Lifers with Christina Farr

Paul Brient on Athenahealth's AI-driven EHR strategy, open ecosystem, and automation of prior authorization

57 min Listen to the original
Healthcare ITElectronic Health RecordsArtificial IntelligenceRevenue Cycle ManagementPrior AuthorizationDigital HealthHealthcare Administration

Summary

Paul Brient, CEO of Athenahealth, discusses the company's evolution from a disruptor to an industry leader serving 180,000 providers through an open ecosystem strategy. He highlights the inevitability of AI in reducing administrative burdens, citing a partnership with Humana where 70% of prior authorization transactions were completed without staff intervention. Brient advocates for 'gold carding' to reward compliant providers and suggests using open-weight AI models for stability in healthcare compliance. He also notes a shift in engineering skills toward managing AI agents and emphasizes direct customer feedback loops to drive product development.

Episode sections

  1. 0:00 – 2:59

    AI in Healthcare and the Future of EHRs

    Paul Brient discusses the inevitability of AI in healthcare, noting that patients are already using tools like Claude for medical information. He envisions a future where AI handles administrative burdens, allowing doctors to spend 95-99% of their time on patient care rather than documentation.

  2. 2:59 – 4:08

    Athenahealth's Origin and Evolution

    Brient recounts that Athenahealth was founded by Jonathan Bush and Todd Park, who initially attempted to create a chain of OB-GYN clinics in California. They pivoted to revenue cycle management (RCM) after finding it difficult to get paid, eventually evolving into the EHR company it is today.

  3. 4:08 – 6:30

    From Disruptor to Industry Leader

    Brient argues that Athenahealth has transitioned from a scrappy disruptor to an industry leader serving 180,000 providers. He emphasizes an open ecosystem strategy, inviting third-party innovators and other EHR companies to connect, rather than maintaining a closed walled garden.

  4. 6:30 – 8:00

    Target Audience and Ecosystem Strategy

    While maintaining an ethos of supporting independent practices, Athenahealth now serves a wide range of entities, including large telehealth companies and specialized care providers like menopause care startups. These customers often choose Athenahealth for its open ecosystem that allows for custom integration.

  5. 8:00 – 28:49

    Administrative Burden and AI Solutions

    Brient acknowledges that healthcare administration often prioritizes billing over clinical outcomes due to financial necessity. He believes AI can eliminate much of the administrative burden, allowing providers to focus on clinical care.

  6. 28:49 – 30:18

    Humana Prior Authorization Expansion

    Athenahealth announced a nationwide expansion of its electronic prior authorization agreement with Humana, covering 180,000 providers. Data from September showed that 70% of 35,000 transactions were completed without staff intervention, highlighting a reduction in administrative burden.

  7. 30:18 – 35:08

    Gold Carding and Provider Quality Improvement

    Paul Brient described a 'non-zero-sum' approach to prior authorization, using AI to filter out 70-95% of routine cases so staff can focus on outliers. This mirrors the concept of 'gold carding,' where compliant providers face less friction, potentially enabling continuous quality improvement and cost reduction.

  8. 35:08 – 39:00

    Patient Access to Medical Records and AI Interpretation

    The speakers discussed the impact of 'open notes' and patient access to medical data. Brient suggested delaying access for serious diagnoses like cancer but emphasized that AI tools help patients interpret medical jargon, leading to better-educated patients and more effective consultations.

  9. 39:00 – 44:20

    Brient's Early Career and Practice Management Origins

    Brient recounted writing billing software for his father's surgical practice at age 15-16, becoming the first person to send an electronic claim in Florida. He sold his company Patient Keeper to HCA in 2014 and credits mentors who talked him out of medical school for his career in healthcare IT.

  10. 44:20 – 48:34

    AI Strategy: Open-Weight Models for Stability

    Brient explained that while Athena uses frontier models from OpenAI and Anthropic, they are exploring open-weight models for stability. Proprietary models change frequently, requiring constant revalidation, whereas open-weight models offer control over updates, which is critical for compliance and reliability in healthcare software.

  11. 48:34 – 52:01

    Engineering Culture and Thrive Conference Feedback

    Athena is sending engineers into exam rooms to observe workflows directly, contrasting with competitors whose engineers do not use their own software. At the Thrive conference, customers provided 1,000 ideas for software improvements, 200 of which were already built, demonstrating a rapid feedback loop.

  12. 52:01 – 56:26

    Shift in Engineering Skills

    Brient noted that traditional coding skills are becoming less valuable than the ability to architect systems and manage AI agents. He described the new role as similar to managing human employees, requiring astute judgment to guide AI outputs and correct errors, marking a shift toward generalist problem-solving.

  13. 56:26 – 56:52

    Podcast Outro and Contact Info

    The host thanks Paul Brient for the interview and asks for his preferred contact method. Brient states he is primarily a LinkedIn user and not heavily active on other social media.

  14. 56:52 – 57:22

    Show Notes and Sponsor Read

    The host wraps up the episode, directing listeners to the 'Second Opinion' newsletter for more digital health content. The segment concludes with a sponsor read for Hippocratic AI.

Assets discussed

AssetKindStanceContext
Anthropicprivate companyneutralOkay, so you just mentioned anthropic and Claude, and that's been the big news over the past few weeks, is that you know, Athena is is integrating and so tell tell us a little bit more about that.
OpenAIprivate companyneutralInteresting. I assumed just because I was looking at the um the alternative with um OpenAI OpenAI um and in Epic that you it had to be provider mediated.
NVIDIANVDAequityneutralUm okay. So NVIDIA is in the news at this moment in time, you know, not really for great reasons, but um now prior CEO, Jensen Wang, was one of the, I would say, most vocal people saying that,

“Stance” is how the speakers talked about the asset, not a recommendation.

Claims worth verifying

Athenahealth has an agreement with Humana to expand electronic prior authorization nationwide to 180,000 providers.

Unidentified speaker (speaker_1) ·29:09

How to verify: Check athenahealth investor relations press releases or Humana investor relations announcements for partnership details and provider counts. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

In September, 6,000 providers ran roughly 35,000 EPA transactions with Humana, and 70% plus were completed without additional work from practice staff.

Unidentified speaker (speaker_1) ·29:22

How to verify: Review athenahealth case studies or press releases from September 2026 (or relevant period) citing Fierce Healthcare or similar outlets for specific transaction volumes and automation rates. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Paul Brient was the first person to send an electronic claim in the state of Florida.

Unidentified speaker (UNKNOWN) ·40:47

How to verify: Search Florida state medical board archives, historical healthcare IT news, or athenahealth corporate history pages for biographical details on Paul Brient. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Paul Brient sold his company Patient Keeper to HCA in 2014.

Unidentified speaker (speaker_1) ·39:17

How to verify: Check SEC EDGAR filings for HCA Healthcare (HCA) regarding acquisitions in 2026, or review Patient Keeper's corporate history/press releases. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Athenahealth has an engineering team of approximately 3,000 people.

Unidentified speaker (speaker_0) ·50:39

How to verify: Check athenahealth's LinkedIn company page, recent job postings, or corporate fact sheets for employee headcount. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

At the Thrive conference, customers pitched 1,000 ideas, and 200 were already built.

Unidentified speaker (UNKNOWN) ·50:58

How to verify: Review athenahealth Thrive conference recap materials, blog posts, or press releases detailing customer feedback and product updates. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Athenahealth works with both OpenAI and Anthropic.

Unidentified speaker (speaker_0) ·45:39

How to verify: Check athenahealth technology partner pages, press releases, or technical whitepapers for mentions of OpenAI and Anthropic integrations. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Paul Brient wrote billing software for his father's surgical practice at age 15 or 16.

Unidentified speaker (speaker_0) ·40:35

How to verify: Review Paul Brient's biography on athenahealth's leadership page or interviews in healthcare IT publications. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

OpenAI and Anthropic are hiring large teams in healthcare.

Unidentified speaker (speaker_1) ·45:00

How to verify: Check OpenAI and Anthropic career pages for job listings in healthcare, or review industry news reports on their hiring trends in the sector. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

OpenAI is working with Epic.

Unidentified speaker (speaker_1) ·45:00

How to verify: Check Epic Systems press releases or partnership announcements for collaborations with OpenAI. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Paul Brient is mostly a LinkedIn user and not hugely on other social media.

