Latest / Behind the Founders / The Hospital Failed His Brother. So He Built AI to Save Lives.

Transcript

The Hospital Failed His Brother. So He Built AI to Save Lives. · Behind the Founders

• speaker-0: Hi everyone. My name is Carmen Durvad and I will be hosting today's episode of Behind the Founders. And today I will be sitting down and interviewing Sal. Sal, welcome. Can you introduce yourself? speaker-1: Thank you. Thank you very much for this awesome podcast, Carmen. So about me, I would say almost 20 years in the field of marketing and AI. currently I lead a startup called ILABPP Health. It's a healthcare AI startup which aims to solve the kind of problems we face at the hospitals, healthcare units, and also at the clinics. And I've been leading the startup from last two years, 2024, we started this basic groundwork of the startup and all. And the actual work started in 2025, and almost two years, and I'm kind of running this as a CEO and founder, along with my co-founder, right? So that's all about me. Before ⁓ working for the startup, I worked for many ⁓ like I worked for Cargill. I worked at Cargill's. ⁓ handling entire Asia-Pacific division of car deals, digital and social, all this, they were handling that. Before that, I was with many companies in Silicon Valley, in New York, marketing agencies, startups in California, in Florida, Miami, right? All these startups, I was the first employee to join there because I love building things from scratch, right? I mean, there no one in my marketing team or AI team, I was the first one, first hire and build the things from scratch. Before that, was with GE Healthcare, leading the entire GE's product marketing in Australia and New Zealand. So likewise, I love challenges, right? So that's why kind of brought me to where I this thing, RQQT. Apart from all these things, I'm also consultant to many startups, as in fractional CMO, many of the startups, because I love also helping other startup ⁓ grow, right? So whenever it's possible, I consult them, advise them on how to set up their startup, how to do. venture funding, how to ⁓ do GTM, to establish those SMTGI workflows, everything that you sort of have been consulting them from last so many years. And I'm also a consultant to many, I would say tier one consulting firms. So when they need any my support, they kind of seek my support. I'm also seated at advisory board at many companies including Pinterest. I'm part of the Pinterest advisor. likewise, so this is all about me and my career and ⁓ journey so far. speaker-0: Yeah, I can see that you definitely have a lot of passion for creating things, building things from scratch. and I see that, you know, you, as you mentioned, you've built and you've been founding RX GPT as well as I see on your LinkedIn that this year you also co-founded or started higher X AI as well. So you do have a couple of ⁓ of those businesses running. Could you ⁓ first of all share with the listeners out there what inspired you to start? going into the more oriented direction. speaker-1: Yeah, So I'll say most of the founders, you when they start, the journey started, starts with a logical, with a, no, I would say mostly with the aim to change the world and everything. My journey was personal. I'll tell you what happened in 2021, around 22, I would say. Around that time, COVID was picking, and I was at the highest peak. And me and my wife and my brother, we three were, know, affected by COVID, right? we were admitted to a hospital and there lot of chaos at the hospital. You know, right, because of COVID, but still nothing was streamlined, nothing was proper at that time. So what happened that one day, ⁓ we all on same, I think, same level and the next morning when I woke up, I could not find my brother and I asked the hospital, like what happened to him? Then they said he passed away yesterday. So, I could not find brother Then they said he passed away yesterday. Now he why he passed away because he could not get those kind of, you I would say all this kind of facilities, all this kind of thing on time. Right. So that kind of motivated me that, I lost my brother to something called hospital operational negligency. Like all this thing were not available. He didn't have the proper lab reports. He didn't have the proper all the things. Right. So he passed away and somehow we survived because because of maybe our, you know, our I lost my brother to something called hospital operational negligency. I think was not that serious, but he was on a serious note, but he passed it. So that kind of motivated me that can I have been working at corporate like Cargill, I've been working at many other companies, to solve this AI problems, marketing problems at the highest level, can I go down and use my skills to solve problems at ground level, like fixing the healthcare, right? Healthcare is something that is essential to everyone, but no one... But everyone takes it like a granted right? When you go to any hospital, go to anywhere, everything is kind of chaos. Nothing is structured. So that's ⁓ nothing, it was not even, never thought of making a billion dollar company or making or sending something that would change world. So I thought, can I use AI to fix this operational inefficiencies in healthcare? So that's the