Latest / CareTalk: Healthcare. Unfiltered. / Will AI End The Payer And Provider Billing War? w/ Hamid Tabatabaie, CEO, CodaMetrix

Transcript

Will AI End The Payer And Provider Billing War? w/ Hamid Tabatabaie, CEO, CodaMetrix · CareTalk: Healthcare. Unfiltered.

• 0:00
  SPEAKER_01: When somebody orders a chest x-ray to rule out pneumonia for a patient, and the radiologist is reading that chest x ray, they rule in or rule out the pneumonia, but they're responsible for, and you and I count on them for writing the rest of the diagnosis they find.
• 0:19
  SPEAKER_01: And if they find an enlarged heart or metastasis of cancer in the lungs, they'll comment on that.
• 0:26
  SPEAKER_01: And if there's any action to be taken, they advise on that.
• 0:30
  SPEAKER_01: But in order to get paid for that chest x-ray, all you need to write is chest x-ray ruled out pneumonia.
• 0:37
  SPEAKER_01: There's no mention of the enlarged heart because it's a fee-for-service environment, and the payers assume that the patient is going to come back for an enlarged heart.
• 0:47
  SPEAKER_01: Then we'll catch up with that point.
• 0:49
  SPEAKER_01: We don't need to be bothered with the facts today.
• 1:25
  SPEAKER_00: Welcome, Hamid.
• 1:27
  SPEAKER_00: Thank you.
• 1:27
  SPEAKER_00: Thank you very much for having me.
• 1:30
  SPEAKER_00: So Hamid, before we get into the wonderful world of simplifying and lowering costs in healthcare and getting coding right, how did you get here?
• 1:39
  SPEAKER_00: Like what's your backstory in terms of getting involved in healthcare?
• 1:46
  SPEAKER_01: It started right after I came out of college.
• 1:49
  SPEAKER_01: I was an engineer and I started working at a hardware company at that time.
• 1:54
  SPEAKER_01: That was high tech in the beltway around Boston.
• 1:58
  SPEAKER_01: And I got assigned to health systems around Boston, which as you know is a pretty dense healthcare environment.
• 2:07
  SPEAKER_01: My job was to help doctors and clinical workers to use hardware in the early 80s and develop software on it.
• 2:18
  SPEAKER_01: And a few years later, wet behind the ears, I asked the CEO of this multi-billion dollar company if he thought it would be a good idea to start a healthcare division.
• 2:31
  SPEAKER_01: After he looked up and said, I don't know who you are, but go ahead and get it started.
• 2:36
  SPEAKER_01: I became the director of a healthcare division for that hardware company, and we basically sold particular products to Meditech and Epic and at that time HBOC and so forth.
• 2:52
  SPEAKER_00: So that was the earliest stage, though, of course.
• 2:55
  SPEAKER_00: Absolutely.
• 2:56
  SPEAKER_00: Just to give people a sense of how old we are, Hamid, that there was a period before which electronic medical records really didn't exist.
• 3:06
  SPEAKER_00: And those are the early, still very large, Epic being the dominant players in the field of electronic medical records, which which I think doctors and patients think should be about health care, care, and quality, but were actually the basics of health care billing.
• 3:24
  SPEAKER_00: And that was how they were pitched to hospital administrators and CEOs.
• 3:28
  SPEAKER_00: And so you were early on in this sector.
• 3:31
  SPEAKER_00: I mean, perhaps I mean, th this sort of dates you, Hamid.
• 3:34
  SPEAKER_01: Aaron Ross Powell It definitely dates me.
• 3:38
  SPEAKER_01: We were early on enough that at that time it was fashionable to have best of breed.
• 3:44
  SPEAKER_01: So you bought a lab system from Serner, you bought a pharmacy system from someone else, and that was the best thing to do.
• 3:52
  SPEAKER_01: And then in the early 2000s, things turned around, and it was you wanted one monolithic system so you can recognize the patient no matter where the care was delivered and so forth.
• 4:07
  SPEAKER_01: So in 93 I started my first company, and that was to help implement Meditech and Epic uh facilities because PCs were brand new and everyone was coming from the terminal side.
• 4:24
  SPEAKER_01: So we needed to help people with all the transformation and change management and so forth.
• 4:31
  SPEAKER_01: And then three companies later, Mass General and Brigham, were both customers of and eventually became investors in my prior company, whose job was to find imaging exams for any patient, no matter where the images were.