Unidentified speaker (speaker_0) ·56:44

How to verify: Check Paul Brient's LinkedIn profile for activity frequency and absence of links to other social platforms. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Hippocratic AI builds safety-focused AI agents that reach patients health systems don't have the staff to reach.

Unidentified speaker (speaker_2) ·57:05

How to verify: Review Hippocratic AI's website (HippocraticAI.com) or press releases for product descriptions and target market claims. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Hippocratic AI CEO is Munjar Shah.

Unidentified speaker (speaker_2) ·57:05

How to verify: Check Hippocratic AI's leadership page or SEC filings (if public) or press releases for executive titles. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Do your own research

  • Verify Athenahealth's current provider count and market share against recent SEC filings or company press releases.
  • Review the specific terms and scope of the Humana prior authorization agreement to assess the scale of automation claims.
  • Investigate the definition and implementation of 'gold carding' in prior authorization workflows across major payers.
  • Check the regulatory landscape regarding patient access to electronic health records (e.g., 21st Century Cures Act) and any exemptions for sensitive diagnoses.
  • Analyze the technical differences between open-weight and proprietary AI models in the context of healthcare compliance and data privacy.
  • Assess Athenahealth's competitive position against other EHR providers like Epic and Cerner regarding open ecosystem strategies.
  • Review the financial impact of administrative automation on revenue cycle management efficiency and cost reduction.
Full transcriptexpand

0:00Unidentified speaker (speaker_0): The reality is people are getting healthcare from Claude today, right? Yeah.

0:03Unidentified speaker (UNKNOWN): We can't

0:03Unidentified speaker (speaker_0): stop that. That genie is out of the bottle and you know,

0:06Unidentified speaker (UNKNOWN): like

0:07Unidentified speaker (speaker_0): AI and other topics, sometimes it gives you very good advice and sometimes it gives you copy advice. And um, we're not gonna endorse that. That w that is definitely a line for us. You could imagine a world, and I'm only imagining, and this is well out of my purview, whereas a patient you could go query the information exchanges yourself about you and get your get all of your stuff. I kind of hope. that the doctors are just taking care of the patients. Like you can imagine a world where, you know, the doctor is briefed via a computer in a thoughtful way, and then most of their time, like 95 to 99% of their time, is engaging with patients and providing care.

0:44Unidentified speaker (speaker_1): Any practical, tactical advice just on people who are wondering, you know, listening in and maybe thinking over. oh gosh, my skills sound a bit redundant. What should you actually do to stay relevant in this new economy?

0:57Unidentified speaker (speaker_2): Hi everyone, and thank you so much for your support of Lifers. It genuinely means the world to me. Before we begin, I want to quickly thank our partner. Lifers is made possible with support from Hippocratic AI. And honestly, I'm thrilled to have them. They're building safety-focused AI that reaches patients a system has never had the capacity to reach. My conversation with their CEO, Minjal Shah, is one of my favorites on the show. Head to the link in the show notes to check it out. And for anybody watching on YouTube, please make sure you're subscribed. Your support helps us keep bringing you incredible conversations about the future of health. So go ahead, hit subscribe, and share this episode with somebody who'd enjoy it. Now, on to today's episode. Hi,

1:37Unidentified speaker (speaker_1): everybody, welcome back to Life Is. I have Paul Bryant with me today, who is the Chief Product and Operations Officer at Athena Health, a company that, if you are tuning into this show, I know you know well because they've been around a long time. Uh, but we're gonna talk about Athena's origin story, which uh, if you don't know it, it's fascinating, and you would probably not guess what the first uh product for Athena was. Um, it's gonna be surprising. So tune in for that. Um, all the way through to questions of how healthcare companies are using AI today, um, independent practices and where it's all going. The recent summit um that they hosted called Thrive, where they gathered up thousands of their customers and had um 3,000 engineers on call to solve problems in real time, which I think. is just, you know, an amazing use case. And then lastly, some of the work that they're doing with Anthropic and other companies to surface information to patients in real time. Paul, welcome to the show.

2:41Unidentified speaker (UNKNOWN): Thank

2:41Unidentified speaker (speaker_0): you. It's great to be here.

2:43Unidentified speaker (speaker_1): Can you tell us? Because I don't think most people know this, and this is one of my favorite Athena stories. Uh, what was the first idea? that the founders of Athena Health had before it became Athena, the EHR? Um, what was it before that?

2:59Unidentified speaker (speaker_0): It is a great founding story. Um, so the founders of Athena actually set out to found a chain of OBGYN practices.

3:06Unidentified speaker (UNKNOWN): And

3:06Unidentified speaker (speaker_1): we're talking

3:06Unidentified speaker (UNKNOWN): about Jonathan

3:07Unidentified speaker (speaker_1): Bush and Todd Park, so two guys.

3:08Unidentified speaker (UNKNOWN): Jonathan

3:09Unidentified speaker (speaker_3): Bush and Todd

3:09Unidentified speaker (UNKNOWN): Park,

3:10Unidentified speaker (speaker_3): they

3:10Unidentified speaker (speaker_0): they set out to go to California to revolutionize the way babies are are are grown. delivered. And what they found, which is what everyone finds in the US healthcare system, is that it was extremely hard to get paid for their services. They delivered great services to mothers and babies, but could not get paid. And they set out as good entrepreneurs to figure out how to get paid and kind of figure that out. And then someone had the idea of wait a minute, that's actually more valuable than a whole bunch of uh OBGON clinics, and they moved back. to Boston and founded what is now known as Athena Health, the at the time the Revenue Cycle Management Company and now the EHR company.

3:46Unidentified speaker (speaker_1): Yeah, and I mean I think of it sort of as both because you know having the EHR is really how you get into RCM and and billing.

3:54Unidentified speaker (speaker_0): Absolutely. Um which is uh how I now have Now the the the I added a new thing to my title, which is operations officer. So I'm responsible for both the product, the EHR piece. as well as all the RCM operations. And we finally put those together after many years. And that's worked out really well.

4:08Unidentified speaker (UNKNOWN): So

4:08Unidentified speaker (speaker_1): I think of Athena, you know, as being a company that really wore the disruption, disruptor kind of title for many, many years. There was even a program called More Disruption Please from the Jonathan Bush era. And I remember going to some of those events all around the country. And, you know, I think it made sense because in the EHR space, you've got some mega giant incumbent companies that have been around for many, many decades, like Epic, like Cerna. And so, you know, it makes sense to think of Athene Athena as being quite a scrappy company in relation to those giants. And then the same is true on the RCM side, lots of big, big, big players that have been around a long time. Um so again, like you could think of Athena as being scrappy. But at this point in time, like do you still think that, you know, Athena do you think of Athena as a disruptor, or do you think of it more as um, you know, a connector or a more of an incumbent just because of how long Athena has now has now been around and and the scale that it's gotten to as a business?

5:12Unidentified speaker (speaker_0): Yeah, it it's a really interesting question. And and you know we we we now serve 180,000 providers, so we're not small and scrappy. We like to try to still be forward-looking. I I think of us more as maybe a leader than a than a disruptor. And you know, a disruptor, you're kind of this small thing over in the corner trying to do something. Um we're now trying to kind of I think show the industry the way. And we're, you know, many of those same tenants that the founders had around more disruption, please still apply, which is we are rapidly open.

5:40Unidentified speaker (UNKNOWN): Like

5:40Unidentified speaker (speaker_0): we're like, hey, the way to solve the healthcare problems is to connect with everyone and invite everybody in. Like we do not have a monopoly on innovation. We think we're innovative and all that good stuff, but we invite everyone into the ecosystem, our customers, third-party uh innovators, other EHR companies, company is like come play, come connect, and come solve the problems of the US healthcare system. And so I I hope hope we've kind of transitioned from that disruptor to um you know a more sustainable, maybe more impactful kind of leader in the industry.

6:10Unidentified speaker (speaker_1): And why why don't you speak to an ideal Athena like customer as well? Because I think it has changed um over the years. I still think of you mostly as working with independent practices, but I'm I'm sure that that has also evolved. So walk us through who who are right now is like Athena's core target. And then how has that shifted?

