Episode Transcript

CJ: Hello, everybody! Welcome to another episode of compliance conversations. I am CJ Wolf, with Healthicity. And I'm excited today for this topic. This topic has kind of been discussed in all sorts of facets of our lives. We're going to talk a little bit about AI, and we have a wonderful guest and expert, Dr. Adam Robison, who is here with us. Hello! How are you? 

Adam: Hello! I am doing well, thank you! 

CJ: Thank you for joining us. Adam we always like to have our guests kind of introduce themselves a little bit. Tell us a little bit about you know your professional career path, what you're doing now, and how it kind of relates to today's topic. 

Adam: Yeah, my name is Adam Robison. I'm an internal medicine physician by training. I did my training at the University of Louisville. I'm a big Cardinals fan. I have been in a practicing position as a hospitalist clinically for about seven and a half years now. So, I do work still clinically today. I founded a company called Aimedica about 5 years ago. Actually, come from our fifth anniversary in 2024 to solve some of the issues, we run data interoperability with EHR as well as applying artificial intelligence to physician and clinician workflows using these data interoperability standards that have become widespread now in the EHR and in the modern healthcare landscape. 

CJ: And when you and I first kind of met each other and talked a little bit about what you do, I thought, wow, this is like perfect because you are practicing, you know, like the pain points, you know what you're trying to achieve clinically. And then you also have expertise and you have a team that has expertise on this technical side and the AI side, and that to our advantage. So, I'm excited to talk about this a little bit. Anything else you want to share before we jump into our topic? 

Adam: Yeah, no, I think that's something I talk to people about. I like to be in the Hair Club for managing those old commercials. I'm a member at them also, you know, a user of the product, so to speak in fact. So, I am a practicing physician. I still work quite a bit, and I see patients on the regular in the hospital. And so, I use the product. I use our products in my own clinical practice all the time, which I think is being great to help refine the product. But also, for me to really, truly understand the problem we're trying to solve. 

 CJ: Yeah, absolutely. And when they let's maybe jump into some of those problems. So, what are some of the problems, the major problems that you're seeing currently with electronic medical records, or we may say EMRs and maybe the lack of ability to use all that information like there was such promise, right? When we moved from paper records to electronic medical records of; "Oh, you'll have all this data at your fingertips and this and that," and some people, some docs I talked to were like, "Oh, it's overwhelming! We don't use all this information. It starts to bloat everything." And so, what are you seeing as major problems with EMRs and the ability to use that information? 

Adam: What we're seeing right now, and I agree with you and when electronic health workers first came on the scene. There's a lot of promise around, "Hey, we’re going to have all data, and they'll do all these amazing things with it." I trained it kind of that switch over, right? So, where I trained it happen right beginning my training, I was using paper charts. We had a computer-based ordering system but it was still mostly documentation was still paper charts. And then, towards end of my training, we converted over to electronic records. So, I've kind of seen both these worlds evolve over time. 

And when the electronic health records first came out, there's this all this promise we're in all this data that we could do all these amazing things with, but really what we just basically saw, and even to some degree right now, the electronic health records, which is basically a digitized chart. 

CJ: Exactly! 

Adam: We didn't really involve the functionality that, I think a lot of us had our minds, we would see and I think that we're now that next evolution of what the EHR will do or what these, we hope had a promise, you know, five, ten years ago, when these electronic records were still coming on market and becoming widespread. We're starting to see some of these potentials, but we're still far, we still long ways to go. But I'm really excited where we're going to be in the next 10 years. I think there's going to be a lot of changes over time.  

Some of the biggest problems I think we're currently seeing, though, is still around the idea that a patience chart is very siloed. I recently have a really good friend of mine. He and his wife had a child that had a lot of medical issues and they had to go and received initial care at a community hospital. But do the nature of the medical problems the child had, they had it go to a tertiary center to have high level of care.  