• 4:49
  SPEAKER_01: We practically had a directory of where a patient's images may be, even across health systems.
• 4:58
  SPEAKER_00: Which is critically important because A, it's it's an elemental part of treatments for cancer and and uh and obviously orthopedics and others.
• 5:10
  SPEAKER_00: And it was a real problem when you were trying to crack it.
• 5:12
  SPEAKER_00: It sounds it sounds like it should be simple, but it's actually rather hard because of all these different systems and the way they're configured at a hospital and a doctor level, um it it it sort of feels like you were early on in the somewhat sysophyan task of trying to make healthcare data intelligible for practitioners and administrators.
• 5:37
  SPEAKER_00: Um why didn't this frustrate you so much that you went back to basic hardware and software, Hamid?
• 5:43
  SPEAKER_00: Like what what kept you going?
• 5:46
  SPEAKER_01: I think it's purely a genetic malfunction.
• 5:49
  SPEAKER_01: Uh once you've done a few companies, you're golden for punishment.
• 5:54
  SPEAKER_01: Yeah.
• 5:55
  SPEAKER_01: I am really interested in fixing things and or creating things that help people fix things.
• 6:03
  SPEAKER_01: Uh that's a very loud voice in my head.
• 6:07
  SPEAKER_01: And that resulted in me getting to codometrics.
• 6:11
  SPEAKER_01: Uh, when I was my tenure was finished at the last company, Mass General asked me if I wanted to take a look at what they had developed internally for billing purposes.
• 6:20
  SPEAKER_01: And that was uh essentially the early codometrics.
• 6:26
  SPEAKER_01: And I, after doing six months or so of looking at it, very quickly recognized that the billing side of things is a mess.
• 6:38
  SPEAKER_01: Everybody really is.
• 6:40
  SPEAKER_01: And some of it by design, some of it just tradition mostly, but a lot of it because the clinical data is separated from billing data.
• 6:51
  SPEAKER_01: Trevor Burrus, Jr.
• 6:51
  SPEAKER_00: Yeah, just so people can understand it.
• 6:53
  SPEAKER_00: So you know, roughly 15 to 20 percent of all of the$6 trillion we spend on health care is administrative costs.
• 7:03
  SPEAKER_00: A lot of that is in the billing, rebilling, E.O.B., just figuring out and the and the ground war between insurance companies that only want to pay for certain things and don't pay for a lot, the complexity of benefit structures, and then on the doctor and the hospital side, the things that they can bill for, and within that there are the clinical pieces, you know, what your what your diagnosis is, and your every everything from the specifics there to the doctor's notes and the administrative side of just getting the bill out the door and having it conformed to an insurance contract.
• 7:43
  SPEAKER_00: It's both the one of the probably the biggest deadweight costs on health care, but it's uh but its complexity is embedded in every doctor's office, every hospital, in fact, any clinical facility.
• 7:57
  SPEAKER_00: And getting at that clinical and administrative piece is, again, just dense and kind of hard.
• 8:03
  SPEAKER_01: I mean, absolutely is.
• 8:17
  SPEAKER_01: And it has continued to be a very manual process.
• 8:21
  SPEAKER_01: So not only is it wasteful in terms of how much how many people it takes to do it, but it's also very inefficient even when uh the people are trying to do their best.
• 8:32
  SPEAKER_01: The data is all over the place.
• 8:35
  SPEAKER_01: So the reason I started codometrics is because of honestly, one lunch with CEO of the Mass Journal Physician Organization.
• 8:44
  SPEAKER_01: I didn't want to do a revenue cycle.
• 8:46
  SPEAKER_01: It looked too messy to be helped at the time.
• 8:50
  SPEAKER_01: Um he said, with AI, we have a chance to bring coding back to its clinical origins.
• 8:59
  SPEAKER_01: In the 1980s, managed care came by, and very quickly, codes that were invented or created for purposes of clinical follow-up with patients, whether it's diagnosis coding or it's procedure coding as to what healthcare is doing and why are they doing it, became elements for billing.
• 9:21
  SPEAKER_01: They became very much a dumbed-down version of themselves in order for billing to be able to be more successful.
• 9:30
  SPEAKER_01: And the reason was uh it was too costly for human beings to read an entirety of a chart to be able to really be accurate, quote unquote.