6:31Unidentified speaker (speaker_0): Yeah, you know, you know, uh we still think of ourselves our ethos around that independent practice. Like, you know, how do we keep an independent practice that once remained independent, independent? And in the US, it's really hard. Um, we make it extremely difficult to practice medicine here because of the the non parts of practicing medicine, all the administrative stuff they're to getting get paid. So that's still very much part of our ethos. And we've expanded, though, to serve everything from the single doctor practice that's trying to stay independent and do the OG kind of private practice, all the way up to some of the largest and frankly most disruptive companies in the industry. So we're seeing a whole advent of people that are trying to disrupt the delivery system, maybe much like Jonathan and Todd were thinking they were going to gonna do early in the days of Athena. world. And they often pick Athena because of our open ecosystem. Like if you want to do something really different, you can't go take a nice closed system that kind of works only one way. You take Athena and then you build the pieces on it that you need to be different. So we have a ton of uh nationwide telehealth companies, people with very different models.

7:32Um, I don't know how many of these I'm allowed to say on on on air, but we have the uh uh a a group out in California that's doing uh um a menopause care, uh kind of midlife uh care for women. That's a really underserved market. They're growing like crazy, they're in Silicon Valley, they feel like a startup, but they're in healthcare. Um and so it's really neat to see all of these um companies come and innovate with us uh in the healthcare space.

7:55Unidentified speaker (speaker_1): Yeah, yeah, I love that. Um so I'm gonna tell you something depressing, Paul, just cause just

8:00Unidentified speaker (UNKNOWN): because

8:00Unidentified speaker (speaker_1): we're I'm gonna go there, which is that um I recently had a had a survey, a little informal survey, and then I turned it into a tweet that led to a lot of replies that uh we can drop a link to the to this tweet in the in the comments for those tuning in. But it basically asked physicians a question of do you feel like your leadership um cares more about whether you're good or bad at your job, whether or not you're um taking care of patients in the way that you should and achieve achieving, you know, the the best possible outcomes, or does your uh administration care more that you are correctly billing and and documenting? And and Paul, I I see you nodding, so I think you can guess what the overwhelming, overwhelming response was to that question. Um and why don't you shoot Should shoot me a guess?

8:47Unidentified speaker (speaker_0): I I would imagine they're focused more on the billing and coding.

8:50Unidentified speaker (speaker_1): Yes. Yes.

8:51Unidentified speaker (speaker_0): Um and uh and look, I I I it is depressing and it is, I think, the way the world is and has been. Um and in some ways of necessity. I mean, you go back to the origin story uh of Athena, right? What did they find was the biggest challenge in their practice? It wasn't the nuances of taking care of newborns or uh making sure that high high-risk risk uh uh pregnancies are are managed appropriately. It was getting paid. And so it it is a little bit of the that is the necessity and the very unfortunate necessity. And I think we're on the verge of changing that, I hope, um, with the advances in AI and things that we can do that we just couldn't do before. And and to get all of that administrivia out. So you also talk to doctors and say, you know, do you want to keep practicing medicine? And you often get the answer of, no, I want to get out of this game. because of all the not, you know, I came here to practice medicine and I end up not practicing medicine. I end up doing all this administr administrative work that is not part of what I signed up for. And so I think if we can get rid of that and get rid of back to everyone focusing on practicing medicine, boy, wouldn't that be a good place.

9:56Unidentified speaker (speaker_1): Okay, so let's let's speak to that a little bit. Because Azina was was very much historically part of this, that you know, you had this generation. of EHRs that came around and then we had all these government rules like like meaningful use and billions of dollars of taxpayers' money being put towards digitizing um all these healthcare systems with the idea of you know at some point we can capitalize on that by making it really easy to share information and we'd have this kind of glorious future. And as you know, Paul, this this did not transpire in the in the way that we hoped. And we create a lot of closed systems. We still don't have interoperability truly, um, even after decades of of trying with lots of new standards and various um approaches um from both the private and and public sector. And we're in a a position still where physicians are doing a lot of documentation and note taking. Um, now here we are in the AI era, and um, there is hope, you know, after many years of I would say not that much hope, um, that some of this is going to start to improve. And we've seen some technologies gain very widespread adoption, like the ambient scribes, that can take on at least one aspect. of this for

11:16the physician. So, Paul, why don't you could you speak a little bit to that from just the perspective of what you're seeing at Athena?

11:24Unidentified speaker (speaker_0): Yeah, I I I totally agree. And uh honestly, I've never been as excited uh as I am today around the potential for the EHR. Um, I've been in the healthcare IT business my entire career, and for the longest time, I'm like, hey, I'm never ever gonna work on an amuletary ambulatory EHR because. it doesn't really solve any problems for doctors. Um and meaningful use was was came out, and everyone had to use it, and it didn't solve any problems for doctors. In fact, it created more problems than it solved for doctors. And now we're at a point where we can actually use it to solve the problems of the doctors. Um and you mentioned ambient scribe, um, you know, that's certainly part of it, right? Let us write your note node for you, but it goes way beyond that. With AI, we can actually create interoperability that works. Because one of the big challenges with interoperability is we actually do move the data around mostly, but we move it in all sorts of different ways with different formats. And redundant data comes in and data that's coded weirdly and differently. And humans can get that, and then they have to go read hundreds of pages of stuff to have interoperability, and no one's gonna do that, right? But AI can go eat that for lunch and go summarize it and present it and digest it and deduplify it and and and and send it to the doctors in a way that they can deal with. Um and it can do all your coding, it can I mean it it really can do the work that previously we needed essentially something

12:43someone to read. That was the problem. Right? Computers didn't do a good job of reading comprehension, right? You send a send a fax, you know, we had some. old school AI that would figure out that the fact you know what the fax was and get it to the right patient. And that was kind of it. But now I've got AI that reads the whole fax and can do all the things that you know a person who read the fax could go do. And it just opens up a whole new world. I'm a little worried that we're gonna like cement faxes in as our new interoperabilities. Interoperability is too

13:09Unidentified speaker (UNKNOWN): much.

13:10Unidentified speaker (speaker_1): I was just gonna say we had a column come out in second opinion from a founder who was like, I actually love faxes. Sorry to say that as an AI founder, I'm I'm because I can read the faxes. You

13:22Unidentified speaker (UNKNOWN): can read them, and it's like, okay,

13:23Unidentified speaker (speaker_0): I mean, like, yeah, why not?

13:27Unidentified speaker (speaker_1): Maybe maybe like there's a world in which healthcare was so slow that there's even some kind of bizarre advantage of this AI from moment it, because you can read faxes. Um, like you said, and um, you know, we never really went through like a true cloud moment or a true, you know, mobile social moment. We just

13:46Unidentified speaker (UNKNOWN): skipped that.

13:49Unidentified speaker (speaker_0): It's like some of the the some of the uh like third world countries that skip land lines on their phones and all just all went to to to mobile phones, right? They actually have better infrastructures than we do in the US. And it's like, huh being late is an advantage sometimes. May Maybe maybe that'll be the case.

14:03Unidentified speaker (speaker_1): It was it we meant to do it. It was strategic all along. Um, amazing. Well, I really hope that is true. So, um, okay, so EHR revenue cycle. I want to talk about both of those areas. Um, both in both areas, you're seeing tons of interest in AI and what AI will mean. Um, in you know. the EHR space, people are talking now about systems of record versus systems of action. I'd love to hear what you think of that and what the fate of the EHR will be in an AI era, um, especially an EHR that's not open and not playing ball with everybody else in the industry. And then from there, I'd love to tackle a little bit more of the the revenue spice spike cycle space because you you look across both. Um give given, you know, people have said revenue cycle, what an obvious smash hit application for AI. That was certainly like a a true, you know, in this first exuberant wave. And now folks are saying, ee, maybe not quite as simple as we once thought. And I think we'll get into some of the complexities there.

15:08But let's start with my favorite topic, which is the EA EHR. job.