Even though both sides were using the one of the premier EHRs on the market, they had two different records that really didn't talk to each other, which was a frustrating experience for the patient who basically, even despite, you know, having a system, that are the same type of EHR, they did talk to us, they had two different types of personal records, and we still haven't overcome that big problem within the EHR space. I think we're starting to get there and talk with that in here a little bit. I like to talk a little bit, that's what my company is trying to do.  

But that is still a huge problem. Until we until we can get around that data interoperability issue, I'll be applying artificial intelligence is something will be somewhat limited because of data some sales will be somewhat limited. 

CJ: Yeah, it's like, you know, as a user of health care, right? As a as a patient myself or my children. It's like I have to fill out the same information like 20 times I'm like, isn't it already in the system! "Oh, our systems don't talk to each other!" It's like, Well, what's the whole point? It's like it have to make things easier, right?  

Adam: Yeah! No, I laugh, because people say; "Look to my chart." And I'm like, "I don't think you understand. There is no chart. There's many charts, you may guess, and maybe in this computer system, but not in that computer system." And I think that's really glossed over. When people talk artificial intelligence in the health space, artificial intelligence by stage and very nature requires large data sets for it to be providing accurate information, accurate models. And right now, we have lots of data, but it's very siloed in different chunks that it's not readily available for artificial intelligence model to use in kind of a broad way.  

And so there that I think there's a lot of misunderstanding on bad data interoperability standards and correcting that and really making that fixed is a huge barrier to really making really meaningful automation and artificial intelligence in the healthcare space. 

CJ: Yeah, no, that makes a lot of sense. I can see those. I have a question that's kind of going to take us a little off the topic of problems, but are there any other problems that you wanted to discuss before I kind of switch gears a little bit? 

Adam:  I think the other issue is the electronic records themselves are very complex and cumbersome to use and I think that's another issue that we've got to figure out. They're like, I think they're often try to be everything to everybody, and a lot of that can be done by fixing these processes in the background. We can really make them a lot better. And really improve the user experience. I hope one day we'd like to evolve the electronic record to something that's not a barrier to me and my taking care of my patients which sometimes feels like it is, works as something as critical to my ability to care for patients like, you know, I'm an internal medicine doctor, so I do use a stethoscope, that's a critical part of my tool belt with care for patients. I feel like, eventually we hope to evolve electronic medical record to the point where it becomes something as critical to providing care much like a stethoscope is. 

CJ: Yeah! Makes a lot of sense. I do some teaching in the patient safety and quality care space at the medical school where I graduated from. And I was reading this research study, and I got; "Oh, my gosh, it only in today's day and age I would have never thought of this 30 years ago, when I was in med school!"  The study was how to relieve click fatigue in the EMR and it's a real thing like docs and other providers have to click through a bunch of different things. And it's like; "Is this really necessary? Is this system helping me? Is it adding burden? and that sort of thing," I thought, "My goodness, what an interesting study!" 

Adam: Also, we are responsible, we as physicians, are responsible for every date in that, every data point in that EHR, and that let record responsible for that. And it's, you know, it's something that we have to be. But if we don't look, or if we don't know what's in there, or if we don't want to click around too much, we're busy. It is a liability issue there as well. 

CJ: Right! Yeah, that's such a great point. So let me ask you a little bit about some recent legislation. You know, you and I have communicated about the 21st Century Cures Act, and I'm curious what some of the impacts of that legislation might be and how it's changing the way that that we think about healthcare data.  

Adam:  Yeah! Thank you. So, 21st Century Cures Act, I can't remember when it first came out. It was, I don't know about a year ago. I think or two years ago, but I remember when it came out because I do practice the clinician, a lot of us physicians were somewhat nervous because it made the chart readily open, right? Any patient could view all their notes in real time, which really don't have a problem with that. I don't say anything inflammatory. I don't think most doctors say any inflammatory than notes, but I think a lot of us were more worried about asking a whole bunch of questions by patients that didn't really understand maybe some of the nuances of, or some of the thought process that go on our notes as we're talking about what we're thinking about the different things we're considering for a given patient's care.  

