EP 26 | With Richard Kedziora | How are AI Agents Revolutionizing Healthcare, spanning from Diagnosis to Companionship?

looking at that patient's medical record saying, hey, we see X, Y, and Z, maybe you should start asking about who this should start looking at who this or other conditions, and then diagnose it faster. You can then treat the patient faster. - There is a lot of edge cases where this can turn really, really sour, especially with like, we can vulnerable people who are not adept at AI. - Or yeah. - There was a specific example, probably a year ago now where a young male was using an AI chatbot system, and they unfortunately committed suicide. - Yes. - And that's tragic. - Yes. - But that system was not designed for that. It's in a right condense. We do have to recognize that as well, and you can condemn the entire industry for this one example, in a week's, that takes a lot of effort, but you can go into AI now and say, generate me a glucose profile for patients, male, 45, that uncontrolled diabetes for one reading a minute for two weeks. - Hi, everybody, and welcome to another episode of what about AI Agents podcast? Today with us, we have Richard Kier-Tiosa, who is a seasoned professional in digital health and AI sectors. He's a co-founder and CEO of a standard solutions that focus on digital health. And his career has been there for 30 years. He's also a three athlete from my understanding. So it would be really interesting to speak to Richard and get his background, like his experience and what people should be building in AI Agents' space in healthcare. So Richard, would you introduce yourself? - Yeah, thank you for having me looking for the conversation. I am co-founder of a company called a standard solutions, as you mentioned, that's focused for over 20 years now on digital health. And that's a broad range of potential applications. We've really focused down to, and I think of as the front end of healthcare, and how do we develop applications, systems, processes with other companies, healthcare providers, to improve patient health and wellness. So as a doctor that uses an algorithm to improve recommendations, as a patient, use the application to improve their own health, or even medical device manufacturers, where they're gaining a better understanding of how their device is used in the real world. Each of those, they're aware, helping impact our fellow humans. - Right, and you've been in the space for over 30 years. So what are some of the most drastic changes I've seen in the last few years, within the production of AI Agents, because I know it's not mainstream, but people are saying that people are getting better advice from charge of the PTO, and longstanding health issues that they've ever got from their GP. So obviously there's something about that, even though it might not be yet certified. - Yeah, the biggest change, or I don't know, I'll go back a little further in the last two decades, has really been around data. Healthcare is the last industry to really embrace the idea of data, data science. They use technology quite extensively, really well. You see it in MRIs and X-ray machines, and from that perspective, but just the idea of data being able to provide information and value is they're sort of late to the game. Now, in medical records systems, only came out. In earnest, in 2008, 2009 kind of thing, they existed before then, but now everybody has them, has access to data to those patient-charts. And that's what's now enabling AI to make a difference in the world, because there is that data that's available. It was all locked up on patient-charts and paper records stored away in a filing cabinet. That's a black hole. We can't use that information really effectively. So now that we have, as an industry embraces idea, data AI is now on the breakthrough. And particularly with the large language models, and now when it's talking about these agents, what can we really do? It's part of me is like, I can't even imagine where this is going to take us in the next couple of years. I'm excited about the possibilities. There's also cautionary notes. You reference, you do, you hear stories where people are like, I was struggling with my doctor and not getting the answers. I was like, I'm forward. I asked chat, GPT about it, and got a answer in seconds. My provider couldn't do that in seven years. You're going to have stories like that in every industry with every technology. And they are, they're just, they're those one-offs. Is it very capable? Yes, absolutely. Is it going to change industry? Yes, absolutely. But there is that note of caution. You do need to talk to your general practitioner. Especially since it hallucinates. And I write a lot of code. And I use a lot of AI disease to write this code. And AI hallucinates and makes a lot of mistakes. And your health is not a place where you want AI to hallucinate and make that mistake. It's almost like faith healing when it works. It's great. [LAUGHTER] When it doesn't, yeah, you can go off the rails very quickly. Unfortunately, and it appears trustworthy. It's like when you read what it responds, it's inherently trying to please you and give you an answer to the