The Tech Glow Up - Fabulous conversations with innovative minds.

The Accidental Entrepreneur Who Built 50 Apps Before Knowing What a Startup Was – PJ Park

• Nathan C Bowser • Season 2 • Episode 8

Share episode

Death by a thousand clicks. That is the problem facing clinicians who spend hours navigating scattered tools for documentation, billing, guidelines, and decision-making. One doctor-turned-founder accidentally built 50 apps trying to solve it before realizing he was an entrepreneur.

PJ Park is co-founder, chairman, and chief product AI officer at Avo MD. He came to the United States from Korea about ten years ago and joined a residency program barely able to speak English. On his first day, a senior asked him to call a dying patient's family. He missed everything. 

That experience drove him to start building software on his own to make his "imperfect doctor" perfect. He built app after app during residency until he had created 50 different tools. His friends finally told him he should start a company. He had to Google what that even meant.

Avo MD is an AI clinical copilot platform for clinicians. Unlike scattered point solutions that each solve one narrow problem, Avo MD builds shared components that work like Lego blocks across workflows. The platform handles admission, discharge, rounding, and charting by combining patient data, hospital guidelines, and evidence-based protocols. AI makes recommendations, then doctors discuss and decide. The goal is a meaningful doctor-AI relationship rather than just more clicks.

Highlights from PJ Park at Avo MD:

  • Built 50 apps during residency before friends told him to start a company. He had to Google what a startup was. His only goal was making his imperfect doctor perfect.
  • Partners with content and IP companies like MCG for evidence-based guidelines. Turnaround time is 10 days versus six months to a year for larger companies. AI consumes proprietary guidelines to make better outcomes.
  • His new iron triangle for healthcare: patients get better, doctors go home early, hospitals make more money.

His insight about the industry is that AI scribes are the first AI solution clinicians actually love because they were not built by administrators forcing compliance. But scribes only cover patient encounters. Most clinical care involves connecting dots between guidelines, protocols, documentation, and billing without any recording to transcribe. That is where Avo MD focuses.

Healthcare gets better when AI takes care of the technical checklists and lets humans do the thinking.

Live from HLTH 2025 - Watch on YouTube.

A "glow up" signifies a positive transformation, reflecting the journey of becoming a better, more successful version of oneself.

At The Tech Glow Up, we humanize the startup and innovation landscape by focusing on the essential aspects of the entrepreneurial journey. Groundbreaking ideas are often ahead of their time, making resilience and perseverance vital for founders and product leaders.

In our podcast, we engage with innovators to discuss their transformative ideas, the challenges they face, and how they create value for future success.

If you're a founder or product leader seeking your own glow up, or a seasoned entrepreneur with stories to share, we invite you to join our guest list via this link.

Nathan C:

Hello and welcome to the HLTH Tech Glow Up. I'm Nathan c and today I'm talking with PJ of Avo, md. PJ, thank you so much for joining me.

Pj:

Great. I'm PJ, one of co-founders and chairman and chief prop product ai. And nice being here. we are working, I'm working on core, a company called Avo MD. This is the, AI platform, clinical copilot for clinicians.

Nathan C:

Mm-hmm.

Pj:

And then I have an international background from Korea around 10 years ago. I also a clinician from my side, Doctor Harvard Medical faculty, but I decided to dedicate myself to the intersection between AI and medicine. So excited about our future. We are trying to build a platform where all of their are solving problems. There are so many. Problems in the world. So, but in a scale of a way so that, we are expecting and scaling, like building solutions and label like fashion in our platform.

Nathan C:

I'm always very interested about people's origin story, about why they begin an entrepreneurial journey. Being a clinician is a full day job, right? so I'm curious what. Sparked the inspiration to start working in AI and technology in addition to, your other practice.