And so that was a lot of concern about that, however, buried in there, if you will, and that was that got all the press that got all the attention was; "Oh now our notes are widely available," being thoughtful you document. Part of that, though, was just the one of the biggest intents of the cures act that was actually to make the data, make the chart more transparent. So, part of that transparency is actually making the data itself more interoperable.  

And so, part of that 21st Century Cures Act was to mandate that all certified EHRs were basically any EHR that's on the market essentially abide by the certain data interoperability standards called SMART on FHIR. And you know, with HL 7 this was kind of that coming out of that group and being mandated that all data has to be able to be, has to follow these standards. And now and that means every data point in the EHR has to be, what has to be available through these interoperability standards.  

Electronic Medical Records have thought this data siloed for quite some time. They've liked it, it's given the competitive advantage, but that's now been removed. And so, by the end of this year, everybody has to be compliant with these standards. And if not that, will be penalties.  

Now, who knows if that, how widely that will be and it just takes a while for that to happen the market, but that is we're going and I think that's a critical thing that has to happen for us to really see some of the promise of the electronic record when it came out to the marketplace several years ago.  

CJ: So, are those compliance requirements that's on like the physician practice or is that on the vendor who creates the EHR?  

Adam:  It's on the vendor that creates the EHR. They have to be able to compliant with that. This is not, this type of thing, isn't something that you as a clinician who buys the EHR has any control over. It's something that the software vendor itself has to has to have turned on and have available.  

CJ: Interesting! As you were talking, it made me think about a recent encounter I had as a patient with my doctor and the medical assistant before the doctor came in and said; "Just so you're aware, when the doctor comes in, he'll be talking. It might sound like he's talking to you, but he's actually talking into his phone as well because he's using this new software to help him, you know, input data more quickly and this and that," I thought, "OK, that's interesting!" And then my encounter went on and yeah, he like stopped and was like talking to his phone and it was a little weird, but I could see how he was trying to increase his efficiencies with that technology. But that was kind of interesting.  

Adam: Yeah! I think with the with the electronic record, one of the side effects of that was actually the amount of documentation mandated by CMS to meet all your metrics for quality metrics. So you see a lot of these things happen and these things are good, but they're just they add a lot more burden like a lot of in the primary care space, a lot of about making sure your patients have adequate depression screenings, OSA screenings, you know, on and on that require communication and conversation with the patient. So yeah, those are can be quite onerous at times, for sure. And that's why I think the promise of artificial intelligence with access data is going to, hopefully automate a lot of those things. 

CJ: Fascinating! Well, let's take a quick break, everybody, and we'll be back shortly.  

Welcome back, everybody! I have been talking to Doctor Robison here about AI and we're going to get into a little bit more specifically about some solutions that he has been involved in. Before the break, we were just talking about the 21st Century Cures Act. Anything else on that topic that you might want to cover? Before we kind of jump into some of these other topics.  

Adam: I think the last thing I'll say, the 21st Century Cures Act is I think we are still ingesting what this is actually going to mean for the market and for patients. Broadly, I think there's going to be a lot of changes on how people ingest their personal healthcare data, because this will have huge impacts on the patient to have a kind of a portable chart that is their own, that will now be feasible with this type of technology, which is going to have a huge, I think, change in how people interact with doctors and with healthcare systems.  

CJ: Yeah, because wouldn't that be great? It's like I'm on vacation, you know, I live and work out of Utah, but maybe I'm vacationing in Southern California and all of my records, right, like previous EKGs, previous imaging, all this sort of stuff. And what if I have chest pain when I'm in Southern California? Wouldn't that Doc love to be able to see, you know, my prior EKGs, recent EKGs or any other type of tests, it'd be cool like if you could carry that on your keychain or something, right? Is that kind of what you're suggesting? I'm not saying that specific example, but that portability.  