• 9:46
  SPEAKER_01: And therefore, it we would be uh swimming upstream if we pitched it that way.
• 9:53
  SPEAKER_00: Trevor Burrus, Jr.
• 9:54
  SPEAKER_00: Just so people can understand, I mean, I mean a chart isn't just a one-page simple set of diagnosis codes alphabetized with numbers for cost.
• 10:04
  SPEAKER_00: It can go on for tens or hundreds of pages with with things with procedures, things that happen to you in a complex case, your lab values, your you know, your diagnosis when you come in through the emergency room, to how you're handled on the floor, you know, how your numbers are changing, and all of which ties into billing and much of what which is open to both complexity between and manual errors.
• 10:35
  SPEAKER_00: And and and it's uh in a revenue cycle, which you talked about, is probably the biggest poorly managed administrative costs in a hospital.
• 10:47
  SPEAKER_00: And coding, which you're working on, is the biggest opportunity, frankly, the biggest cost within revenue cycle that I think drives hospital administrators and CEOs crazy.
• 11:01
  SPEAKER_01: Every study I've done shows that coding is about a quarter, if not larger, than the entire administrative cost of billing and so forth.
• 11:13
  SPEAKER_01: So that one sentence got me started, and uh we've stayed focused on let's fix that problem.
• 11:20
  SPEAKER_01: Let's go from subjectivity to very objective coding.
• 11:25
  SPEAKER_01: I'll give you an example.
• 11:27
  SPEAKER_01: Um because I had a radiology background, this one was the one that amazed me.
• 11:32
  SPEAKER_01: Uh when somebody orders a chest x-ray to rule out pneumonia for a patient, and the radiologist is reading that chest x-ray, they rule in or rule out the pneumonia, but they're responsible for, and you and I count on them for uh writing the rest of the diagnosis they find.
• 11:53
  SPEAKER_01: And if they find an enlarged heart or metastasis of cancer in the lungs, they'll comment on that.
• 12:00
  SPEAKER_01: And if there's any action to be taken, they advise on that.
• 12:04
  SPEAKER_01: But in order to get paid for that chest x-ray, all you need to write is chest x-ray ruled out pneumonia.
• 12:11
  SPEAKER_01: There's no mention of the enlarged heart or because it's a fee-for-service environment, and the payers assume that the patient is going to come back for an enlarged heart, then we'll catch up with that point.
• 12:24
  SPEAKER_01: We don't need to be bothered with the facts today.
• 12:27
  SPEAKER_01: So that simplicity uh really causes redundancy.
• 12:33
  SPEAKER_01: As a result, when they're trying to take care of a patient on a long uh longitudinal record uh or at risk, they hire nurses to pull charts and recode things.
• 12:46
  SPEAKER_01: Uh so it's not only wasteful to start with, it becomes more costly as it goes.
• 12:52
  SPEAKER_00: Aaron Ross Powell And it's also, I think one of the things that I find attractive about your approach is much of this information is standard if you can find it and should be machine readable or interpretable, which would suggest that you could standardize and simplify and pull errors out and remove that manual error piece.
• 13:17
  SPEAKER_00: Is there a other than lowering cost, is there a clinical benefit to the patient if we do this, if we leverage your company and AI and machine learning to get the information right?
• 13:30
  SPEAKER_01: No doubt.
• 13:32
  SPEAKER_01: When codes can tell the story of the patient, uh gotta take an example of an inpatient stay.
• 13:40
  SPEAKER_01: They are seen by many, many different physicians depending on the reason for their admission.
• 13:46
  SPEAKER_01: And these doctors, as much as they collaborate together and with one another, they're pretty busy people.
• 13:54
  SPEAKER_01: So if the codes aren't accurate and the codes aren't correctly reflecting the acuity of the patient, you're at the mercy of the doctor missing reading one of the 20 pages of uh yesterday's notes.
• 14:10
  SPEAKER_01: So uh accurate coding leads to a lot of goodness.
• 14:15
  SPEAKER_01: One of them is better patient care, certainly, no doubt about it.
• 14:19
  SPEAKER_00: The other is it will help payers create the Before you get there, Haminda, you made you made a very generous comment about doctors collaborating.
• 14:27
  SPEAKER_00: I don't see how doctors well, A, I think it's hard for doctors to collaborate given the way practice is set up, but it's impossible if they don't have the standard information.