question you're posing. There's early examples of how many stones a day should I eat? Clearly, you shouldn't be eating rocks. But the AI comes back and it's like, oh, three stones a day. It's recommended by the American Medical Association. So it sounds good. It's like, oh, the AMA said that. Oh no, just made that up. So you do have to do your due diligence and talk to your GP. And don't eat the stones. Oh, the stone. Oh, yeah, it's true. So what are some of the most alarming things that you've seen in the health care space? That for you was like, hey, well, this is changing the way we do things. I know there's some what exerjories. And you know, they're saying like five years, there's not going to be human surgeons anymore. So I'm going to be raw. But you know, it's people are trying to raise funds. So they have to say what it takes to raise those funds. But like, so what, what are some of the things that you saw that like, I don't know, obvious to everybody, but had like an extreme impact on the space and like, have changed lives, maybe have changed your life? Yeah. There is, it's about scalability for one. People don't scale. There's an obesity crisis. There's a mental health crisis. There are not enough trained professionals to provide the care for the people that need the help. Even those that are just interested in talking about it and don't necessarily have a mental health challenge, but want to talk to somebody. The level of expertise is not there. It's just not available. It's very geographic, dependence to some extent. Yes, there are virtual encounters and those capabilities. But you know, there are so few professionals trained to get to that. That's where I think it's going to make a big difference. But it's interesting as I've been out there and it conferences and talks and just discussing one of one with people. I've had this idea, which has been interesting. So one of the biggest use cases right now in AI in health care is the idea of AI not taking an ambient listening. So you're a doctor sitting there typing away at the keyboard, trying to record everything you're saying, and not necessarily paying attention to you, because they're struggling with their medical record system. The AI can listen to what you were talking about and can put that into the medical record to provide a good set of records and information that the provider can go later review. Fascinating, very low risk, which I think is important, and it's being quickly adopted across the industry. Now, take the other side of AI where it has extensive knowledge about medical research and what the care and treatment guidelines are for various diseases conditions. Put those two things together is a very powerful thing. And you're starting to see some companies out there do this. So you think of rare diseases. There are even ones that aren't rare. This is a disease called lupus. It's an autoimmune disease. It's been in my family. And today, it takes roughly seven years on average to diagnose lupus because it presents in different ways and different fashions. The symptoms can shift over time. It's hard to diagnose. Particularly if you're a general practitioner in a rural area, you might be early in your career, haven't seen it. Talk to the large language models. Have an AI agent running in the background looking at that patient's medical record saying, hey, we see X, Y, and Z maybe you should start asking about lupus should start looking at lupus or other conditions. And then diagnose it faster. So should this also then lead? So what you're implying here is basically, it's a very good tool in the hands of a good GP. be in a good practitioner because it helps them expand upon the things that they haven't seen and it's medicine is a huge field of this. It's very difficult to know everything. But would then disemplied that AI should also help us shorten the education cycle because one of the problems we have in Canada, for example, is it takes a very long time to produce good doctors and once you produce them, they go to US because they pay more. And Canada's been struggling with big shortage of doctors because nobody wants to work in Canada. They don't pay anything and you know, you can go to America and make like a million bucks a year or you can stay in Canada and make 150,000 US a year. So like, what would you do? Like a few more doctors. So like my point to you is, like, do you think it's a little short on the venture to the cycle of education? Should the medical like education change? Like, do they need to know all the things that they have? Yes. So I think medical education, education in general needs to change, needs to incorporate how to effectively use AI. Hands down, it is changing the world, not even just medication. It's changing the world, changing society. We can all argue for you for better or worse, you know, and what's going to happen. But yes, we as humans can't keep up with the pace of change in medicine. I heard a stat yesterday at a conference that a doctor was rough as the idea that clinical knowledge doubles every 72 days. 72 days for the level of information to like just fast to grow. You and I can't keep up with that. Another person at that conference