Adam: Yeah, that's exactly what I'm, yeah, we will be able to eventually, because right now I talked about with other people at a time that doctors don't own data, patients don't own data, healthcare systems don't own data, it really is owned by the EHR vendor because of how it's set behind their infrastructure. It's not readily available. If you switch EHR vendors, there's a huge costly and time intensive process to switch over from EHR vendor X to EHR vendor Y. And then the data structure is so different it takes forever to kind of maintain that, and it really doesn't transfer over neatly into that new EHR just kind of left in some sort of archived format and so in the future this will eventually go away where the EHR data will basically be like an excel document which can be readily available or an Excel spreadsheet or word doc that can be readily accessed through any sort of software that you like to choose from. 

CJ: So, you talked about this data interoperability standards. Why are those so critical for what you're working on for artificial intelligence, you know to be used successfully and applied broadly in healthcare?  

Adam: I think a lot of people now with the advent of ChatGPT are somewhat aware of artificial intelligence is and it's a good thing, but also bad thing, because I think a lot of people's understanding of what it is a little bit different than or a little narrow, which should say from what it actually is. And there's a whole bunch of different uses we've been, I like to say we've been using artificial intelligence for 25/30 years now and just ChatGPT is a large language model is one of the, you know, the most newest recent fat of this, but before that, the previous iteration that everybody was aware of was IBM Watson and IBM Watson when it came on the scene there was a lot of hype around how that was going to revolutionize healthcare and obviously it hasn't. And it's now just kind of still being out there people trying to figure out to do with IBM is.  

For artificial intelligence to be useful in the healthcare space or any space that's deployed to, it's really reliant on accurate complete data and without that you're going to get useless information. That's why IBM Watson failed was really because it didn't have really complete to operate from, they made false assertions and false assumptions.  

CJ: That's so interesting. As a hobbyist, I'm into photography and I use some Adobe products like Photoshop and others and they are launching and I've been using just recently some of their AI features to do what they call Generative Fill, and I'm not a technical person you might understand it better, but essentially this technology looks at the entire picture and you can ask it to generate maybe clouds or generate, so if I took a picture and it was just a blue sky and I want clouds, you know, so it's just fascinating and I think it does what you just said, it requires like mass amounts of data to come up with some good things. It comes up with some funky things and then you just you can kind of tell it; "That's pretty close or way that's way off," and then you can have it redo and it just kind of learns as it goes. And I just, I find that super fascinating and I know you work on this more in the healthcare space, not photography. And I just wonder that is going to be some power tool if...  

Adam: It is, but actually you actually highlighted one of my biggest heartburn with conflating large language models and generative AI with like image generative AI and because in the situation you described it's okey if it gets clouds wrong, right? Like we really don't care what type of clouds it shows you just care if you get clouds, right? With the healthcare space we absolutely care what type of cloud it showed us, right? We absolutely, it's the right cloud, not the wrong cloud.  

And that, I think is the pitfall of right now broadly applying artificial intelligence without really thinking how it is generating the recommendations it's giving to you as a clinician, difference between a white box versus the black box algorithm. We have to be able to understand why a suggestion was made or why a process is automated and what's doing if we can't, we can't trust it. And then that's the biggest problem I have with just broadly applying artificial intelligence to whatever process because it is fraught with issues.  

Now where I think it's already being making splashes I think, when I think it's appropriately is within the imaging space.  

So, radiology, pathology, there's been a lot of study there showing that it's good or if not better than people reading those images, there are still some stuff that requires a pathologist and radiologists look over and say, "Yeah, that looks right to me," because they have that wealth of knowledge to kind of pull from. But I think why those situations are really well suited to artificial intelligence as it currently stands because when you have a, let's say an x-ray in front of you, you have all of the data, right? All the data that are in that picture, all the pixels are there because it's in the complete X-ray. For radiologists, their job, yeah, little history is helpful, but they can still look at the X-ray and go; "Yeah, we think it's XYZ," right?  

