I got to meet him when I was at Walmart. And when he saw that Walmart was shutting down, one of the conversations with him was, "Hey, I'd like you to come take a look at Lumarice." And I'm like, "Yeah, I know, I've done a lot of value-based care in my life. I'm not sure that's the next chapter." And he's like, "No, no, no, no, man." And he goes, "Have you heard of Sam Aldman?" And do you know what Chad G.P.T. is? And he goes, "That's what we want to do. We want to harness the power of this new technology. Trust me as a technology investor. This is as exciting a time as it was when the internet was being born." And I think the technology is going to be amazing, but it will yield no value to this country and healthcare. If we don't have really smart people who understand healthcare, help us figure out how to smartly deploy it around the right things. And I'd like you to come and do that. So for me, it was like that allure plus its DNA and the fact that they learned how to manage data. And then, frankly, they also kind of already aggregated lots of really smart technologists and other healthcare leaders that I really enjoyed interacting with. And so for me, it was kind of a natural to come, to join them. Yeah, and looking at it from an external point of view, I think that makes a lot of sense because there's like a lot of like tech first companies that have been born in the last few years and they're trying to enter healthcare, which can be very difficult because healthcare is a very rigid environment. But if you've been in that arena for years and years and years and you know what exists and what the flaws are and the roadblocks, and then you add the tech on top of it, it seems like it may be a little easier to kind of solve some of those problems. Yeah, it's not easier. I don't think I mean. Yeah, it's never easy. But what did it, so I will say this and I think you probably you know, talked about this at times. The technology in and of itself is valuable only when it exists within workflows that drive outcomes. And theoretically, it should make those workflows more efficient or frictionless or more effective or more pleasing to either, you know, to users of the technology. I don't know how you get this right if you're a pure technologist who wakes up one day, harnesses AI and says, I see this five trillion dollar industry of healthcare, I'm going to go after it. I think that, you know, it's a differentiator for our company to say, we are clinical operators first. Our whole history has been in managing risk and partnering and working with providers, understanding workflows, understanding how technology gets deployed in those workflows. That's our starting point. It just so happens now, we're harnessing the power of AI to go do that work. And I think it gives us more credibility, especially in health systems when we tell that story. And so for me, it's an advantage. I don't know whether it makes it easier or not, it's probably depends on the day you asked me, but it certainly makes the conversation more credible. And I think it gets us in the room and gets us in the door. And at least people start with an open mind because they understand that DNA, they understand that history. And it certainly has helped in our early days of of selling and building and selling top. Absolutely. Familiarity and trust is what healthcare is built on. So I'm sure that goes a long way. Yeah. So one of y'all's offerings, the agent that you're offering is called Tom. And you all refer to it as a care team member, which is to me, interesting, right? It's not like a tool or a button or some, you know, dashboard. That's a very kind of strong term, care team member. So tell me more about Tom and kind of what it is and how you see it as being different than, you know, an extra app you have or an extra tool. But actually calling it that level of, hey, this is a member of your care team. Yeah. We thought that was really important. At least at this point in time and the technology is nascent, there are a lot of Americans who don't trust AI. Now it's interesting, right? There's a whole bunch of us who flocked to chat GPT and Gemini and others and are diving headfirst into it. But there are people. There's a big chunk of America who's not exactly sure what to make of it. And we thought that for a while, we were going to need a technology endorser. And so you get, you mentioned the word trust. And I think that's everything in healthcare. And so even today, even post-COVID, even in the politicized world we live in, most people still put a relatively high degree of trust in their own primary care physician. And so we felt like for AI to be adopted and accepted as part of a care reality, it needed to be endorsed by the clinician. And so for us, the whole thing starts with a provider introducing to his or her patients, Tom as a member, an agenteic AI member of the care team. You know, it signals that Tom works from your data as a patient. So it's not some anonymous, large language model. This is a data driven, coordinated, overseen AI tool that is going to help you the patient, be healthier and stay connected to me, your provider. And so I think that was number one, because I think early on, folks being willing to answer a call to engage is like the gating item for any value that's going to be created. You know, I used the analogy of I was on an airplane not long ago and we were flying into a fog, you know, kind of a dense fog environment. And it was at midnight and the pilot came on and he basically said there's zero visibility at the field. But fortunately, we're going to still be able to land night. We're going to use the auto land feature. We train for this. We don't get to use it very much, about 1% of our landings. But the FAA mandates in conditions like this that we have to use it because a human would not be able to land a plane in this condition. And it was unnerving. But the pilot had also come out before the flight and introduced themselves to the folks. He basically just said, I'd like to meet the, I like to know who's flying in the