journey started. started from a personal call and now it has become my purpose to help hospitals streamline and reduce their operational inefficiency. speaker-0: Wow, ⁓ first of all, Sal, thank you so much for sharing this personal element of the story as well. And I first of all want to acknowledge like I'm really so sorry ⁓ for the loss ⁓ your brother. I know that it must have been such a tough time, especially with a lot of changes going on. ⁓ it's really commendable that you were able to take that tragedy and transform it into solving the problem so that it does not happen in the future when it comes to patient negligence. negligence or the missing the data and everything like that. So having said that, I would love to learn a little bit more about if you can explain to the listeners out there who have never came across RXGBT, what is the purpose of RXGBT? What is it for? speaker-1: Sure, sure. Let me just make it very simple, like I explained to you, people get it right. So let's say, you you walk into a hospital, any hospital, maybe in the US or maybe UAE or India, anywhere, right? The moment you walk in, right, you see, you know, the nurse sitting at her desk, she's supposed to be taking care of patients, but she's running behind the admin. She's trying to get the, you patient's number, trying to get the patient's, you know, I would say reports. She has to fill out those intake forms, verify those insurance coverage. Nurse has to get assigned a hospital bed. Same thing happens at the small clinics. You go to any small clinics in the West Coast, East Coast, even in India, even in UAE, that they have one doctor and one executive assistant. The doctor, instead of treating patients, the doctor is busy in paperwork, right? They're getting out patients' last six months records, last one year records. trying to do the billing, invoicing, right? Now imagine the time they should be spending in treating patients. are spending on fixing the paperwork, right? Now because of that, normally the same thing could be done within 15 minutes. The patient has to wait for one hour, two hours. Likewise, it happens to every patient. And that kind of, it's a... simple situation. Imagine you are at a kind of critical emergency and whereas the patient needs immediate attention, whereas you are just filling out the form, you're just kind of trying to do this paperwork. How lives are lost because of this thing? And to a data and survey, it of costs a of billion dollars of revenue losses to hospitals because they waste lot of ⁓ time, a of ⁓ things on Rxgpt does simple, right? Let's say, you know, there's a nurse, there's a patient coming in, it's like charge PT, Rxgpt, just go to our tool, charge PT, like Rxgpt and type in the patient, X has come in, please send me his last year, last one year, you know, kind of records, right? You just want to know, right? One prompt like you're writing on charge PT, it sends you everything within one second. And then you say that, patient Y has come in, can you just, you know, do his I would say registration to his thing in kind of thing. So it's like an agent again platform that does everything with one command like cloud coworkers. Likewise, it does everything likewise. Similarly, like a doctor is seeing a patient right now. Sometimes he has to go through this papers, everything to turn the pages. I mean, do you have six months report, one year report? Now imagine in Rx2PT, he just goes in and type in the mixer, the excess come. to visit me. Can you just show me, me analysis of his last one year, two year reports and you have all the reports stored in it. I know I will say maybe a drive, maybe cloud data. Now it can fetch all the data from that analyze, say that, okay, patient has this BP, blood pressure very high last year, it came to normal. We see an abnormality. So, our activity does not consult or advise doctors for medication, but it helps doctors to find out the know what's happening with the patient. And doctors normally know they have to go to the paper, then talk to their figure out what's happening. It takes you know, hours and hours. I imagine within two minutes, two minutes, you get the output. So how much time the doctor, doctor is saved. So then like length of seeing this one patient in one hour, can see 30 patients in one hour, right? Now it can also in terms of revenue, can help all the small clinics, health care, to even increase their revenue. Also, lives, saving lives, because imagine a critical patient who is waiting outside, and he can be checked within 30 minutes, within 15 minutes, right? So that's the kind of thing it does, So likewise, know, RIT has, we have built around seven agents. We're not going to go very thin, very vertical, vertically thin. First is lab report. any doctors and patients, they can... check their lab reports immediately, know, the radiology, all these things. And third, booking and that we have invoices and the, all the clinics, all the healthcare, healthcare are not call center. You don't have to have a, know, especially outsource call center to take a call and talk to a customer. You should be only focusing on saving lives. Now we have invoices and not like you and I are talking, right? So we can talk to customers, talk to patients. give them real time answers. It's