15:13Unidentified speaker (speaker_0): So, um I yeah, I think this the this the notion the EHR is a system of record, right? We we maintain the medical legal record for our clients. And that is a very important role. And there's a bunch of regulations around it, a bunch of case law and all that stuff, right? And that, I mean, I don't think that's going to go away. I think the question is, okay, the HR also has this part where the humans interact with

15:34Unidentified speaker (UNKNOWN): it, um,

15:34Unidentified speaker (speaker_0): and the agents that interact with it uh increasingly. And and what does that become in and and and and and do they do they get separated? And I and I do think that if you are a closed system, um you run the risk of being separated

15:48Unidentified speaker (UNKNOWN): um

15:49Unidentified speaker (speaker_0): and having, you know, people not really use your system anymore. Um

15:54Unidentified speaker (UNKNOWN): and

15:55Unidentified speaker (speaker_0): I think that, you know, you can argue that's a good or a bad thing or not. Um I think the right way to do it, at least the way we're doing it, so I I won't judge, but it's we're putting all of our money is, you know, the goal is to open the system up and also bring AI in into the system so it's native. And so we have this notion of an AI-native EHR where we're basically rebuilding it as if AI already existed when we started. And I think if you do that, the need for someone to go, you know, go contract with 20 companies from Celical. Silicon Valley to go make their clothes system kind of sort of work goes away. And they go, Oh wow, I can I can have agents right in here. They're right here for me. I I I don't care whose agent it is. Um I just switched from Muse to Dot this morning. Um I love my Muse agent and uh we got dot connected to our corporate structure, and now dot is my agent. Like I don't care wh whose agent it is. Uh I just care that it goes off and books my travel and you know sets of appointments and stuff. Um and so I think that if we bring the the HR vendors bring the gentic capabilities in the right way to to their to their customers, I think that the the this notion of system of action versus system of record, it won't be a distinction. But I do think for the legacy systems and the people that are trying to keep walled gardens up, they're gonna

17:10face this tension where people are gonna be like, no, I'm just gonna go get a bunch of agents and kind of build another user. experience and just have the AI populate your system of record.

17:19Unidentified speaker (speaker_1): Yeah, I think that that makes so much sense. And I will say as a as you know, I have my own um AI that is like my executive assistant, one of my great frustrations is that while it will book travel and if it's amazing at that and it's um sending out, you know, calendar invites and whatnot, um, it's getting blocked. whenever I ask it to do vaccinations for my children. I have three of them. There's a lot of vaccinations, I gotta

17:46Unidentified speaker (UNKNOWN): say.

17:47Unidentified speaker (speaker_1): Um, you know, and like any basically anything related to healthcare. And I I get it, you know, like you don't want a bunch of agents taking up all the appointments in the day and whatnot. Um, but like gyms and studios and you know, the whole wellness industry is around. has already embraced it and is seeing a big spike in people like going to classes and whatnot, because the uh the EAs, the AI EAs are booking all this stuff on

18:11Unidentified speaker (UNKNOWN): the case.

18:12Unidentified speaker (speaker_1): Yeah,

18:12Unidentified speaker (speaker_0): they make it work out. Like, yeah, you uh you told me you want to work out. You gotta go to the gym today.

18:16Unidentified speaker (UNKNOWN): Yeah,

18:17Unidentified speaker (speaker_1): like why is it bad that that we see increasing utilization of services? So what you know, what what's your thought on this? Like, is this a temporary blip? where healthcare is saying like we're just gonna block all these like AI agentic companies from like letting consumers manage medical bills or letting consumers manage prevention or any any other healthcare related service?

18:41Unidentified speaker (UNKNOWN): Well

18:41Unidentified speaker (speaker_0): look, I mean I I think across the industry people are trying to figure this out. I mean United Had just announced yesterday that they're blocking uh the all the agents from booking travel which seems highly unfortunate. Hopefully they will reverse that. Um look, I I don't think it is bad. I think the question is what's right and what's you know, what's appropriate and what's not appropriate.

19:02Unidentified speaker (UNKNOWN): Um

19:03Unidentified speaker (speaker_0): and I it's such a new technology. I I don't know that we've all worked that out. We're taking a pretty aggressive crawl approach to this.

19:12Unidentified speaker (UNKNOWN): Um

19:12Unidentified speaker (speaker_0): we just put a MCP server in front of basically the all the APIs that we offer to PHRs. So a personal health record can access a bunch of stuff on a read-only basis right now by your health. And if you're using Claude, you can now do that within Claude. And that's proving to be controversial, frankly, among our customers. I don't think it's inappropriate, otherwise, we wouldn't have done it. Um And we are contemplating enabling. that technology to do things like book appointments, uh, understand your bill. Um, why not? Right? Now, you know, people go down the okay, now I'm gonna get my healthcare from Claude. And the reality is people are getting healthcare from Claude today, right? Yeah. We can't stop that. That genie is out of the bottle and you know, like AI and other topics, sometimes it gives you very good advice and sometimes it gives you crappy advice. And um we're not gonna going endorse to that. That is definitely in line for us. And we have our own agent. If you use our application, so we have a you know a portal application. We we have an AI agent in there that will do a lot of these things, book appointments for you, and all that. And that's guardrailed, so it won't give you medical advice. So if you say, hey, you know, I want to change the dose of my medication, what should I do? It's like, call your doctor. That's what you should do, right? Um, but in the wild, people can ask that question and get all kinds of answers. And

20:28I don't yet know what to do with that. I generally think it's an awesome thing. I I I frankly use it personally. Um I've uploaded um you know, I got I I issue with my left knee. I got a a C a MRI, I uploaded the MRI to uh to Chat GPT and said, Hey, I want to modify my exercise routine, giving Given these results, and it did. And you know, I'm gonna go see a physical therapist too, but I can't do that for another week and uh I'm in the meantime taking chat's advice, and so far it's doing good. So um it it's it's it's a it's a world that we just have never seen before, and we don't know what the playbook is. Um, but I think it's up to all of us to ask, you know, what if, as opposed to just saying no.

21:06Unidentified speaker (speaker_1): Okay, so you just mentioned anthropic and Claude, and that's been the big news over the past few weeks, is that you know, Athena is is integrating and so tell tell us a little bit more about that. How did that come to be? What does that mean? Um I'm assuming you know it's really focused on the providers, but then is there a playbook for for patients as well? What do you think is going to be the future there for the patient?

21:29Unidentified speaker (UNKNOWN): Well,

21:29Unidentified speaker (speaker_0): the the the big thing we're doing actually this week is is for the patient. So we're we're making it so that now if you're a patient and you're in CLOD, Claude, you can um with very explicit permission say, Hey Claude, go access my medical records in Athena. And Claude will have full access to everything that you would have if you logged onto the portal or you were in the Apple Health app or any of the other PHRs out there. And uh and it will protect that information. It will not use it for training or anything like that. But when you ask questions about your health, it will have that as context. And we did that in part because the regulations basically say we have to, like if you know, Claude is essentially a personal health record. So it's appropriate for them to go do that, obviously, with all the controls and all the all that. But also, if you think about the alternative, is I'm in Claude and I'm cutting and pasting my medical record into Claude and asking it questions. Like you can argue whether Claude's going to give you a good answer, but if you put part of your medical record in, Claud is not going to necessarily give

22:30Unidentified speaker (UNKNOWN): you

22:30Unidentified speaker (speaker_0): the right answer because it's not going to have the full context. And so this is the, in my mind, the right way to do it. It also enforces all the protections around health data. If you just paste stuff randomly in Claude, it's not protected.

22:43Unidentified speaker (UNKNOWN): That's

22:43Unidentified speaker (speaker_1): interesting. I assumed just because I was looking at the um the alternative with um OpenAI OpenAI um and in Epic that you it had to be provider mediated. But the fact that a patient can do it on their own is is very cool. I guess like I would run into the same problem as a patient, which you know isn't your fault, but like curious how you would think about this. That like I might be getting some portion of my care from a provider that uses Athena and then another portion of my care from a health system that uses Epic. And so there's always going to be those missing gaps and aspects of things being incomplete for most people, unless you're like an Athena only shop. And it's very hard to control for that because most physicians don't even tell you what EHR they use. So, you know, do you think you're going to get like this pain from Athena that feels the most complete through a Claude? And then from everything else, you're sort of still gonna have to be copying and pasting for the near future?

23:43Unidentified speaker (speaker_0): Well, this gets back to our interoperability discussion. So right now with Athena, we actually have all that information from the other systems about you in our in our in our data stores.

23:55Unidentified speaker (UNKNOWN): Um

23:56Unidentified speaker (speaker_0): we have some discussion about whether we would also make that available uh in the portal and therefore available. in uh in Claude. We haven't made that decision yet, and there's all kinds of both legal and just conceptual discussions about that. Um, but it's increasingly technologically possible. And you know, you could imagine a world, and I'm only imagining, and this is well out of my purview, whereas a patient you could go query the information exchanges yourself about you and get your get all of your stuff.