night. My name's Captain Smith. I've been flying for 34 years. I was former Air Force. We should have a good flight tonight. You know, sit back relax. We're going to dim the lights and we'll talk to you on our way into New Orleans in this case. And then when he came back on to tell us that we were using auto land, it was not a big deal because we trusted him. He understood the technology. He knew it was better than him. And I just think that's a cool analogy. And that's kind of how I think about using AI today. Now maybe 20 years from now, maybe we'll feel very differently. But today, I think we still want a human kind of overseeing it, telling us it's okay to use it, endorsing how it's going to be used. And that's how we do it. And so for the, for, you know, I think hopefully it becomes apparent at what we're doing, which is we take your data. We connect to you through a series of interactions that are autonomous based on your health needs. We summarize, we alert, we push information back into the physician's environment. If it's urgent, it pops up quickly. If it's just something that he could see later, it's filed in the EHR. That's part of it. So, so, so Tom's helping take care of folks. And then the other thing that Tom's doing is Tom is taking what he learns from a patient and knows about a patient and marrying that up to the world's evidence of what should happen for someone with diabetes in an A1C of 8.4 on these meds with these blood sugars, this kidney function. And, and creating a contextual nudge or clinical prompt with the evidence if you want to see it in a way that kind of upskill those of us, you know, to reduce the cognitive burden of those busy practitioners or, or maybe even junior practitioners and making a world class so that they can take on more care. So it's a, it's a bi-directional tool one that connects to patients, one that connects back to the care team. Supports both in their journey of primary care and towards greater health. And that's kind of how we've positioned it for today and and and that seems to resonate with the clinicians that we're talking to. Yeah, really, really cool. That that plane example is pretty wild. It's crazy that the planes can do that. And I mean, these things can essentially fly themselves. There's auto-filing. They can land themselves and you're like, oh, wow. And that he not told me, I would have never known it. Other than you didn't see the runway likes include the the tires we're hitting, which was kind of crazy. But, but you wouldn't think about it right. You do imagine one day like we're using AI today. And this, we go after the, you know, administrative use case is the things that are kind of low risk because it's new technology. There's probably a day where the technology is improving to be so good that either, I don't know if it will ever be mandated that we use it. But I think consumers and physicians will want to use it because they'll know that the capability of AI is beyond the human brain. And as a patient, I think we'll always want a human. But man, you're going to want them to be supported with world-class technology. I think I don't think it can be either or in the future. Yeah, I think about it as, you know, placing a central line, right? Like back in the 90s, you do blind lines all the time at attendings that would do them. But now, the standard of care is doing it with an ultrasound. And say you don't do it with an ultrasound and something goes wrong. Everyone on that committee is going to be like, what the wide-end you use standard of care. You're going to be in all types of trouble and whatnot. So one of those things, I'm sure there's going to be certain parts of AI and these tools that are essentially going to be standard of care and not using it's going to be almost reckless. So I've done a lot of podcasts. You are the first person who just made me feel extremely old because I was one of those doctors in the 90s doing central lines without ultrasound. So thank you for that. I appreciate that. That was not my intention. But respect to you all, because I could not do one without one. And I would be helpless. We created a few hematones along the way. One of the things that I kind of personally nerd out over is wearables. You know, I'll be honest, a lot of times they give you data, which I don't know what to do with that data. But I love looking at it. I love looking at what can, you know, make certain biomarkers and numbers better or worse or whatnot. And I know you all have some of the kind of wearables. And I think you may have a agreement with one of the manufacturers to kind of integrate into Tom. Tell me about that and how you see the role of all of that data that we're missing out on and can be captured in the home and other places in in one's health. Yeah, I mean, it's a really interesting space and early days for us. I mean, I'll, you know, I mentioned the name. I mean, we do have a partnership with Aura as kind of a to think, help us think through what this could look like. But, you know, there's such a rich source of information that comes from patients historically, you know, we work from the data that exists largely within hospitals and health systems. And that's fine billing codes and historical data and labs, et cetera. But, but, you know, the patient now is such an extremely rich source of information through sensors and wearables specifically. The problem is, and you know, I mean, I look at my Aura app every morning. I sleep. I'm curious like what happened last night. So I get in my app and I look at it. But, you know, other than maybe some hints that are baked within the app, I mean, no one else sees this information. I mean, it's like, it's a consumer experience for me. But we have been experimenting with this notion of what if we tried to make it an extension of the office, right? So, we run a Medicare Advantage plan. And last year, we created a benefit for Medicare Advantage seniors to give them an Aura ring as a benefit as part of their kind of their membership and the plan. And then many of them were willing to give