They can see that and I'm seeing that the pathologists, the history is a little bit helpful there for sure. Like the age and some of the things that does help flavor in or color in some of that information so that they can make that accurate diagnosis. But you still have 95-99.9 percent information in front of you, whereas broadly applying artificial intelligence to a patient chart as we established earlier in our conversation, it's only, it could be one of 100 different charts and so that's the problem.  

CJ: That makes complete sense because again, I'm a little naive to the technology, I'll kind of go back to my photography example. It seems like and I'm guessing here that, you know, Adobe uses when they do this generative fill, they're not just using the data from my pictures, it's from everybody who is asking for clouds and so it can kind of come up with some pretty good things and you can click regenerate, but when it comes to like faces and stuff, oh does a terrible job because and I don't know, it may be limited in what it's pulling from, right? Like it needs data to come up with something I think is what? I'm hearing you say and if and if in the patient's chart like we talked about, if it's only a small portion of their history and medical record, you're not going to get as good information from AI is if AI had access to everything, is that right?  

Adam: Exactly! So, imagine in the future state where you have because of the data interoperability standards, you have a complete record of somebody from the time they were born to the time they're 85. That's a complete data record. That's a complete record, right? And it's all mapped appropriately. You could theoretically start making predictions on how likely they are to have heart disease. We can start doing really meaningful screening for patients, right? That's really targeted to that patient as opposed to everybody greater than you know, 45 I think is the current age for colonoscopy to have a colonoscopy, or 10 years from their first early diagnosis first relative. 

We can start making a lot more targeted screening recommendations a lot more targeted prognostications on disease outcomes for a given patient. Right now, those are not very specific we know they applied broadly speaking, but we really can't get really, really finely tuned on that patient from our just, with regards to a prognosis and treatment algorithm.  

Once we have a more complete data set, that's not just, not over a whole bunch of different EHRs throughout the patient's life, but it's one complete record which we will get there. We will be able to start doing some really amazing things with artificial intelligence that we can feel confident in because it is a complete data set. It's not a very specific, not a very small part of that person's record.  

CJ: That's so interesting. So, you kind of mentioned these pitfalls and the promises, any other pitfalls or promises of artificial intelligence that you want to discuss before I kind of move to a question about regulations a little bit.  

Adam: Yeah, I think the other big pitfall I would say, I kind of touched on it earlier, is just not really understand the problem you're trying to solve with artificial intelligence. Just applying and going well, it's going to make this better. I think that's like I said, it's really critical to understand how the decisions being made, why it's being suggested to the user. So, we as clinicians and whoever's using the software in the healthcare space can have confidence those suggestions are correct if we don't know why it's making those suggestions, we're going to be critical, but it's going to fail, I think. I think that's really critical to really understand what is going into that thought process on why suggestions being made.  

CJ: That's helpful. So, Adam, you know, a lot of our listeners are kind of in the compliance space, right, healthcare compliance and the name of the podcast compliance conversations and coding and billing and all that kind of stuff. What regulations are currently, you know, out there as it relates to AI implementation with the modern you know, electronic health record landscape. Are you aware of any?   

Adam: Yeah! So, this is actually really interesting. So, there aren't a lot. This is the Wild West I think right now. So, in the past healthcare software, as long as it has been somewhat limited from FDA oversight and regulation. It wasn't considered a medical device per say. So, the FDA has really stringent requirements for any sort of, you know, medical device; knee replacements, your hip replacements, you know, the pacemakers, you name it, you know CATS, stents, those type of things are very well and vetted and regulated as they should be.  

What we have really seen is that totally touch the software landscape, this clinical decision support because previous clinical decision support was just taking what was widely available in the literature and just making it available to the physician, you know that was kind of where we're at right now.  