24:26Unidentified speaker (UNKNOWN): Um

24:27Unidentified speaker (speaker_0): because we have the plumbing there. You you know, all the the I don't know how many billions of dollars we spent on all the plumbing. We have the plumbing. We just don't use it right. Um, but you can tap into that in theory. Now that's that's a little bit far out there, but I I wouldn't

24:42Unidentified speaker (UNKNOWN): I

24:42Unidentified speaker (speaker_0): wouldn't say that that isn't going to be part of the world because as a patient, you would like that, and it is your information. Everyone talks about, you know, is it my information if I'm the provider or not? It is the patient. patient's information. Their medical record is their information. And they should be entitled to have it, in my view.

24:55Unidentified speaker (speaker_1): Yeah. Yeah, I I completely agree. And it's crazy how like it still is. I mean, I was just going in for um I had a recently had a procedure at a health system in the billing department called me and they were like, Can you send over snippets of clinical information for like when we have to contact the insurance company? And I was like, Why why? don't you have that? And they're

25:16Unidentified speaker (UNKNOWN): like,

25:17Unidentified speaker (speaker_1): the clinical and the financial side don't necessarily speak. It's like, how, how are we still in 2026? Like, I'm even talking about two different departments within the exact same health system, let alone two different health systems. And like the the impact to all of it, Paul, is is just dizzying, given like we're so worried at this moment about spend and affordability and having just these massive information gaps. Like it's so clear to me and everybody else in the industry that like if we don't have the information, we're gonna repeat the the tests and procedures to go to go get it. Um, or we're gonna just outright make mistakes.

25:57Unidentified speaker (speaker_0): And absolutely. And we're gonna spend all this money on kind of useful. useless, you know, in the sense, you know, not providing healthcare, all this administrative tax on the system.

26:07Unidentified speaker (speaker_1): 100%. Um, okay, maybe we should flip a little bit to revenue cycle because that's another area that I think is AI is coming for revenue cycle for sure. So you see in the news um poll, like, you know, some random health system has negotiated a rate increase, like they're doing that year after year. The payers agreed to it. But what what you don't see in the news headline is in response to this rate increase, the payer has just gone and done a bunch of really complex things in the background to make it much harder for that health system to get paid. And that there is a whole industry around recouping that payment for a service that a health system has already run. rendered sometimes months later, and that millions of dollars are just never recouped. So um to me, this is like, you know, we we want to talk about waste? Like let's talk about that. Um, what's your view on it? And then how do you think um AI plays role? And I think is already starting to to play an interesting role

27:11Unidentified speaker (UNKNOWN): here.

27:12Unidentified speaker (speaker_0): Yeah, you know, it it it it it is a challenge. And I think in and this has been since you know since we kind of invented health insurance, right? In a given transaction, it's kind of the zero sum game. You you described it very nicely, right? It's like um every dollar that the payer doesn't pay to the providers is a dollar that they keep, and every dollar that the provider capture captures is a dollar that the payer doesn't get.

27:33Unidentified speaker (UNKNOWN): Um

27:34Unidentified speaker (speaker_0): I'm hopeful that somehow we'll be able to able to work our way out of the zero sum notes, because I think there is a a game that isn't a zero sum game. But you know, what we are seeing right now is that payers are under a lot of stress. Um this was a very bad year for them. They all lost money. Um I think all is a pretty fair characterization, but certainly most of them lost money. Um and they're doing things to pay less money. And you know, one of my least favorite things that they're doing is saying, hey, you don't need prior auth for this, but then you go perform the service. and they go, Oh no, but that didn't meet medical necessity, and now we're not gonna pay you.

28:06Unidentified speaker (UNKNOWN): And

28:06Unidentified speaker (speaker_0): so now we have to go fight about whether it meant me meant medical necess necessity after the doctors already provided the service. Right? Like I think everyone would rather just do the prior off. Like like like wait a minute, that that wasn't that wasn't the game that we thought we were playing. We thought we said no prior off, we could just do the service and you'd You pay us. Um and so there are you know, there's there's there's there's a lot of stuff going on that, you know if you're in the payer's seat, you're like, hey, let me tell you why we're doing this, and that makes sense, but it's for a small portion of the problem, and it creates a lot of expense for the rest of the industry. Um and I I do think AI can help. I I I I I think you know, a more fundamental like let's solve this together,

28:45Unidentified speaker (UNKNOWN): I

28:45Unidentified speaker (speaker_0): think has it's always been the opportunity. We just haven't figured out exactly how to do that yet.

28:49Unidentified speaker (speaker_1): Okay, so let's dig into some of their more recent news because Athena just recently held an event. Um, thousands of people came. It was called Thrive. It was in Nashville. Um, next year I definitely want to be there. Um, poll. But um let's talk about one of the big pieces of news that that came out of that. And this very much relates to the topic that you just brought up around prior authorization and the whole revenue cycle space. And that that is um an agreement with Q. Humana to expand electronic prior authorization uh nationwide. And this is to 180,000 providers, right, Paul?

29:21Unidentified speaker (speaker_0): Yes.

29:22Unidentified speaker (speaker_1): Um, via Athena Payer. And some of the numbers that you've already shared from September, um, and this was with um Fierce Healthcare and that outlet outlook in the space, is that 6,000 providers ran roughly 35,000. EPA transactions with Humana, and 70% plus were completed without any additional work from practice staff. Now, um, that's a little bit grocky, but you know, in my mind, a really big accomplishment, um, especially with just what front desk staff have to deal with at independent practices around the country and beyond. Um so Paul, can you share a little bit more with us um related to some of you know what you've been able to accomplish with with Humana and how did you how did you get them to that payer um to be so willing and and ready to jump on board with you?

30:18Unidentified speaker (UNKNOWN): Well

30:19Unidentified speaker (speaker_0): I I think this gets to the you know an example of kind of the non a non-zero zero sum game, right? So if you think about prior auth, you know Prioroth prior auth is there for a reason. but it's there for a reason for like a small percentage of the transactions. And you know, what we did with Humana and I think what we can do with other payers is we say, wait, we have all of the information about the patient, right? Why don't we send that to you? And let's figure out what the algorithms are and AI can play a good role in this and separate out the 70% or hopefully 80 or 90 or 95% of the things where you you really don't need Pyroth. Like it's like it's it's like it's okay. It's not the reason you put it in place. You put it in place because you go look at your stuff and go, these 5% over here are like weird outliers, and we would like to not pay that because it doesn't make sense. And and you're right. Um, and so it's that notion of now using the power of the computer, the power of the EHR, Or the power of an agreement between payers and and and providers and their ER HR providers to be able to say, let us help you. sort the wheat from the chaff so you can focus all your prior off work on

31:19the people that need it, like the the the outliers, and the rest of us who they're kind of doing the right thing can just go out about our b daily business without all this this tax and overhead. And um I I think that notion can start to be applied more broadly to take all this friction out of the system. It's like you know, there are bad actors. If you're if you're the payer you're sitting there looking across the United States. at all the claims you pay. You know, some percentage, it's you know, single digit percentage of them are claims that are like these are really weird. You know, there's fraud, there's abuse, there's people doing things that they shouldn't to to patients. And it's like there's always a a weird tail. It's the problem is the only way to catch that weird tail has been to impose a lot of work on the people that are doing the right thing.

31:59Unidentified speaker (speaker_1): Yeah, I mean, I've heard had this concept um of an idea of goal. gold carding, which

32:04Unidentified speaker (UNKNOWN): is

32:04Unidentified speaker (speaker_1): sounds very similar to what you were saying, where it's like, you know, just take ninety two percent of providers are just correctly following evidence based care and guidelines. And those people should not have excessive burden to prove um that they're doing the right thing by patients in the form of prioroth. Um, you know, maybe you have your outliers, but in I in my mind, I think a health system would want to know or a provider group would want to know about who those outlier cases are, where it's a doctor that is, you know, going and doing a bunch of things that they shouldn't necessarily be doing that are driving up costs and expense. Because sometimes those are also things that are not good for the patient overall. And so, you know, I love the idea of moving in this direction. I could see even collaboration between the payer and provider around that eight. 8% of like, hey, just so you know, this is happening with this cohort of individuals, here are the cases in which it's happening. Um, and then, you know, you as a practice could start to kind of, you know, get some alignment around it. And then maybe that individual in a in a number of years starts to practice more evidence-based care. And you could see this being a system that just evolves um continuously. What what do you think about that concept?