us an access to their data. And so we're interested in now looking at signals, for example, of potential sleep apnea. And when we see it, you know, have Tom reach out to these members to do structured questioning. There's some standardized questions that we as clinicians might use to discern whether someone may have sleep apnea. Well, Tom can do that, right? And then if positive, you know, connect that back to the primary care physician. So, this is a really cool example of how you could see this working in primary care where if I have a garment or an Apple watch or a Fitbit or whatever device, what physicians don't want is don't give us all the data. We don't have time to wait through it, even if we understood it or we're interested. We just don't have time to deal with it. But if AI could take that data in and understand it and and marry it to the traditional data that it has access to and incorporate the totality of that information into recommendations or alerts or guidance, that would be that would be very helpful. And that's kind of how we see it. So in that world, then primary care, you know, I've said, you know, stops being randomly episodic and in person and becomes continuous and connected and proactive and frankly largely virtual. I do see a day pretty quickly where the in-person visit for primary care is the exception, not the rule. And that providers have AI, you know, engaging patients, accessing remote monitoring devices and information, monitoring them, and team up for clinicians, asynchronous kind of activities, maybe to titrate a medication or to follow, you know, to order a lab test, and that the human synchronous connections, whether those are virtual or in-person, you know, represent a subset of the patients who maybe have complexity or need that human touch. And I think that's the work of primary care in the future. And I don't think that would be possible without wearables and sensors in the home that give us information and insights that are more real-time than what we would get from labs or claims or other traditional data sources. Yeah, absolutely. I mean, I go to my primary care doctor every year and it's a pretty simple and quick visit and hopefully it stays like that for the time being. But like in my mind, I'm like, I'm not wasting his time, but in my mind, I'm saying he could probably be using this time much better for, you know, that complex patient and be thinking about, you know, various treatment options and what not versus, you know, me. But unfortunately, the system right now is built in a way where that's the only way I can get that care and have that relationship. But like you said, you know, for someone who's maybe in their 20s or 30s, you may not need to physically see that. person for three years and then maybe something comes up that he wouldn't see creeps up or cluster numbers and you see him on year four but that doesn't mean you're not getting care during that time. Yeah and we just have to figure out and honestly the only thing that stands in the way of that being the model is payment because I think a provider would say is today unfortunately mostly I get paid when I see you in Bill and Ian M. Cote and that means largely I'm seeing you you know either virtually or I'm seeing you in my office and so all of this AI facilitated monitoring that could be really value. How do I get paid for that right? I still have liability. I'm still going to take responsibility for you but I don't get paid and so I do think we need to continue to evolve payment and think about that and I you know the problem that the the reality of it is is the cost of delivering AI facilitated care is very low relative to the in-person care right and so I can see a world where there's like a base layer of access to primary care that is you know paid in a monthly installment that looks kind of like a low cost you know almost like your Netflix bill or something it's you know you know ten to twenty dollars a month and that just is for routine AI monitoring and connectivity and oversight and then you pay for the encounters that maybe once every four years but then you are part of a primary care environment you're just not consuming the key real estate and time which is the human time of the you know the position that really should be reserved for for the most important kind of critical times in someone's care and I think that's we just have to work to that model to go and I think we can get there I think we'll get there fairly quickly actually. Yeah that's awesome you kind of let right into my next question you read my mind so you are a recently named to the CMS healthcare advisory committee which is really really cool and obviously these are some of the things that you all probably will have conversations about and have input on is that you know I'm not very familiar with kind of the deep in the weeds regulatory environment are these conversations that are starting to happen in terms of payment for AI based care and how that would look like. Yeah for sure I mean exactly right well thanks and thanks for mentioning it's really kind of certainly an honor and I mean extremely humbled to be able to do that I mean and really hats off to the administration to look into the community to look at people who've been living truly in the front ends of these problems they can bring our insights and learnings back to DC it's an amazing opportunity and I'll take it very seriously but yeah the CMS is clearly the largest payer in the United States when you think about Medicaid Medicare or Medicare Advantage and so what they do matters and they're very much thinking about how do we change payment to free up innovation to make it take it full advantage of this point in time and in my conversations with Dr. Oz specifically he was extremely interested in our view of how we're approaching primary care and I think he one of the reasons I think I was selected was largely was to bring those ideas into this environment to kind of see what they can do to kind of amplify that you know for that to come to fruition everywhere