And my company does that to some degree. We take what's available in the literature and provide it to the clinician in a smart way with the patient's chart and go, "OK, this is the information that applies to that patient's chart based of what's in the in literature, right?" That's reasonable to do. So, we automate that, we review the patient's chart and we find what's in the literature and we provide that back to the clinician.  

Where I think this is where, and I think that's fine for right now, I don't think that's a big deal where I think we will see this in the future as artificial intelligence becomes a lot more savvy as we start looking at this interoperability issue. So, we have more complete data record of the chart we can do some of these really interesting things that I think we'll see in the next 5 to 10 years of predicting in a more robust way than we're doing right now. Where it's really predictions tailored to that specific patient as opposed to recommendations are kind of this is what you know was what's in the literature.  

We're going to see FDA regulations around that. I'm almost positive we will because based on some stuff they've been put out recently, they will be, because it has to be, right? All of a sudden, these black box algorithms are suggesting you on three patients and that are very care for that patient and that's something we will see in the future, but it hasn't been yet because I still think we're in the very early stages of this currently.  

CJ: That's a good point. For some of our listeners, you may remember a while back we had a guest on who talked about software as a medical device. I mean, there's a whole industry on software as a medical device. Now I know that it's probably different than AI a little bit, but I think it's starting to touch that. So, if this topic interests you, you might want to look back at that episode. But you know when you were talking, it made me think of, so I used to work at MD Anderson Cancer Centre and you know, years ago and even today, the big thing is personalized medicine, right? So, you don't treat breast cancer the same for everybody anymore. You're looking at markers, cellular markers and those sorts of things to treat specific, your personalized cancer. And that's what I'm hearing you say a little bit. Is that with access to the medical record you're getting, and maybe combine it with what's in the literature, you could get some personalized medicine in a way.  

Adam: Absolutely! And I mean think about what with this idea of data being this more longitudinal view of a patient as opposed to just this, you know, they've seen this physician for the last five years they have that five-year chart, but really their whole lifespan and then even more importantly connecting to their family history, right? So, to your father's record, you threw if they could do that in this new world we're talking about. Where you can base off of their own history, you know exactly when they were diagnosed with cancer and know exactly when they're diagnosed with a heart attack or whatever. And you can make screening recommendations based off of that, right?  

And so those are those, it becomes a very interesting new world we enter into when that chart becomes more open and widely available, which is, and then that is where I think that regulation is going to be really will have to be done because this is going to be entering a whole new world of personalized medicine then is it really feasible currently.  

CJ: Yeah! So, tell me a little bit more about what you do at Aimedica, which is and you mentioned a little bit that, you may pull in the medical literature a little bit apply it to patient's chart, but you know, I'm sure there's more to it and you know, in our last couple of minutes here, can you kind of give us a high level overview of what you're doing?  

Adam: Yeah! So, several years ago when I founded this company, I was sitting in front of a patient's chart and waiting, "This is ridiculous! I have to go to all these third-party sources outside of the EHR to make it to treatment decision my patient. I needed this to be integrated into my chart." But I not only it should be integrated dressing the chart which, there are links that they're doing, but I want the software to look at my patient's chart and say what stuff should I be looking at? Present me the information that is reasonable for this patient to be considered given they have Afib, cirrhosis or whatever.  

And so that's what we did. We came up with it. We have a platform that looks at the patient's chart, our AI medical platform looks at the patient’s chart, looks at all the data. Again, it doesn't look at longitudinal, it's just that snapshot current patient's chart, we want to get there eventually, which I think with the Cures Act, we will be able to get there. Where we look at that patient's chart and say "OK, what information, what do you need to consider as a physician from a coding and a billing standpoint and from a documentation standpoint?" And so, we're able to provide that information, but more so now we're able to also harvest that metadata and give clinicians and hospitals insight to how evidence-based practices and coding is being applied to their patient population because it has direct access to patients' chart.  