33:15Unidentified speaker (speaker_0): I mean, I think that's I think that's spot on, right? I mean, you know, the the the reality is, you know, there is somewhere out there is the worst doctor in America, right? We have one of those. Um you know, I it it it's I I in a past life, uh what we used to do, um put up provider report cards in in a in an old manual way to try to, you know, figure out who the good providers are and the bad providers are. And it was driven by payers. It was just what you said. And I would come to a group of doctors and I would say hey, raise your hand if you think you practi you're above average in terms of your your quality of care you provide. And guess how many hands go up in the room?

33:46Unidentified speaker (UNKNOWN): Every

33:47Unidentified speaker (speaker_0): single one, right? It's like, and how many of you take care of patients that are sicker than average? You know, every hand goes up. It's like, you know, okay, I'm here to tell you that's not true. Half of you are below and half of you are above. And I'm gonna and worst of all, I'm gonna tell you which ones. And you know that it's an uncomfortable thing to talk about in a lot of ways, because first of all, as patients, we would like to assume that we're going to the best doctor that there is. And we don't want to think about the fact that there are actually doctors that aren't very good. But in any population of any set of people, there are people that are better than others, right? And uh and I think that what you're talking about is hey, this is actually a way for continuous quality improvement to occur in the healthcare system, in the healthcare delivery part. Like now you can take this thing that is a painful financial thing. um, that everyone just would like to avoid and it's like, like, please make that go away if you're a provider, right? To a wait a minute, this could be used to actually raise the quality of healthcare delivered in the United States and lower the cost of it. Like, isn't that what we're all about? Um, so I I I actually think that if you kinda go a couple of turns, um, that we might be able to start

34:47doing that. Um it's a little challenging 'cause challenging because the payers are all fragmented, right? It's back to my payer days. You know, you show up at a at a at a practice and you know you're 5% of their patients, and it's hard to get them to change their practice because of the 5%. Um so it will require some structural or some collaboration. But I it it feels like the right answer here at some point.

35:08Unidentified speaker (speaker_1): And you know, speaking of philosophical changes, I think it's it's really notable that Athena is thinking about the process

35:14Unidentified speaker (UNKNOWN): of the

35:15Unidentified speaker (speaker_1): patient in this context because, like I said, like everybody else is thinking about the provider for the most part. And, you know, obviously fears abound um whenever you start you mention the patient of like, oh, you know, they're gonna get the the wrong end of the stick, they're gonna willfully misinterpret, the patient's gonna, you know, have access to all this information and be pushing for all these unnecessary tests and procedures. and that will drive up utilization. I mean you saw this with open notes, which you and I remember well, Paul, of um this concept back in the day that, you know, shouldn't a patient be able to access their own physician notes. And you you saw so much anxiety around this idea that like, you know, what if a patient saw like SOB, which means shortness of Subress breath, interpreted, or uh what if a patient saw written down and they were clinically obese, you know, how would they react? And um, you know, Paul, I'm I'm happy to share with that. I've so far have not heard any physicians telling me that it was a disaster when they shared their notes with with the patient.

36:15So do you what's your sort of view of the world as you think about the concept of really giving patients their own information? You talked about that as a right, but then what are some of the ramifications or consequences? of

36:27Unidentified speaker (UNKNOWN): that?

36:28Unidentified speaker (speaker_0): I mean look I I I think you know the the the open notes thing was was very interesting, right? You think the world was going to come to an end uh when we did that. Now I do think there are a couple tweaks to it. If if I if if I were King for day I would say hey you could delay for a couple of days because if you're gonna if you have cancer or you have AIDS or you have something like that, like like like if you have if you have a d uh a a bad diagnosis, I think it's nicer if it comes from the doctor than if you read it on the screen. I just that's my personal view. And if we had a couple days, the doctor would call you, that'd be nice. But other than that, it's been great, right? And and frankly, with all these tools that are out there now, it's even greater because it is hard to understand some of the medical lingo, right? You can misinterpret it or I mean it's not be able to understand the thing. Um and you put it into your favorite, you know, chatbot and it will interpret it for you. And it's really nice. So I think an empowered patient is a better patient. Now, you know, doctors have been dealing with patients that you know got their advice from Dr. Google for a very long time. And you know, sometimes that can be challenging, but in general, I think it is it is better. Um, my uh my brother is a patent

37:31attorney, and what he is finding is that his clients are now much m better clients because they use the GPS. GPT to help think about patent applications.

37:42Unidentified speaker (speaker_1): Oh my

37:42Unidentified speaker (UNKNOWN): God,

37:42Unidentified speaker (speaker_0): they're so much better educated. And you know, it's got pluses and minuses and they want discounts and whatever else. But it's like, I think the same with patients. Like a better educated patient, I think, is a better patient on average.

37:51Unidentified speaker (speaker_1): Yeah.

37:52Unidentified speaker (UNKNOWN): And

37:52Unidentified speaker (speaker_1): most doctors, like when I've gone into the doctor's office, they're like, any questions? And like before, you know, some of these tools, I my questions are probably terrible. And now I feel like I'm asking really good questions.

38:04Unidentified speaker (speaker_0): So

38:05Unidentified speaker (speaker_1): it's not like I think I have all the answers. It's more that I now know enough, you know, to be to be dangerous. Um and I maybe dangerous is the wrong word, but it's like I know enough to be able to like, you know, ask the doctors things that I think will really inform my care and maybe even make them think a little differently. Um, which is uh to me that that can only be a good thing, especially when we're saying to people like, by the way. you have 10 minutes to speak to this person.

38:30Unidentified speaker (speaker_0): Exactly. Well, I mean, and look, and the reality is there's a lot of medicine that is just judgment. Like there's not evidence-based, right? And so you can get the evidence-based stuff out of the way before you get there. Like get all that, get all that in your brain. And it's like, hey, but you know, what do you recommend? Like you do you physical therapy, do we surgery? Like, you know, the the the the outcomes don't seem to be different. Like, and in and that's where you want the clinician, the person who knows you and understands the situation to provide advice, and you can get there much faster. And then you got you know you got 10 minutes. So

38:57Unidentified speaker (speaker_1): Yeah, make it make it work, make it count. Um okay, so I want to touch a little bit more on your background because you've been an entrepreneur in the past. You founded a company, you were the CEO of a company called Patient Keeper, um, which you sold to HCA in 2014. Um, and you've been working, you know, in in a variety of senior roles at um places like McKesson and Boston Consulting Group. It seems like the idea of physician workflow is uh is one that has kind of held true for a lot of your career. Um, how did you end up like caring about that so much that you wanted to dedicate decades of your life to it? Um, and then what's what's like a uh enduring legacy? that you'd want associated with your own time, kind of pushing for change here that, you know, in 10, 20 years, like maybe something is different because you've been a piece of that change?

39:49Unidentified speaker (speaker_0): Um, that's a great question. So it's it's a weird industry to kind of kind of devote your life to. It's it's more popular today than it was. But I got my start actually when I was in high school. Um my dad was uh a general surgeon and he left academic practice to found a pr a solo practitioner uh when when when I was uh like a I guess a freshman in high school. And guess what? He couldn't figure out how to get paid. Um and this was a time when like there weren't really computers or like Apple two E's and things like that. And you'd fill out claim forms on an IBM selectric typewriter and send it off and money would hopefully come back. Um and so I helped him figure out how to get paid. And I then being the crazy person that I was, uh, wrote a whole bunch of software to get him paid. And we would now call that a practice management system. And then I wrote it and got

40:34Unidentified speaker (speaker_1): Wait, how old were you?

40:35Unidentified speaker (speaker_0): I was uh fifteen, sixteen years old.

40:38Unidentified speaker (speaker_1): So you built a practice management system when you were like fifteen, that's what

40:41Unidentified speaker (UNKNOWN): I

40:42Unidentified speaker (speaker_0): did. And in fact I was the first person to send an electronic claim in the state of Florida. Oh

40:47Unidentified speaker (UNKNOWN): my

40:47Unidentified speaker (speaker_1): God.