was talking about menopause and providing, you know, the care and treatment for women going through medical, menopause. And she's like the level of education the typical doctor gets in med school is like a luncheon learn. They don't teach about nutrition in medical school beyond maybe a class if you're lucky. And that's changing over time, which is good. But yeah, you need to think AI first. Now, how is that advancing? It can help you study. You know, it's like my kids grew up, you know, creating flashcards on the computer kind of thing. It's like, well, AI can generate all sorts of questions to help you study to explore areas of knowledge that you're not familiar with. And there is, you know, it's a skill using AI. How you prompt it, how you talk to it, how you use it is a skill. You need to do play with it. You know, as I talk to people and like, yeah, I'm just, I'm not too impressed. I was talking to somebody yesterday about its graphical capability to create images. Like, oh, yeah, you know, I was in there and I was playing with it and the hands are all messed up and it can't do words. Like you clearly haven't gone and played with it in the last three weeks because it's gotten better. So you do need to continue to play with this technology experiment, learn and adapt. And what I find interesting is in your classroom, you can ask your teacher like, I don't quite understand that. Can you explain it in a different way? Chat GPT and large language models do the same thing, like, you know, in medicine. It's like, okay, I don't clearly understand this concept. Can you explain it in a more simple way? Yeah, sure. It can do that. So not just the doctor, but also the patient. You know, God forbid, you get diagnosed with, you know, leukemia, cancer and you go home, you're just struck. You didn't hear half of what the doctor said after he said the words leukemia. And so then you can go home and do some research around this. And so you can have that conversation with a large language model with an, you know, agent, you know, chatbot type of feature to be able to help you better understand what that is, help guide you as you go to that next follow-up appointment. What questions do I need to ask? And what I find fascinating is so many things. But as you can ask that teacher, can you help me with this? You can ask the AI to do the same thing. Help me use you better, you know, how I create a better prompt to get a better and it will help you do better, which is a powerful capability. Yeah, I mean, some people are using a for cell development. They're asking AI to help maybe better as a human, like, you know, because, like, for example, chatch of petite knows everything. If you've been promising it for like the last two years, it knows everything about you. And with the last update, it has literally access to your whole conversation history. And then basically you can just ask chatch of petite like, what do you think my IQ is? What are some of my bad personality traits? How do I improve them? Like, and it will tell you all the things that it thinks that you're doing wrong in life. And it is, you can use it for self-reflection. It's such a, I don't know if you've tried it, but it's such a powerful tool. And it'll surprise you. It's one of those things because you're like, really, can it do this? And it can be, you have to be using it to be able to generate those, you know, good responses. But it does. It surprises you from time to time. Yeah. I did a webinar a while back when I'm using AI and in product development and asked it to create me requirements for an application that helped people get over the fear of peanut butter being stuck to the roof of the mouth. Something, you're like, huh, that's unusual. I read it in a book. I'm going to think, so I threw that at it. And its response was really, you know, it was just such a human sort of response of like, why are you doing that? Like, you're getting me, you know, that's what its response was. And it's like sort of that human moment that breakthroughs. Like, wow, interesting. And I said, yes, you know, I want to do this and then generating me a whole bunch of ideas, which we're meaningful, you know, see you think about, you know, behavior change. Here you go. Yeah, I think like one of the things that will change is to your point. GPs will become great prompts and engineers and medicine. And, you know, and patients will become, they'll learn from GPs how to engineer their own problems. It's kind of scary to a certain extent, but it's like, if GPs are good enough to like prompt AI to give them the right answer and they are educated enough to know the things. So like that's AI being assistant to a GPs and very powerful, I think, agent, but it's like, like, as we just talked, to me AI is like, it's almost like a friend. It's an assistant that's always there with me and it helps. And I speak to it probably more than I speak to my wife at some time. Boy, I don't know if you started talking to it yet. I've downloaded the app and so I talked to it. It trying out like a different method of voice. The voice. The voice mode. Yeah. Which is fascinating. And then you can turn on the video capabilities so it