• 14:37
  SPEAKER_00: It really is.
• 14:38
  SPEAKER_01: And uh you often find out that the patient is reminding a doctor of what they told the doctor yesterday.
• 14:46
  SPEAKER_01: Uh I happen to be a frequent flyer uh as a result of some genetic things, but we've all visited patients, and invariably you come across a significant number of circumstances where the patient says, I can't believe the doctor didn't know the thing I said to the other one.
• 15:04
  SPEAKER_01: Uh so codifying information can do anything but to simplify extraction of information and uh reflection on it.
• 15:14
  SPEAKER_00: So before we get into the fund payer side of this about how you're going to solve the payer provider arms war over billing, uh are you seeing the results of using codomatrics in hospitals, improving clinical quality?
• 15:29
  SPEAKER_00: And how do you how do you aim to show that?
• 15:31
  SPEAKER_00: Because I think that would be quite compelling for patients.
• 15:35
  SPEAKER_01: We keep track of every behavior of the system from data elements all the way to any predictions all the way to submission.
• 15:46
  SPEAKER_01: And we count the number of times that codes that arrived to our system, maybe upstream coders coded it, maybe physicians coded it.
• 15:56
  SPEAKER_01: We show in an analytical way how many changes took place.
• 16:01
  SPEAKER_01: And then you can categorize those changes to risk factors.
• 16:05
  SPEAKER_01: Some things risk the clinical treatment of the patient all the way to some things risk the reputation of the health system.
• 16:14
  SPEAKER_01: Never mind the denial and the dollars and cents and so forth.
• 16:17
  SPEAKER_01: So we can show where we've been incredibly effective, and invariably we are.
• 16:22
  SPEAKER_01: We handle roughly about 35 of the largest health systems in the country.
• 16:29
  SPEAKER_01: And when you take their pulse, somewhere around$200 billion of net patient revenue, as the lingo goes in our industry, which is how much they charge for patient revenue separate and distinct from anything that has to do with research or administration of medicine and so forth.
• 16:51
  SPEAKER_01: And there's a lot of data going through, and the patterns are also striking.
• 16:56
  SPEAKER_01: The patterns are roughly 56% of the time, the diagnosis codes are specific enough to be actionable.
• 17:06
  SPEAKER_01: Trevor Burrus, Jr.
• 17:07
  SPEAKER_00: So you're driving faster and clearer decision making by doctors for patients.
• 17:12
  SPEAKER_00: No doubt.
• 17:14
  SPEAKER_00: And now let's get into a little bit of the payer-provider stuff.
• 17:19
  SPEAKER_00: Because one of the promises of codomatrix for a hospital is you're gonna lower my costs of coding because you're gonna take out the manual errors and automate, and so I'll be able to actually get through more claims to get them accurate and to have these clinical insights.
• 17:36
  SPEAKER_00: But hopefully I'm gonna be able to bill more effectively.
• 17:39
  SPEAKER_00: Insurance companies are very public right now.
• 17:44
  SPEAKER_00: They're saying there's all this up coding happening.
• 17:47
  SPEAKER_00: And I think a lot of hospital administrators that I talk to believe they're just catching up to where the insurers are, but the insurers are also putting billions into their own AI and machine learning to to I guess to be more generous to them, to manage the new, more accurate billing and to basically match their AI to yours to keep their trends down.
• 18:15
  SPEAKER_00: What's happening there?
• 18:16
  SPEAKER_00: Who's winning?
• 18:18
  SPEAKER_00: And how do you think this is all going to work out between managed care and the the managed care, if you will, hospital war over billing?
• 18:28
  SPEAKER_01: I'm afraid the war will go on for some time, particularly the way it has just started.
• 18:35
  SPEAKER_01: I believe payers are ahead of providers in deploying AI and interrogating the claims that come through.
• 18:45
  SPEAKER_01: Their interest is to find anomalies in either patterns or in lack of compliance with their guidelines.
• 18:56
  SPEAKER_01: As a result, their AI systems are tuned for f finding spelling errors as far as I'm concerned, because they don't have an incentive to take a look at the entirety of the patient record to decide whether or not the doctor made the right decisions and or whether the doctor's documentation was accurate, the coding was poor.
• 19:22
  SPEAKER_01: So on the provider side, we're making the codes accurate now.