40:47Unidentified speaker (UNKNOWN): That is my real

40:48Unidentified speaker (speaker_0): distinction in life. Uh I said

40:53Unidentified speaker (UNKNOWN): Secret talent.

40:56Unidentified speaker (speaker_0): Secret notable thing. And so I and I built a business when I was in high school and I took some time off before I went to college, uh, selling practice management systems to doctors in northern Florida. And so uh I had a whole bunch of clients. The the weird coda to that is, you know, I then went off to college and went to BCG and did all this other stuff. It wasn't until my second son was born in 2002 that I got the last of my clients off of an Apple IIe-based practice management system, which is kind of nuts because you couldn't like like the Apple IIE was long since out of like you couldn't even get the hardware, right? Yeah, I yeah, anyway, a long story. So that's how I got into it. Uh I'm very passionate about physician workflow, uh, both because of my dad and because. of all of the early people that influenced my life, which were mostly a bunch of doctors and a variety of specialties, and they were kind of mentors to me as I built my business and got to know their businesses, because as a practice manager system, you kind of helped them run their business, which is also a little strange to be doing when you're in high school. Um, but uh but it's really to their legacy. And and they all um oh uh it's one other thing. I was a pre-med pre med in college. And in my junior year, a group of my clients got together and met with me and talked me out of being a doctor.

42:12Unidentified speaker (speaker_1): That's

42:12Unidentified speaker (UNKNOWN): so

42:13Unidentified speaker (speaker_1): funny.

42:13Unidentified speaker (speaker_0): They did it because they thought medicine was, you know, start all all this we've been talking about, all these challenges were mounting, and uh and they thought that I could have a bigger impact on on healthcare by uh doing something more entrepreneurial. And so it is it is to them that I dedicate what I do. Um and uh hopefully that if they look down now, they'd say, Yep, okay, that was good.

42:35Unidentified speaker (speaker_1): And you have um you sound like you have two sons?

42:37Unidentified speaker (speaker_0): I do, yes.

42:38Unidentified speaker (speaker_1): Okay, so and they they maybe are all grown up now because you said 2002, but

42:42Unidentified speaker (UNKNOWN): they are

42:43Unidentified speaker (speaker_0): they they are out and about and and and in between members of society, yes. So

42:47Unidentified speaker (speaker_1): I have a I have a son who was born in twenty-one twenty one. and um, you know, let's imagine into the future when he's uh becoming an adult and going off to college. Uh what do you think will be different about physician workflow then?

43:03Unidentified speaker (UNKNOWN): Oh

43:03Unidentified speaker (speaker_0): my god. Um you know, I kind of hope that the doctors are just taking care of the patients. Like you can imagine a world where, you know the doctor is briefed by a computer in a thoughtful way, maybe, you know, like some screen as they walk in, or who knows, maybe some neural implant. And then most of their time, like 95 to 99% of their time, is engaging with patients and providing care. As still a human, right? Even with probably all the AI stuff that people are going to have. But I think that human to human connection is. is why people go to med in the medicine and it is why people go to doctors. And I think even with the best, you know, AI agent and assistant, we're still going to want to receive care from a compassionate, thoughtful, caring provider. And I think if we if it if we've got all the stuff that we've been talking about kind of just it's just evaporated, right? There it's behind the scenes, it it just works. And the doctors can do that. And the people, the practice staff can focus on the patients when

44:05they come in. Like imagine that. And imagine no faxes and no clipboards in my

44:09Unidentified speaker (UNKNOWN): process.

44:10Unidentified speaker (speaker_1): I I truly can't. And yet, like every other industry seems to have moved on. So it's really just healthcare. For a while it was healthcare and the drug dealers, but I think even they now have signaled and and telegram and and other such such technologies. So it's really just us at this point. Um okay. So NVIDIA is in the news at this moment in time, you know, not really for great reasons, but um now prior CEO, Jensen Wang, was one of the, I would say, most vocal people saying that, you know, open source um is really is really worth looking at versus using some of these closed models. Um healthcare seems uh relatively split on this, as I've I've talked to people on it. Um, I think one of the reasons that people like open source is because you, in theory, have more control over your uh data. Um, that then there are you know safety concerns and and um and I think OpenAI and Anthropic have made real headway in the space. They're hiring up big teams in healthcare. Um you're obviously working with Anthropic, OpenAI is is working with Epic, and I think this. is just the beginning in terms of partnerships that we're gonna see across the industry and the ecosystem. So,

45:20any thoughts just to this debate? And what have you been seeing and hearing?

45:24Unidentified speaker (speaker_0): Um, you know, and just with the caveat, this is probably only valid for the next like you know week because the world keeps changing so much. Um, but um so the the foundational models are great. I mean, we work with both OpenAI and Anthropic, we use their models and um and boy, we can do things that we've not been able to do before. However,

45:43Unidentified speaker (UNKNOWN): the

45:43Unidentified speaker (speaker_0): thing that I look at is they change all the time. And when you have a system as complicated as Athena and we have all this stuff built on AI and they change their models, we essentially have to go revalidate.

45:54Unidentified speaker (UNKNOWN): Like

45:54Unidentified speaker (speaker_0): we've got all the data protections and things like that. That's not that's not an issue. Like but but the like we have to revalidate our software when those models change. and they change all the time. And so I look at open weight to say, hey, I want control over when things change. And I don't, frankly, you know, to read faxes, like last generations models read faxes just fine. I don't need to go solve some crazy math problem that no one solved before. I just need to read a fax and understand what it is. And the current models do a fine job of that, right? So we're looking at it from a hey, we want to get stability. Um, and because we have to assert, like you know, when when we go to our customers, we're like, hey, our AI works, and in order to assert that, we have to go validate it, and if it keeps changing, we have to validate it a lot. And so that's the the the what that's where I'm really starting to focus on open weight models, and we are we we're we're actively exploring where we use them and look in some areas where we do need that cutting edge frontier model, great, we'll use those. But for many of the things we're doing, the current generation or even yesterday's generation of models works just fine, thank you. Um and so that's kind of where we're we're we're at least we're thinking, but you know, you if you ask me that in in a month, I may have a different answer.

47:07Unidentified speaker (speaker_1): Yeah, it seems like it's it's kind of a mix across the industry, which is really what you were saying. Um and it depends on the use case. Like I think that's our 100% true. What do you think is the AI story? So I feel like there's like a million AI narratives like out there. Like every some people think the story is a safety story, like the dystopian, you know, are we all gonna die in five years story? Um, some people, and I I'm kind of on this in this camp, um, think I there's like a more interesting story around economic entanglements amongst these companies and and can concerns about the domino effect, like if NVIDIA falls, what else falls with it? Um, and I'm closely following that um portion of the story just because of how much AI is like propping up a lot of the US economy. Some people are really focused on the story, you know, the global story around the On the big geopolitical powers that be, um, you know, whether it be China or the US. Um, some people think of it as a story where we really

48:07have barely scratched the surface in terms of what industries can be doing with AI and that we should just pause development and let like real application work happen. There's lots of different narratives out there, and everybody you ask has a has it has one that they think is the real story that's not being told enough. I'm curious what your real story is. that you don't you don't think is being talked about enough, but is is uh something very defining about AI.

48:34Unidentified speaker (UNKNOWN): I

48:34Unidentified speaker (speaker_0): don't know if I've got one that hasn't been talked about enough, but I where where where I where I focus my energy is is is is twofold. Um most importantly, reimagining the software. Like just reimagining software. And I think you know the SaaSpocalypse is out there. There's all this like people reimagining software in in all sorts of dooms ways.

48:52Unidentified speaker (UNKNOWN): I

48:53Unidentified speaker (speaker_0): think in our industry, a lot of what we talked about, the reimagining is wow, I can actually bring software to our primary user, the physician, in a way that delights them for the first time. Like that is to me, that that's what gets me super excited. The second thing, which is really cool, is that we can build software differently. differently. And for us, differently is solving the the fundamental problem we have in healthcare IT. You ask, you know, why is healthcare software kind of not so good? And the the reason I believe fundamentally is because none of the people that write our software use it. I have no no engineers that are out taking care of patients. I have a chief medical officer who takes care of patients every now and then on an Athena platform, but that's kind of it, right? Whereas if you work at you know Facebook or you work at uh at Netflix, like you use your software.