can see what's going on the room and stuff like it is, it's incredible what it can do. It, you know, recognizes me, recognizes my wife, you know, and I found a turtle, a little baby turtle. It was about an ounce, an inch, a one inch big baby turtle on my back yard the other day. And I turned on the the slight mode, the video mode and, you know, and showed it this picture. And without me saying anything, it's all that saw this picture and it said, oh, how cute. Such a human reaction makes it such that it can be a natural companion. There was a study I saw recently that is one of the most quickly rising uses of the large language models as a companion of just that, edging into the idea of mental health. And you do have to be careful of it when you edge into the area of mental health because these are not designed for mental health care and treatment. There are one is fat pod, large language models that are specifically designed and tested, you know, the experts have made sure that they do work accurately for mental health challenges. So you're careful of that. But yeah, digital companion. Imagine psychiatrists prescribes you an LLM. You have to use this LLM for six months. You'll get better. All right. If it's happening today, whether in I knew the psychiatrist psychologist, you know, like there is hesitancy. There is some concern for that. There's, you know, lack of trust. But the technology is getting better and better. And you and I can only talk to so many people one on one. We don't scale the technologies. And does an replication and cloning of individual personalities and then applying them. And that's how you clone a good salesperson. That's how you clone anybody. And yeah, to your point, vulnerable people, like, remember, there was sometime back somebody married Siri or something like that. There was some like one of those, like they're vulnerable people out there who will. Wasn't there like some AI that advised the child to go murder his parents or something or like, like, there is a lot of edge cases where this is, this can turn really, really sour, especially with like, we can vulnerable people who are not adept at AI. Or yeah, there was a specific example. Probably you were going now where a young male was using an AI chatbot system and they come out, unfortunately, committed suicide. Yes. And that's tragic. Yes. But it's that system was not designed for that. It's in rice and incidents. We do have to recognize that as well and you can condemn the entire industry. that's one example, you look at self-driving cars, which are another huge example of AI agents. Self-driving cars are safer than the humans out there on the road. It is a little, you know, interesting. You get out of this, you get out of the time, you're at Cisco, or it's places where they have the self-driving cars and get in them. It's a little surreal. There's nobody behind the steering wheel and it goes where you need to go. But yeah, they are safer than humans. The same thing with doctors. Doctors, nurses, they get tired. They make mistakes. There's hundreds of thousands of preventable medical errors that happen in health systems every year. People get tired. It's a natural reaction. You know, in the AI's, the computer doesn't get tired. No. So I think that's really, really powerful. I think being able to clone this people across the board, being able to clone GPs, surgeons, imagine the GP can be cloned into like 10 different versions of this GP, like an alarm operator's alarm, so then GP just fact checks and trains the LLMs. So you've just scaled your GP 10 fold, right? I don't know if that's good or bad. I think you'll need a lot more data. Yes, it is definitely. I'd like to think of AI's not as artificial intelligence. I think that has a lot of loaded weight behind it. I like ambient intelligence where it's helping you be better in that perspective. It's like how do you be better using this AI? It's a partner in providing the service to individuals. It's a sounding board. It's the second opinion. It's like how has AI and individual can be better. I had to do user interviews. We recently got a grant around creating a repository of smell test data. Smell testing, all factory, the loss of your sense of smell has been a concept and test been around for decades. But with COVID, definitely when started losing their sense of smell, so now there's a lot more money that's being dedicated to this. So we're going to create this research data repository to explore and find things across different studies and uncover interesting patterns and trends. I've been in the industry for 30 years now. I've done hundreds, if not a thousand user interviews. I'd never done one of a PhD expert in smell testing. I went into CHGVT and was like, here's what I'm doing. Here's what I want you to do. Generate me 10 questions to ask. Generate me 10 more. Generate me 10 more. It is something I've been doing throughout my career. It does take some time where using a language model, I was generated 50 questions in a matter of minutes and then I can use my experience and okay, I want to take these 15 questions and I'm going to ask them in this order to guide the conversation. But I was done in an