• 19:29
  SPEAKER_01: Therefore, when providers were getting penalized for poor coding, it's going to translate into better claims, potentially better reimbursement.
• 19:40
  SPEAKER_01: So before we get into the debate of accuracy, it's just a question of A, do we have a subjective versus objective?
• 19:50
  SPEAKER_01: We need an objective way to measure accuracy.
• 19:54
  SPEAKER_01: B did the documentation support the codes?
• 19:59
  SPEAKER_01: If they did, then it's not overbilling.
• 20:02
  SPEAKER_01: If it didn't, then that's a problem.
• 20:05
  SPEAKER_00: Just so folks have a feel for this, I mean there is a tremendous amount of denial and delay simply related to human errors like misspelling or incomplete records that presumably with AI and uh with a machine, you can you can fix that before or never or ha or not you know not hit that before the bill goes out the door.
• 20:30
  SPEAKER_01: Oh doubt.
• 20:32
  SPEAKER_01: And unfortunately, there isn't straight incentives to fix it on both sides simultaneously.
• 20:42
  SPEAKER_01: Payers would want to fix it to get rid of their administrative overhead for the parts that they're gonna pay for, anyways.
• 20:50
  SPEAKER_01: And providers have an incentive to do it because human costs are gone up, uh number of transactions are higher, and doctors are sick and tired of dealing with administrative burden and so forth.
• 21:04
  SPEAKER_01: So these are two different objectives.
• 21:08
  SPEAKER_01: And the thing that brings these objectives together is the commonality of accuracy driving better care.
• 21:17
  SPEAKER_01: Because uh at least we've all drank the Kool-Aid that better care means lower cost, ultimately.
• 21:26
  SPEAKER_00: Aaron Powell Well, and and that better care uh let's start with it, provide it it means we live longer and live a healthier have a healthier lifespan.
• 21:32
  SPEAKER_00: So I think we all agree that that the human a human benefit is is is worth a reasonable investment.
• 21:40
  SPEAKER_01: It really is.
• 21:40
  SPEAKER_01: I'm being cynical, it also means more insurance payments over time, but yes.
• 21:46
  SPEAKER_00: You've you've raised, I think, uh almost$110 million.
• 21:50
  SPEAKER_00: You're at the cutting edge of AI, at the cutting edge of what one of the first practical investments are artificial intelligence, machine learning in hospitals that provides an immediate benefit.
• 22:01
  SPEAKER_00: How do you think about the world right now from how important AI is to running a hospital and improving your business specifically given the rapid pace of advancement of the large LLMs, both kind of open weight and closed weight?
• 22:17
  SPEAKER_00: I mean, whether it's here or or but you know in the U.S.
• 22:20
  SPEAKER_00: or outside, there's so much happening.
• 22:23
  SPEAKER_00: How big is the this this AI revolution we're in and how important is it to healthcare in specifically your business?
• 22:34
  SPEAKER_01: Specifically in our business, uh coding is uh twenty to thirty billion dollar industry and bringing the cost for that down by magnitudes of four, five, six is not at all unreasonable.
• 22:50
  SPEAKER_01: Yeah, that's a lot of billions.
• 22:53
  SPEAKER_01: Uh the problem is not just technology.
• 22:59
  SPEAKER_01: Yes, LLMs are phenomenal inventions, and every other day there's a new uh progress being made in one front or another, but they're not tuned for this battle between the regulations and the messiness of data.
• 23:19
  SPEAKER_01: They need people like Codometrics to sit on top of them and address that layer and use the intelligence of LLMs for identifying data that's important to then massage of that layer.
• 23:35
  SPEAKER_01: That's what we do.
• 23:36
  SPEAKER_01: We have a pretty large pipeline of multiple modalities of AI models.
• 23:45
  SPEAKER_01: We start with LLMs and learn a lot about the patient and then make a lot of decisions along the way of what the code should be.
• 23:53
  SPEAKER_01: For instance, LLMs don't understand today, anyways, that if a patient was seen 30 days ago, such and such a payer regulation doesn't allow you to rebuild for that again if you haven't accessed and read the prior data.
• 24:12
  SPEAKER_01: So I think there is a lot of room for synergistically living with large LLMs if you pick your battles right.
• 24:20
  SPEAKER_01: The mote that we have tried to build around codometrics and hopefully create value is that objective quality measurement.