not like a pre-fed AI bot, it's a rag-based voice agent, which can understand your past and ⁓ current and advise and kind of tell you what to do, right? And talk to you. Let's say you're calling to the clinic, our voice agent picks up the phone, tells that, okay, Mr. is there. Okay, he's not there. He will be coming tomorrow. can I know a thing and I know a thing. Then it can also book that slot and send you an invoice within five minutes, right? So the time it will, so what are we doing is that we're saving time for patients, doctors and hospitals and saving time in saving lives, right? I imagine you you catch an accident and suddenly you want to figure out, I know, is that? Do I get a O positive, B positive in the blood bank? Now you are making some random calls and everything to figure out and you know, it comes to emergency, time means time means health, right? It's not about having everyone has money, right? But no one has time, right? It's kind of rushing against the time. Imagine time when you call to hospital to our voices and ⁓ it's to all the hospital network, all the kind of clinics across the US, across the thing, it can simply tell you, okay, yes, you should go to this hospital in Maryland, not to here because over there we have availability of doctors so and so. and we have your B positive or O positive plot is there, you can just go there, right? So that kind of thing it can do. So that's the beauty of Rx. speaker-0: Amazing. I think I really love the explanation that it's all about efficiency, but it's also about not only saving lives but making room so that doctors are able to have more time to do their job or taking more patients as well. And something I do hear a lot is that the doctors have such limited time and as a result of that, they won't there aren't they're not able to give the full attention and care as well. And hear this a lot also from my personal experience and also from friends and people and so on. So I think this platform a tool, it is like you mentioned, it is really life-changing. It is gonna save lives as well. So I'd be curious to know as like follow-up question. So this platform is this like integrated in specific clinics and hospitals, or if so, I would love to know in terms of numbers, which hospitals, how many clinics or maybe health insurance companies, how does it look speaker-1: like so far. Sure, sure, So right now, to be honest, we spend around nine to eight months to do research. Healthcare is a regulated industry. It's not like any other open industry like hiring. Now we can just do something to build a tool and try. Healthcare, even if you're building anything either SaaS or AI, you have to follow the regulations. All the compliances within the US, within India, within the UK, a lot of compliances there, right? So it took us eight to seven months to understand. to navigate those compliances or regulation, do the research. Because you are not allowed to, let's say you build an AI. Your AI cannot treat patients. It cannot say that if you are depressed, take this medicine. It cannot do that. In the healthcare, you are not supposed to do all this kind of hallucination. It has to be the precise thing you have to do. It took us seven months to do that. And then we built the first MVP in December, I think around January 2025. We launched a second version in March 2025, almost like four months ago. Now in the month of April, you started at GTM, go to market, right? So which we kind of, we have around 13 plus incoming contracts from big hospital network from India, from UK. US takes time because of the HIPAA compliances. We all know that. So I really kind of respect the regulation, but it takes time to the foot on the door, right? ⁓ we have around, ⁓ 17 plus clinics where they're trying a pilot right now. They are paying customers, but they're trying. No, they're kind of trying that pilot on almost because the moment we, it's not about getting a solution and kind of launching it to kind of getting the customer feedback and improving the solution and then launching big, right? So right now we have like almost, I would say pilot is going on. have like 17 to 18 clinics who are paying for us and we're kind of trying our solution. And then we have around 13 to 14 contracts. But this incoming hospital network, have like 200 plus hospitals in one network. So it's what we were talking about around like 600 plus, no, around like, around like 7,000 plus hospitals, right? To cater to those many hospitals, we need to upgrade our infrastructure. And that is why I'm also raising my series around, because I don't think that the kind of infrastructure that we have, can only cater to clinics. maybe 100 plus hospitals we cannot get it to 6000 plus hospitals we cannot do that right now that is why I'm raising a series day round to get to make us get to there so that's the status of RCTT. speaker-0: Yeah. And you mentioned that currently there's some pilots running in you mentioned like six seventeen hospitals or something like that. Where are they all based? Which locations? ⁓ speaker-1: Yeah, the location mostly are, as you said, like Silicon Valley, West Coast and around California. And there are around four to five hospitals also in the UK. And one or two are, you will be surprised in Nigeria. People