hour, versus several hours, to be able to do the same task. And it just saved me a lot of time. Oh yeah, I mean that I always try to deep research people who come to my show. I don't like to ask it to come up with questions because it's very hard. I haven't found the good prompt that will lead to questions that deep-heating questions. But having said that, so you're a chief of parading on this. So what are some of the most common challenges that helps you address in your role? So you mentioned user interviews. But what else have chat GPC helps automate or replace? Yeah, we're not at the stage of replacing anybody yet, particularly in software development. The industry isn't there yet. It is advancing quickly and we are planning with the tools and technology to better understand what they are capable of and where they're going. But huge advocate of it. From a training perspective, we are ISO 1345 certified, which speaks to the fact that we have a good well-documented software development lifecycle process. How do we develop software? We have a very formal process for doing that. And that's required when you're creating software as a medical device. And the FDA here in America is going to review and approve the software to make sure that it's safe and effective and works as designed. So there's a lot of training that goes into it. And so we started out with Word documents. And okay, training first levels like read these documents. Well, now you can drop them into and create a podcast. We're now you're listening to a discussion of these documents. And then creating quizzes around these documents to make sure that the person understands what they're reading. And so yes, it's something that we as individuals we can all do. But now it's that much faster. And you know, as you have say a 50 page document and you create 10 questions for it. Okay, well, it'd be much more effective if we had a rotating set of 50 questions I ask. That clearly will take more time, but using the AI, you can generate those 50 questions so much more quickly. You do got to double check them and to make sure that the answers are correct to avoid, you know, potential hallucinations. But easy example of how we're using it. I am using it. And there there's this concept vibe coding where it's the latest buzz word for using AI and software development. But it's the idea that you're just prompting a system to generate the code and you're not actually writing the code. And so I'm working on replacing our time tracking system. It's a little old. It's time to replace. I don't want to pay the fees for the various systems that are out there. It's like, let's use this as a good, well-defined example of what I need. So we're using it internally to be able to learn and experiment. How can we apply that to our projects? That vibe coding perspective, you need to create a table maintenance screen and application. You have the admin side of a system. It's in the cloud kind of thing. It's easy enough to create a user admin screen capability. Let the AI code that for you. Vibing, vibed bugging, all the great fun that comes in it. All right, testing data creation. We do a lot of medical device testing and we were doing this in diabetes. And this is data. And so some of the newer devices out there in market will generate a glucose-readambered minute. And if you want to create data for that over a series of weeks, that takes a lot of effort. But you can go into AI now and say, generate me a glucose profile for patient, male 45, that uncontrolled diabetes for one reading a minute for two weeks. And it'll generate you a series of data. And then you can say, "Graph it." Because it's really hard to see what the numbers are. One is it look like. So graph it. Okay, now you have this profile. It's not very interesting. Now I add in the fact that this person eats three meals a day and adjusts the glucose. It understands, crazy enough, that like you eat your glucose is generally going to go up. Now incorporate the idea of exercise, which generally will flatten that curve as it's going up. Incorporate the idea that this patient takes medication every morning. It's going to adjust those profiles. It's not the be all end all of test data, but it's giving you a really good, quick data set to run some meaningful tests with and see what happens. And it's just, you can do that in minutes hours versus days. Yeah, it speeds things up for us as well. We've been able to, you know, we used to have a lot of software engineers. A lot of them have, you know, just become redundant because of, well, AI and how fast things are going. Problem is once you've built out big enough of technology stock, you need engineers to sit around and maintain it anyways. And, you know, and if you're not an engineer, you're going to let AI go wild. It's going to write a lot of garbage on maintainable code. That is going to start to stumble over its own code. Anyways, it will be fixed eventually and we're going to just, you know, be able to build these SaaS products in seconds. Like that's kind of crazy if you think about it. We do a lot of work in healthcare, obviously. And so there's this level of trust and certainty of abocation