• 24:30
  SPEAKER_01: Uh we create a product called Rosetta, as in Rosetta Stone.
• 24:34
  SPEAKER_01: And Rosetta is basically the basis of is the coding correct or not?
• 24:41
  SPEAKER_01: And it's configurable if we learn things we can add or subtract from it.
• 24:45
  SPEAKER_01: And that has been an important component.
• 24:48
  SPEAKER_00: Aaron Ross Powell Those dynamics change daily for on the clinical and the insurance side, in some cases hourly.
• 24:54
  SPEAKER_00: So that's a really important translation layer.
• 24:57
  SPEAKER_00: So that's so you're you're not worried about the large pacoderms of the LLMs trampling you.
• 25:03
  SPEAKER_00: What about some of the smaller or increasingly larger uh um uh velesteraptors coming at you from the um the uh the ambient listening folks, the the abridges of the world, SUKI and others that are that are software that sits in doctors' offices.
• 25:25
  SPEAKER_00: It's interesting.
• 25:26
  SPEAKER_00: They that that that takes away the need for doctors to actually write things down and document the case.
• 25:35
  SPEAKER_00: It appears that a lot of them sold themselves to doctors and hospitals with the ability to simplify doctors' lives, but raised money from investors based on their ability to get into your business of autonomous coding.
• 25:49
  SPEAKER_00: How do you think about some of the competitive dynamics if you're not worried about the large LLMs, the the uh the the it sort of attacks on your business from the other side upstream to where doctors and and clinicians are actually taking that information in a from from voice to text to code?
• 26:11
  SPEAKER_01: Aaron Powell Well, first of all, I've met the founders and CEOs of all the ambient companies because to the one has been interested in collaborating with Codometrics so they can add the two different functionalities together as you described.
• 26:28
  SPEAKER_01: Secondly, at the moment, the ambient documentation is not a fully autonomous capability.
• 26:37
  SPEAKER_01: The doctors still have to sign off on the documents that were generated.
• 26:43
  SPEAKER_01: Third, the ambient environment isn't yet for every setting for inpatient purposes when there are multiple doctors and machines that go bing all over the place.
• 26:58
  SPEAKER_01: That is a different proposition.
• 27:01
  SPEAKER_01: Finally, we put ourselves downstream from all of that so we can objectively become the compliance engine.
• 27:12
  SPEAKER_01: Because some places are using a bridge for one thing, they're using Epic for a different thing for coding, they're using 3M for a different thing.
• 27:20
  SPEAKER_01: You have to have one common compliance environment for every health system.
• 27:26
  SPEAKER_01: So everybody who has lots of money and lots of patience to build things over time can certainly build things.
• 27:35
  SPEAKER_01: But our advantage is probably a five or six year advantage at the moment based on all the data that we have been working on and the focus we've had on the objective uh Rosetta environment.
• 27:49
  SPEAKER_01: But I'm looking actually forward to more success on their part, because I just love the idea of a doctor walking around talking in their ear with a computer to try to find out which patients they're seeing next and what's really important.
• 28:06
  SPEAKER_01: That's a good thing.
• 28:08
  SPEAKER_01: So to the extent that we can contribute to that, we absolutely will.
• 28:12
  SPEAKER_00: Aaron Ross Powell, Jr.
• 28:12
  SPEAKER_00: You've been ranked, I think, number one in class, which means your customers think you're the best.
• 28:16
  SPEAKER_00: It's not like you have even in your sector, you don't have competition.
• 28:21
  SPEAKER_00: What accounts for the success you've had with customers and how do you stay ahead of the customers in your or the competitors rather, Lacasa and others, in your specific sector?
• 28:32
  SPEAKER_01: Aaron Powell We from the very beginning stayed focused on targeting 350 health systems in the country that we believed have the best data.
• 28:43
  SPEAKER_01: Nobody has perfect data, but those that have good reasons to mind their data quality.
• 28:52
  SPEAKER_01: And we created mechanisms for mutual incentives for this transformation to happen.
• 29:01
  SPEAKER_01: That focus is what has won us both the notoriety in the market as well as the penetration and depth in mostly large the largest health systems.
• 29:15
  SPEAKER_01: Other competitors have gone wherever the dollars are, physician groups, uh, rural hospitals, and so forth.
• 29:24
  SPEAKER_01: Uh I think now everybody's coming to the conclusion that because this is a transaction-based business, you need to go where the high volumes are.