though you know you need private payers employers and others to kind of get on board but CMS is such an important influence or policy and payment in this country that I think as goes CMS others are sure to follow and so real excited to get to DC and to share my ideas absolutely yeah really really unique opportunity one other thing I think I was kind of looking through some of the previous kind of interviews you've given in conversations was a discussion and correct me if I'm wrong something about maybe one day AI having its own API number is it tell me about that do you think they'll be maybe like certain tools or algorithms that will have that and how does the liability of that maybe look and would love to kind of hear your thoughts on that I mean now we're talking future future future and I think I was a pasta you know I was kind of thinking about like you know you know certainly you could treat AI as a provider and you know you could pay you could pay for AI directly by almost no different than you would pay me as a position you could you could make AI or certain forms of AI eligible to become a provider that has an NPI number and gets you know regulated payments just like all of us I don't love the I mean honestly the problem with that to me is it still perpetuates this notion of fee for service now you're just transferring that to technology and I don't think that's the right thing I think you want some sort of fixed payment so the incentives are less about volume of care and more about how do we just do the right thing for people to keep them healthy and I think that's still the ultimate payment model but but you know I do envision a day where you know we will have enough clinical trial data we'll have enough experience we'll have enough certainty for certain parts of care and you're already starting to see that in radiology and elsewhere where you know it's very clear that the technology can perform at an extremely high level and in some cases maybe won't need human oversight I think primary care is such a complex thing is it's just really hard for me to imagine a world where all of primary care is being delivered by a computer that has an MPI number but I think there are pieces of care right maybe specific chronic diseases etc where you could at least start to have a conversation and thought about that one day we're not there today and I think what you picked up on was that might be a path that we could go if we need to figure out how to get payment directly to those who are deploying AI that makes sense yeah quick logistical question so currently Tom how what is the operational distribution of it how many clinics is it in and how far reaching is it yeah it's it's early we've just deployed it within the first health system it's being deployed currently at scale we've done a lot of beta testing in small contained environments and pieces of time to test it but the first full deployment is actually underway right now at health system number one we have several others who are in very late stage contracting and we've got some exact you know I think some very exciting announcements that will be coming in in the next one to two months but this is really early because we were building a platform and going after a lot of values so this was a little slower to market intentionally we could have gone to market quickly with narrow solutions we chose to take our time and build out something that's way more comprehensive that can create more value because we're going after such a big problem and so we're in early stages I'd love to come back give give us 12 months and I'll come back and hopefully have some some real user stories and let you know how it's actually working awesome well end on one final question and you don't have to answer it with an exact number but kind of that vision that you were laying out and I think you wrote a white paper and maybe the year 2030 was on there but when when do you think as a system if everything goes right and we get you know the government is on board and clinicians are on board and everyone when do you think we may start to see a first glimpse of that kind of care where it's that continuous care everyone's got you know better access to PCPs and whatnot I imagine 2030 maybe a little ambitious but you tell me what when do you think we'll kind of get to that I don't think that's ambitious I think there'll be on the pockets I think we will see in some contained geographies a pretty substantial version of that come come to life I just think we have to have it and frankly the deployment the development of the technology and the deployment of the technology is actually going quicker than than I would have imagined and the excitement that I'm seeing from providers who are looking for solutions like this is significant and and so I feel like all the the stars are aligning and five years feels forever in today's world so I will stick to in 2030 the premise is that paper is could I envision a primary care position adequately taking care of 5,000 people whereas you know today it's probably on average it's kind of strong down to about 1800 patients per primary care panel I think that is possible by 2030 I really do I love it I love it that that gives me a lot of excitement and optimism and you know for the longest time thing things in healthcare move very slow so it's exciting to actually see these ideas kind of hitting hitting the ground and you're you're a big part of it so thanks so much for joining me and I'll take you off let's do this again in a year and well I would love to get an update then I'd love it thanks for everything you're doing appreciate time absolutely take care thanks for joining us on this episode of non-fungible human we are always open to suggestions on who we should have on next and feedback is always welcome if you're enjoying the show please feel free to rate subscribe and leave a review wherever you'll listen to your podcasts this helps others find the show and we greatly appreciate it until next time stay healthy friends and we'll catch you in the next episode