are actually open to that, right? Why Nigeria is that? Because the barrier to entry is less and people need healthcare assistance there. Because I don't know, if you go to US, even go to Canada, the healthcare is pretty much, pretty much good, right? whereas no one is talking about Africa. We thought maybe, I mean, of course we always think about our profits and all, but at least for the sake of humanity, can we do something about Africa? That's where people need it, right? I understand they are not able to pay that much that US customers are able to pay. I'm willing to take a loss, but I'm just doing it to save lives there. It's not my investors and all, can know. The previous ones, when they read the similar, they kind of question my... vision that why are you launching in Africa is not a money-making region to be honest because they want to be able to pay the kind of customers in the US and UK will pay but I thought I said told them I'm not doing it for making money I'm just doing it to help people right so because a business has to have a sustainability angle it cannot be simply operating on a profit profit profit basis it might be a loss making decision for me when it comes to Nigeria and Africa But I'm proud of that because I'm, to AI, I'm saving some lives there, So maybe in the future, when you have a larger round, maybe a series around, we'll be able to out in Africa. But of course, there is a need to sell healthcare columns in Africa. Let's do it. Let's not just think about launching in those profitable regions in the US and making some money. That's not the purpose of our activity. ⁓ The is always to... help people and save lives, right? That's the best way, that's the best region. Africa, that is where we're launched in Africa. speaker-0: Mm-hmm. Yeah. At the end of the day, it's also about the impact that you leave. And I can see that of this really started because of the personal tragedy that you've been through. And then also that's why you wanted to continue to pass on and leave that positive impact as well. ⁓ ⁓ have a follow-up question when it comes to like the different locations. So I myself, ⁓ half German. So you know, Germany is, of course, really well known for the healthcare insurance, but at the same time. ⁓ They are a little bit slow when it comes to the digital infrastructure in terms of it takes a long time to get appointments. It's very slow. So I'd be curious to know, do have any plans that rolling out in Germany? speaker-1: Absolutely. now we're kind going very focused way. mean, at least have this two or three pockets where you go vertically and we kind of know at least capital markets. maybe in the next phase, obviously, we will go to ⁓ That's not a big deal. But right now, just being very focused to solve these problems at ⁓ very focused level. And then once these problems are solved, once they get some results, then definitely we'll. go to Germany and all these areas where there's a need for this application. speaker-0: Okay. And just out of curiosity, let's say there's a patient, they lived in one country and then they use this platform for all the information about their medical history and then they move to a completely different part of the world. what are your thoughts in terms of because you know there's also the aspect of switching healthcare insurance, ⁓ they need access to healthcare history. Do you I think this would be kind of like the the tool to be able to escalate that. I can imagine some people they might have a very complicated history of chronic illnesses. So I'd speaker-1: So in terms of patients data, have everything centralized in a centralized database cloud storage. So accessing the data won't be a problem because it's the same data that stored in the cloud like the way you're using ChartDipity. Let's say you're logging to ChartDipity now and you travel to Bali or travel to Maldives and you're logging from there, your same statistics will pop up, right? But likewise, when it to patient data, it's fine, but you have to... comes to the reset insurance providers, healthcare providers, that will be kind of changed. So for that thing, we should have a, that is why we need a wider access right now. We have like 13 to 17 clinics and just 13 hospital network, right? So the more hospital network we have, insurance provider we have, then we can switch easily. Or else we'll have to talk to that, let's say new insurance provider, new hospital healthcare in that location, in that geography. and try to set up a contract with them. That's the only way. Or else he has to just, when it comes to incident providers or hospitals, he has to be working with our current partners, not with the new ones. That's the only way. But we are thinking through, that's a good point, I must say. We are thinking through, is there a way we can make it universal like that, but it's a B2B, right? It's not a B2C platform. B2C means you can easily make it. B2 means it has to have some kind of, collaboration, the partnership deals to switch, make you switch from hospital A to hospital B. Right now it only works with the current set of vendors. Let's say hospital A is our vendor. it works with they have ⁓ hospital ⁓ the world, it works with them. But when it comes to new hospital, yeah, we'll have to either