• 29:34
  SPEAKER_01: So they're all playing catch-up to our early decision on staying focused on the larger health systems.
• 29:41
  SPEAKER_00: Well, in the larger health systems, academic medical centers, some of the premier institutions of the country are relying on codomatrix to win.
• 29:48
  SPEAKER_00: So I guess final question, I mean, you're in this day-to-day.
• 29:52
  SPEAKER_00: There's a lot of fear about AI, there's a lot of hype about AI.
• 29:56
  SPEAKER_00: How are you feeling about whether all of this, all of the technology that's becoming available, hardware, software, utilization of it, is going to actually help us deliver better care at a lower price?
• 30:08
  SPEAKER_00: Because it's unclear to me who's going to win the land war between providers and payers.
• 30:21
  SPEAKER_01: And every five years there's a new wave that uh promises to bring the cost of healthcare down and so forth.
• 30:28
  SPEAKER_01: And until the incentives are aligned, uh everybody's just trying to get a piece of the growing pie.
• 30:38
  SPEAKER_01: The cost hasn't really come down.
• 30:40
  SPEAKER_01: AI has the best chance I've seen so far in the last 40 years in evening the playing field, in having the information sorted for good decision making.
• 30:56
  SPEAKER_01: Whether decision making is on the clinical side or it's on how to manage the financial aspects of healthcare, reimbursement, submission, uh, population health, at risk, managed care, all of that, everybody will do better if the data is clearly suggested, suggesting a better approach.
• 31:19
  SPEAKER_01: For instance, payers and providers are well served if they have a common certified set of AI models that they agree it correctly read the chart and it correctly submitted claims.
• 31:37
  SPEAKER_01: That will not happen in the most expensive cases of hundreds of thousands of dollars.
• 31:42
  SPEAKER_01: Everyone will have would like to have human uh intervention of uh those things being paid rapidly.
• 31:52
  SPEAKER_01: But for 80% of the cases, you can bring the data together in a way that both parties agree.
• 32:01
  SPEAKER_01: And I don't think that's a long-term thing.
• 32:03
  SPEAKER_01: I think it's a five-year thing.
• 32:05
  SPEAKER_00: Aaron Ross Powell It sounds like a dark but promising future.
• 32:08
  SPEAKER_00: But that would that was that was actually a positive.
• 32:11
  SPEAKER_00: Is there is there is there is there worse to follow?
• 32:16
  SPEAKER_01: I I I think it's uh positive, but we can't forget two elements, and that's what I was describing as uh uh the hardest thing is adoption on both sides, providers and payers.
• 32:32
  SPEAKER_01: Change management is a very difficult thing in a bureaucracy called healthcare.
• 32:37
  SPEAKER_01: Anywhere from a small healthcare office all the way to the largest health systems.
• 32:42
  SPEAKER_01: You're dealing with people who have the knowledge to bring their worst fear to fruition of losing their job.
• 32:51
  SPEAKER_01: Whereas that's not what this is about.
• 32:54
  SPEAKER_01: This is about changing allocation of resources to the most important things as opposed to the least important, critical, uh least important but highly clerical work.
• 33:08
  SPEAKER_01: Um so how health systems adopted, how payers adopted, how we help them with transformation, I can't just show up with technology and say, this works, good luck.
• 33:20
  SPEAKER_01: Uh we need to be uh incredibly involved in the adoption of it.
• 33:25
  SPEAKER_00: Well, with that, thank you.
• 33:28
  SPEAKER_00: Um if you liked what you heard or you didn't, we'd love you to subscribe on your favorite surface.
• 33:34
  SPEAKER_00: Hamid, it's been an absolute pleasure to have you walk us through this complicated, dark forest of payer and provider, billing and not payment, clinical information.
• 33:45
  SPEAKER_00: Um but you've really incredible progress, and I've talked to some of your customers, and they are really happy with the insights that you're bringing.
• 33:53
  SPEAKER_00: So thank you for doing your your great work and and good luck with hopefully helping us drive to the better side of AI versus the the darker future of just more billing and denials.
• 34:06
  SPEAKER_01: John, it's great to meet you and thank you for all of your contribution w in so many different angles.
• 34:13
  SPEAKER_01: Thanks for having me.
• 34:15
  SPEAKER_01: Thanks for joining us.