above, say, like a book, a hotel website or, you know, you do your mistakes being made on like buying a car and someone was able to trick the car by AI into selling a car for zero dollars. Okay. That's not great for that company, but no harm. You don't get a no foul in healthcare. There are obviously a greater degree of risk involved that you do have to worry about. So can we use it? Yes. You know, but you do have to make sure that you test it out well and effectively to make sure it does what it needs to do and doesn't go off the rails. Exactly. So where do you see this go in the next five years? Because, you know, we've really spoken about it. about like this data and vibe coding. So how will this transform healthcare in the US in the next five years? - I think it's gonna let medical professionals, doctors, nurses get back to the reason that got into healthcare, providing that one-on-one care and service to people. It's gonna let them interact more with people, which is why they got into medicine. So they're not gonna have to take the notes. The AI can do that. Now what is, what are the top 10 questions I need to query this patient with? The AI can help with that. Billing data. How do I generate that to make sure that the billing goes through effectively? Well, a lot of those administrative aspects, you're gonna see happen very quickly if they're already happening now. Secondarily, then it becomes improving the diagnosis of conditions and then after that, it really then treatment comes into it. And very much so that it can help out with treatment, but you want that medical provider, that human to cross check everything that's happening in BBO's decision maker. But all of that, this is gonna get easier. So I think the general health is going to increase, which is an amazing thing. - So and if you were like in the young interpreter, looking to build something in the health space, you can have much of a budget, but you kind of wanted to, so what are some of the use cases that you can think for AI agents that are like ripe for introduction into the space right now? That's some young startups could like pick up on and start building, because it's a very competitive space, but it's very well regulated. So you have a mode of like certifications and what that you can acquire. - Yeah, and you have to be make sure you're have a complying above it. And all in healthcare before technology, it's like you really do need to understand who's using your system, who's buying the system, who's benefiting from it, because you have this mix of, patients, providers and insurance companies, which makes it a very challenging space from a startup perspective. But if you're getting into it, one, yes, AI is sort of a must have. The ambient listening is a huge technology. It's being picked up. I wouldn't even go near that. It's too crowded of a space. I was at some health tech conferences last year. There was multiple ambient listening vendors next to each other at this conference. And this is the worst feeling in the world. - You're not saying it's a huge thing. - I had the very point, like why are you different than the person right next to you and they couldn't really answer? - Yeah, and the conference organizers, like Kudos to them, I remember we were at some documentation conference like a some, at the website, and they put like six documentation companies all next to each other and you're sending there and you're like, "Well, I'm fucked." - Yeah, it's crazy. And the conference is making a ton of money off of these vendors being there. And it's probably not worth their time as a startup to be at a conference like that. You're not gonna be a lot of attention. - Where I think everybody needs to start thinking is how do you marry together in these different aspects of healthcare to use the fact that the AI knows so much about the background, feel the medicine and research and what's happening. And marrying those things together. So not just ambient listening, but understanding that here's the medical research, what interesting are you seeing in the data to elevate everybody's experience and incorporate it into the workflow. That's a huge aspect of working in medicine and healthcare. If it's not part of your workflow, you're never gonna get any traction. They don't have a lot of time. - So sort of like vertical, verticalization of ambient listening. So ambient listening for delivery rooms in the infant department or whatever. And it's trained specifically to listen and alert on specific things that are being said or specific events or well, so anything beyond ambient listening that caught your eye and that you think somebody could build that could be beneficial for the space or at least to kind of get into the healthcare space. I know it's maybe it's something you don't wanna share. (laughing) - Yeah. Now, I think there's an intriguing prospects to now on understanding, I'll call them silent signals. - Right. - So I've seen one company that has done this where it monitors your typing on a keyboard over time. And so over time is your ability to type accurately on that keyboard or the pace, the rate that you type changes over time? - It might be an indicator of a mental