be partnering with them ⁓ doing something else. But yeah, that's a good point, I would say. That's a very valid point you raised. speaker-0: Mm-hmm. Okay. ⁓ tell us a little bit about your behind the scenes. ⁓ how does your team setup look like at RXGBT? speaker-1: Yeah. Yeah. So RxCpt is a big team. It used to be a big team, but right now it's kind of become lean. We have a team of around 23 people right now at RxCpt. And initially for nine to 12 months, we only focused on building the AI thing, right? So we have a CTO under which we have all the SMT. To make an SMT-AL platform, you need a front-end team. You need a back-end team. You need a cloud like AWS GCP. You need also something called RAG. Who understands RAG better? Without RAG, it's nothing, right? Then you also need someone who does very good UI UX design, right? And then you need a testing. I mean, who could test this application before launching it. So that's a core tech team, right? And as a startup, don't think we'll have, we have, need many people. mean, we need, of course we'll need later on, but right now, since everything is clean. We have like very focused roles, very focused kind of thing. So two or three people working on one problem, like one senior, one junior, that's what kind of thing we have. And when it comes to marketing, if you do everything in marketing, you're doing nothing for startup. We have to be very focused on one strategy and one area at a time. So we have a GTM lead and we have an intern, GTM intern. And then we have like finance. finances also required, but he is a kind of consultant. He's not a full-timer because right now it's all about building scaling up to so we don't know the PFO. So we have financing and there's also HR for because hiring is very essential for us. People say AI will reduce hiring. No, I don't believe in that. We still need people. We still need people. I don't think Claude can build this tool within one day. Claude can write some code. But still, to understand people, understand the requirements, we need people. We need people in the room. We need people build. So hiring is very essential for us. We have an HR to do that. And I, as a CEO, I kind of get into all these pockets on the highest So that's all ⁓ that have right now. speaker-0: Yeah, just wanted to acknowledge that point, right? There has been there's been a lot of fear when it comes to AI replacing jobs, but I think at the end of the day it's come to show that AI is sort of like it's sort of like a supplement, but at the same time, we still need that human element as well, right? Yeah, for sure. Yeah. speaker-1: Right, right, right. AI cannot replace human consciousness, intelligence, right? AI just said do, like you used to, you you need a phone to call to your friends. Previously, you used to have those landline numbers, now you have an iPhone, right? So AI is like an iPhone, and iPhone cannot, no, iPhone has no need when there is no person, right? Likewise, AI has no need when there is no people. And if you see a lot of the companies who have fired people to replace them with AI, now they are rehiring. So we need people, we need people, but we need people who are skilled on AI. That is why I advise everyone. Get skilled on AI. We need people who are skilled on AI. yeah, we have to replace jobs, but it will replace people who do not have skills in AI. That's what I could say. speaker-0: Yeah, just the other day I was having a conversation ⁓ with another guest and they were also talking about that AI essentially is kind of the new Microsoft tool, Microsoft Excel. It's like it's an essential tool that everybody needs, everybody needs to acquire as well. So you mentioned that for you in terms of growing growing and building your team, what's really important to you is that they have experience working with AI. So from what you're seeing right now in the team setup, you've had a you have a really structured setup, you know, in in ⁓ know, in the tech world and marketing and so on. What do you think ⁓ the most difficult or role that you have to hire or what is like the operational, the biggest operational challenge that you're facing right now? speaker-1: Yeah, the biggest operational challenge we're facing is, I'll be very honest with you, the market doesn't have the right AI talents, right? People who just use charge to write some prompts, they call them some AI engineers. They're not the AI engineers, right? We don't have a full scale or full stack AI engineer in the market who understands how AI work, what are the AI governance models, what are the AI ethics things, how do we orchestrate AI agent, how do we build an AI workflow? People use N89. N89 ⁓ a, ⁓ like a canvass of MCPs, right? I mean, that's not the AI, right? We need the actual AI engineers who understands how to build this entire solution, right? And let's say, for example, I'm trying to automate HR or ⁓ automate hospitals, right? ⁓ unless I know the operational expects of hospitals, how can automate it? Let's you tell me that, can you make something that would record the podcast and just edit it, delete it, right? Now, if I do not know the process of how the podcast is recorded, how it's edited, I won't be able to automate that, right? So that's the domain knowledge lacking, right? So that