49:38Unidentified speaker (speaker_1): Yeah. Right?

49:39Unidentified speaker (speaker_0): And so and you have some idea of what it's like. I I watch movies. Okay, I have some idea of what it would be like to build a thing. But it's very hard to get an engineer to imagine what it's like to be uh you know a pediatrician walking in to see a patient and taking care of a patient. Like that's a very hard thing for them to imagine. And so the software they write is not reflective of the the process. What we're doing now is we are engaging directly with providers. And we're sending engineers out to practices. And they sit in the practice with the doctor as they're caring for patients and they work together to imagine what the software could be and they see it right away. Before it was like, Yeah, we go do that, and then we go back and three months would go by, we come back and say, Hey, is this what you meant? And they'd be

50:19Unidentified speaker (UNKNOWN): like,

50:19Unidentified speaker (speaker_0): No. And then we go back and three months It's later. like Like now it's like, no, no, I can show you right now what we said or what I thought you said, and you can go, Oh no, well, I know I said that, but I I meant this. And the rate of innovation and the rate of of good software building is just dramatically increasing. And at our Thrive conference that you that you mentioned, um, we actually invited our customers, we brought a bunch of engineers to Thrive, and I had my entire engineering team, all like three thousand people around the world, um, on call and our customers would come down and they would say, Hey, I use your software every day and this thing is annoying, or I'd like it to do this thing, or why can't it do this thing? And they gave us almost a thousand ideas of

50:58Unidentified speaker (UNKNOWN): things

50:58Unidentified speaker (speaker_0): that the software could do better, and we've already built 200 of them.

51:01Unidentified speaker (speaker_1): Wow.

51:02Unidentified speaker (UNKNOWN): So

51:02Unidentified speaker (speaker_0): imagine going to Thrive with a thing, it's like, oh God, I hate it when I when one what we've got actually the the the the the clinic I was talking about, the um the that does all the menopause care, they have to send out specialtys specialty pharmaceuticals a lot. And there's like one field where you can't use a macro, a macro to expand a bunch of things, and they need a macro there, and like it isn't there. And like they're like, I have to type this every time. Please, God, put a macro in. And we did. And so imagine like going back and going, hey, I wanted to thrive. I learned all this cool stuff. And by the way, there's a macro in that field. And now we can like send out specialty scripts much faster. It's like it's like that kind of thing just changes the way we can build. And you know it's such a complicated industry, and there's so many workflows. Um and so I'm really, really excited about that.

51:41Unidentified speaker (speaker_1): I love that. Especially given I've been to so many of these types of events and they always end up being sales pitches um and sales focus versus product and strategy focus, which it to me is like really where the opportunity lies is to like listen to your customers and hear what they want and and build for them. Okay, so one narrative that I think also comes up a lot, which again I think is a big one and a real one. is how human jobs and labor will change. So last question for you is as you think about, you know, building teams at Athena and all the internal work you have to do to make sure you have the right butts and seats, especially in an AI era. What are some of the skills that you highly prize today that maybe weren't as important a few years ago? And then how is how is the job, you know, like what jobs are changing shape and shifting within Athena, where maybe, you know, you had three thousand engineers or you have three thousand engineers now, and two years ago you would have needed six thousand, or have there been any substantial changes that you could speak to across the organization where you've actually done the work of saying, you know, these teams need to be more people, these teams could be less people, and like starting to restructure because of AI.

52:49Unidentified speaker (speaker_0): It's I mean it it is truly fundamental. And you know, the problem is that we have people that are really good at, for example, writing code. Like, what do they do? I write software. I'm really good at synthetic formation of Java. I'm good at that. Like, it's like we almost don't need those people anymore. And that's really sad, right? Because, like, hey, what do I do? I write code. Well, okay, but we don't need people that just write code. What do we need? We need people that understand problems, that can architect. systems, and ironically, that can manage essentially people, but manage agents. My head of engineering and I, um, I've got Richard Bardwell, we were all excited because we're like, hey, we can all become individual contributors again. We can all write software again. And then we were like, no, no, no. Actually, the task is management. It's that I give a task to an AI agent, it does the task, and it does it probably wrong the first couple times. of And you have to be astute enough to know it did it wrong, just like an employee, and give it guidance and have it go off and fix it. It just does it really fast, but it's the same cognitive process. And

53:49what we're seeing is that some of our people are really good at that, the kind of future managers of the world. And we have people, unfortunately, that are not good at that and need to learn that skill. Um And that's where I think there's going to be a lot of disruption because there are, frankly, a lot of people that went to computer science school and became coders. And what we needed them to become was solution architects and you know user experience people and and and problem knowledgements, you know, surgery matter experts. Um and and some of them did and some of them didn't. And I think that there's going to be a a a real change in the composition of what an R RD and D team looks like.

54:23Unidentified speaker (speaker_1): And just um last question then any practical tactical advice just on people who are wondering, you know, listening in and maybe thinking, oh gosh, my skills sound a bit redundant. Like, do you should you take some a sabbatical and go talk to your um boss about it? Should you, you know, should you be coding at the in your evening hours differently? Like, what should you actually do to stay relevant um in this new economy?

54:50Unidentified speaker (UNKNOWN): I mean

54:51Unidentified speaker (speaker_0): I think that the the rewards are going to come from the people that can think more broadly and more integratively. I think it is it's actually the return of the generalist or the you

55:02Unidentified speaker (UNKNOWN): know

55:02Unidentified speaker (speaker_0): the liberal arts education that when debated. Like, like it's it's no longer your specialized skill. Go learn how to program in you know Java or you know, C sharp or whatever. It's you don't need to know those things anymore. It's it's now how. do I think conceptually about putting together a system and what are the components and what are the problems. And and so it really is is more of the the kind of generalist problem solving skills, which are unfortunately the harder ones to learn. If I said go, you

55:28Unidentified speaker (UNKNOWN): know,

55:28Unidentified speaker (speaker_0): go learn how to program in Java, you go take a course in that like this is the classic liberal arts broad thinker people that are going gonna to be rewarded. So um think broadly and you know you can experiment on your own. Like learn learn how to manage your agents. I've got agents doing all kinds of stuff. Learn how to manage them. Like you can you can do that right now. And and you'll see that I mean I I just had a fun exchange I was mentioned setting up our our new dot system and I'm like, I want you to book travel. It's like I don't have access to uh to travel. I'm like, well why don't you have access to travel? And it's like, oh I'm sorry. I do have access. It's like it's it's like you know you you you have to learn how to manage the folks. Uh Um and I think learning how to manage your agents is is is is a good way to start.

56:06Unidentified speaker (speaker_1): Yeah, totally. Um, I'm I I've become, as everybody knows who knows me, like such a big proponent of I've always been a proponent of of liberal arts education, but even more so today, because I feel like critical thinking skills, judgment skills, working through problems where you're thinking about first order, second order consequences and how to mitigate those things. Like I think those skills are uh even more important than than ever. Um Paul, thank you so much for joining me on Lifeers. It's been such a real treat to to speak with you. And um, you know, if anybody's listening in and wants to get in touch, are you a a LinkedIn guy? Are you like an ex guy? Like what's your uh preferred method?

56:44Unidentified speaker (speaker_0): I'm uh mostly a LinkedIn guy. I I'm

56:46Unidentified speaker (UNKNOWN): think

56:47Unidentified speaker (speaker_0): I'm not hugely social media, but uh yes, I'm a LinkedIn guy.

56:50Unidentified speaker (speaker_1): Probably for the best.

56:52Unidentified speaker (speaker_2): That's a wrap on Lifers. If you know someone else grinding it out in healthcare, send them this episode. And if you want more unfiltered takes on digital health, check out the second opinion newsletter. Link in the show notes. Lifers is supported by Hippocratic AI, which builds safety-focused Focused AI agents that reach the patients health systems don't have the staff to reach. Learn more at Hippocratic. HippocraticaI.com and hear CEO Munjar Shah break it down on a recessent episode of Lifer's, linked in the show notes.

This page was produced automatically: the audio was transcribed with nvidia/parakeet-tdt-0.6b-v3+sortformer-v2.1 and interpreted by Qwen3.6-35B-A3B-UD-IQ3_S.gguf. Transcription and interpretation can contain errors; always confirm against the original episode. Nothing here is investment advice.
All insights