condition of Parkinson's disease. - It might be. - It might be. But it's that type of thing that nobody really thinks about. It's like, what else can we learn from the retina of your back of your eye or a technology that listening to you and I talking here and be like, hey, I've been listening to an RJ for about three months now and I've noticed this speech pattern is changing. - Right. - What does that mean? Where does it detect a note of depression overall in my conversations? Powerful technologies that are in the background that could provide meaningful information. You do get into privacy issues of who's listening to this and who's using this information. I don't want my employer having awareness of that, hey, I was really tired the last five days and my productivity wasn't that great. What is that being for? So there are privacy issues that you have to think through and how to use this technology. But yeah, it's things that how do you apply this technology such that it's in the background, learning, understanding, that's probably more machine learning type of technology. But then that interface and use of that data and communication and that information is where you get into the large language models and it's ability to look at that information and share it. Yeah, that's true. We actually, we were a proposal for our CMP in Canada about watching inmates in prison cells that for suicide, for indications of suicide. And I had a good demo and they said, well, we don't think your technology is viable but I had a demo of it where we could monitor them and we could detect signals of depression using the lamps because there's a lot of nonverbal signals and verbal signals that you can say that, this image looks suicidal even more than he can for five or six days. And this is what his normal state is and this is his bad state. So just like, you know, you can, especially now with LLM vision and when you can like, do your point, like you can show it to turtle, you can show it in inmates. These are person suicide, can we help them? Like, you know, that's, because in Canada it's a big problem because you know, they don't treat their inmates well and they kill themselves and like some, they don't want that to happen. Yeah. But yeah, but to your point, mental health and monitoring, silent monitoring is a big space and you can, and this is something that would be interesting to schools. For example, like schools would love it. Like if you could monitor a class and see if some students exhibiting some symptoms of abuse or depression or whatever, to your point, privacy is a question, but if parents consent, I mean, yeah. Yeah. Well, it's, it's fascinating because we, the marketing, I think of the marketing industry, they have such a deep understanding of our patterns of behavior that we know if you walk into a supermarket, you're more likely to buy a product at eye level as opposed to something on a higher or lower shelf. Powerful moment. It's maybe not the greatest use of technology in terms of trying to sell other things, but it's that powerful behavioral understanding that then can be used in healthcare for greater good. Very true. Well, RG, is there anything you wanna tell to our listeners, maybe promote something? We have people from various industries that listen to us. Maybe they should come, maybe reach out to you. I think you have very experienced in the healthcare space. So maybe I'll let you have a word here to promote whatever it is that you have. Thank you for the opportunity. And it's been a great conversation. They're really enjoyed it. We are like, you know, many kind of always looking for that next project. You know, how do we make a difference in the health and wellness of our fellow humans? That's those are the types of projects we enjoy. And we've been fortunate that customers have put our names on patents. Our software developers are named on, as authors in clinical, peer-reviewed clinical research articles. Now, because we are very much part of that cutting edge conversation of how do we drive the next generation of technology. So just love to explore that, whether you're a big company or a startup, love having those conversations. On a personal perspective, I'm writing my first book, all about productivity and how to think about it in terms of energy, not time. Only 24 hours in a day. it's how to use your energy to be effective, to be productive without burning out. It's not going to be coming out later this year. You can follow me at productive harmony on Instagram or LinkedIn. I'd love to read that. I mean, I 100% agree because you sometimes you run out of energy for a specific activity and you overextend yourself and you put too much energy and then you don't have any money to do the next year. How do I re-invigorate myself? How do I get back into that activity? Energy is like, it's not just about the time. You may have the time but you don't have the energy to do it. How are you going to do it? Absolutely. Thank you so much for joining the show RJ. I really enjoyed having you here and hopefully we get to reconnect later time to talk about the changes in the health tech world. Be here. Sounds good. Have a great day. Thank you.