is why, so that's the first thing, right? AI engineer with the right set of domain knowledge is not in the market. And the pressures who are there, they're good, but they don't have a domain knowledge. someone has a pressure who just graduated with an AI degree or maybe anything like that. They have not worked in the healthcare. They have not worked in the insurance sector. They have not worked in any of these things. So how would they have that domain knowledge? And someone who has the domain knowledge, like 30 years, 20 years, he cannot even function in AI. He's like too much, we try to do that. We try to upskill them, but they're like too orthodox to AI, ⁓ Langshan, learn... QA learn land grab learn Python right now. So that's the thing. There is a lack of people who have both the domain knowledge and the expertise you can find in software, you will find people who walked in walked on maybe AngularJS or Java for the last 20 years and work in healthcare. So they have both domain knowledge and skill, right? In AI, that is not possible. You can't find very, very rare talents who have both. domain knowledge of hospital healthcare and they also know how to build AI solutions because AI is new. That's a challenge I'll say. Second thing, sales. not sure. I am not able to get salespeople. People who understand how to sell, how to position a product in the market. Maybe it's new because of that. Maybe AI was always in the field of research, ML ops right now. Now because of agenti AI, it's been, you know, getting commercialized. So we don't find many sales, or sales guys who can not this solution approach customers. ⁓ would say, ⁓ it, position it and close it because they are all into either hardcore fast sales or hardcore B2B sales or hardcore, all this kind of software. The software is different from AI. ⁓ not right. So that is where the finding is even AI cannot replace you. can build a SDR agent using agent AI, but that cannot No, go to hospital and close the deal. Right. For sales you need that human intervention. Even if I introduce one AI agent, like we have built one AI agent there, but it can just send emails all up, but it cannot have that human connection to know. mean, the conviction that all the skills that a sales guy need to convert a customer, right. The sales and the first thing, second thing, these two are the primary, I would say. Yeah. think that we are not able to solve, right? Having a right AI talent with the domain expertise and second, the sales guys who can, sales for us. speaker-0: Yeah, absolutely. No, I can definitely see that those are definitely like some of the operational challenge that you also want to solve next as well. And definitely especially because we're as the world is evolving with AI, ⁓ that's also like it also requires like this transition as well. And I can definitely see that it is very it brings its challenges too. okay, Silas, just one final question. If you could pass one advice to someone thinking of starting a business, what would it be? speaker-1: I would pass so many advices, Anyone who wants to start a business, I'm sure, you know, the audience, younger audience don't like the word AI. Anyone who wants to start a business, get into agentic AI. Agentic AI is the future. Find out any traditional business, maybe logistics. healthcare or maybe supply chain anywhere just go to something called if you know about this site called craigslist is a very famous site in the US right go to craigslist find out those traditional businesses understand their workflow their operations automatic that's the is going to feature because see when when you met know when you made those internet draw all those guys who build those no big that they don't make money. But the guys who build this Google, Microsoft, right, all this on that internet web, they made the money. So anyone who's building, who build those models, or maybe LLMs, all those data centers, they will make some money. But the ones who build those commercial products of MTG, they will be the one who will be the next millionaire to billionaires, right? To get into the field of MTG, first, would say, right? And second, I follow something called the theory of airport model, which means if you go to any airport, right, your flight never wait for you. You might not be ready. You might be not ready. Never wait. So never wait for the right time. You, because AI is new. You might feeling that it's not complete. It's on pilot. It's not even complete. I showed you waiting. Don't wait for the right time. Launch it, scale it, get to the customer, get the feedback. It's okay to have a better product. It's okay to have. That's how everyone starts. And then you kind of that right those two things I can say to anyone who is starting to do something in the field of entrepreneurship. speaker-0: Perfect. Thank you so much for the very concise and helpful advice. I'm sure that it's definitely going to inspire and also like give shed some light for people considering starting a business as well. Well, Sal, it was very nice having you on the podcast. This is Carmen Durvald and Sal from Behind the Founders. Be sure to stay tuned for the next episode. speaker-1: Thank you very much. Thank you. It was pleasure having been part of the podcast. Thank you very much.