We are getting into the health clinical research space because this problem of accessing data and providing contextual information to the users as they're reviewing charts extracting data is a big time. It's a huge, huge expense from a time standpoint, from personnel. And so, we can really with our tools, we can shrink that time and personnel burden reviewing charts and abstracting data. We can automate a huge chunk of that now, so it makes it really easy. So that's kind of where we're doing. It's, there's a lot they're going on. I can spend a whole podcast talking about it.  

CJ: Yeah, that is so exciting like. So, you talked about maybe it's point in time looking at the patient point in time and the longitudinal is not quite there yet. And when you say that is it like, you know, maybe I have a patient with chronic kidney disease and you know, following them for three or four years and I'm, you know, you've got three- or four-years worth of glomerular filtration rate test, albuminuria all that kind of stuff. Is that what you mean is like, maybe taking three years worth of data when you say longitudinal?  

Adam: No, what we can say is what we can say is we can look at whole patient's chart at a given EHR. So, like if you know received all their care one site for 10-20 years, we can get all that data. What I'm saying is as we expand our network and learn more and more sites, we'll be able to link those charts together in a meaningful way. We'll be able to weave all the charts together independent of what EHR vendor people are using. That just requires us to build out our network, but as we do that more and more sites, those records become more and more meaningful and useful. We can kind of weave those disparate charts into one greater hole for that patient and to give a broader view of that patients' chart, it's just having access to data, which is obviously the problem right now. We're trying to solve that. But once we have that, the more data we have, then the accurate those models become and the more we can more meaningful insights, we can provide the users and the patients and clinicians. 

CJ: Wow! Doctor Robison, this is fascinating. We are kind of running out of time, but I want to see if you have any last-minute thoughts, maybe even plug Aimedica a little bit. If you have a website, of course we can put all this in the show notes as well, but any last-minute thoughts and or anything like that you want to share.  

Adam: I think the last-minute thought, is I think we're at the very beginning of a huge change in the healthcare EHR landscape, but I don't think it's going to be just applying ChatGPT, I think that's a really narrow view of what we'll be seeing in the in the next several years. I think we're going to, it's really going to be around data interoperability and then applying, once we have that fixed, which we are in the process of fixing that through that 21st Century Cures Act, we'll be able to start doing a lot more interesting things regarding to how that data, once that data becomes unlocked. Then really the sky is the limit, we can do with it, but I think we're going to see that in the next several years. So, I think there's been a lot of talk about AI like, "Oh, ChatGPT! Large language model going to change things." I think that's a really narrow, myopic view of what artificial intelligence will be doing in the future and liking it more to you know, to tricorder in Star Trek, where we'll be able to do a lot of really cool things with it that we can't do now which we can't even probably fathom we can do with that. So, I think that's where we can go in the future. I'm very excited to be part of that and I hope we'll be around for a long time to come. And so, yeah, our websites, Aimedica.ai and we are, you know, we're excited to be a part of that future that we're hoping to create.  

CJ: Great! And if people want to reach out to you, you could probably share emails, or there's probably ways on their website to get more information we can have that in the show notes as well.  

Adam: And we'll put that in short. Thank you! 

CJ: Awesome! Well, thank you so much. You know, I bet this is a very fast-moving field. Of course, it is. And maybe, you know, we have you come back in six months or something and tell us you know what's already changed from what the last six months. I think either you or your colleagues, I know you have some technical experts on your team as well. It'd be fascinating to talk to some of them maybe.  

Adam: Yeah, it's to your point, it's even in the last few years I've done this, it's been amazing, all the different avenues we will apply our technology and our platform to. So, it's been a very fascinating role to be in for sure.  

CJ: Well, this has been great. Thank you so much for your time and expertise. And thank you to all of our listeners for listening. If you like these episodes, please share them with friends. Please subscribe we're kind of building the audience even more and more every time. And if you have ideas and suggestions for topics or speakers, please reach out to us. We'd love to hear those. So, until our next episode, please stay safe. And have a good day everyone!