Pj:

Let's say I never really seen myself as an entrepreneur. I mean, I never had a dream. Or like, okay, I want to build a company, CO, whatever. I just want a doctor. I be a doctor. And then during, so I, I was an intern. I came to the residency and then of course I was not their smartest doctor during the residency, but there was a language barrier. The first day I was able, barely able to speak English. What I remember is that like my senior was asking me to, Hey, this patient is dying. Contact the family and to let them know that. Oh my God. Like of course I missed everything. Life happens. Yeah. yeah, so that be given my background on engineering, AI research and ui ux, one of the thing I was really focused on was like, okay, building softwares on my own. Let's keep some kind of like relationship between app and human doctor and then make my kind of imperfect doctor. Perfect. So that was my dream. it was just hobby that was doing. Since my residence develops, I ended up with like 50 different apps. Then some of them are featured in the newspaper and everything. So that it led me to Columbia, Machine learning research over there. Also, I was very interested in ui, ux and then eventually I noticed, my God, I cannot, what's the point of building 50 different apps? Those kind of things actually. Mm-hmm. Still happening. at that time I thought that we might able to build a platform where there are some. Common, like shared comp. Share the T between all of these applications so that we can scale and solving more problems. Best and faster and then I had no idea what company is, but my friends and my coworkers, and then my future co-founders, they're coming in and told me that, Hey, you should build the company. What does that even mean? Like I started Googling and then I thought that, okay, this might be the only way I'm so excited about this project. Only way to make it work is to make it a business making money so that this is sort self-sustaining. This is how I started. That's a little bit different from his stories from others, because I know that many people this big dream of like, I wanna make money. I wanna like, make a big business change to the world. Yeah. I went little bit different way. I,

Nathan C:

I love that. Like the entrepreneurial spirit and the like, here's a problem that I'm experiencing in my day job. I'm just gonna go. Sit you accidentally fix 50 different problems. And then people are like,"Hey, like, this is really like what an entrepreneur does. You should probably start a company," I love that you were the last one to know that you had this entrepreneurial mindset. It was just like part of how you're solving problems. can we dive a little bit into the solution with it? AVO MD and, with the product that you're working on now, and, help me understand, you know, the health tech, audience, or the tech glow up audience doesn't necessarily know a ton about, I see health technologies, I see. So, if you can, kind of introduce, your product and,

Pj:

so let me, put it this way. I mean, there are many different sectors in healthcare, like patient side and then the payer and everything. I'm always focused on clinical care. Improving clinical care and then any. Revenue improvement, inefficiency, issues in clinical care where providers, clinicians, and nurses involved. So focusing on that area, what we do is that there are so many problems. And then with complicated workflow, let's say I'm automating a patient. Like. I'm assuming there are so many AI technology over there. so that, like, you're still doing all the like EHL, like epic, like chart review clicking everywhere. Mm-hmm. epic has some features. You summarize some portion of it. Okay. Now, some of clinical care in involve patients doctor encounter so that you might use an AI scrub system, make a report to enhance. your clinical documentation, now you need to make some decision. Your ology is imperfect and then you need to go to like UpToDate, probably not Wikipedia and, and everything, and there's billing solutions so that all of these solutions are scattered and high focus. In a, in a niche, like we call them like one trick 0.4, one solution. and then I know that we have a thousand different problems and then this is not really working in a way we want to change the world how healthcare change behave within five years, four years. But it's,

Nathan C:

it's like death

Pj:

by a thousand clicks. Uhhuh, right? Uhhuh. Yeah. So then we decide to build a platform. Okay. We know. You said that that's by a thousand clicks, that's an example of really, really bad user experience. Yeah. but we noticed that there are some common charitable competence between this workflow, for instance, in clicks. I, I'm looking for your, like, advice on it so that we can come up with a better solution so that it's more deeper ai, Cuban doctor relationship. Those components can be shared between things and then you can build this kind of solution really, really best. Just mix and matches kind of Lego block. Mm-hmm. So that we are sort of many different clinical problems, like better admission, better discharge, better like patient encounter, outpatient, inpatient, or more.

Nathan C:

Ooh, it's always exciting, when you have a new tool and you can be seeing, all the different layers, where you can make an impact. the name of the show is the HLTH Tech Glow Up. A glow up is a notable transformation Or like a rebirth. Let's start general with the health technology or health industry as a whole. do you have a goal or is there a glow up you wanna see in health technology, like in the next six months? What are your big goals or, challenges to the industry?

Pj:

if you say six year, I have a lot more to talk, but you said six months. Yeah. Is that, I'd like to say two different things. One is. In clinical care, like everybody's talking about, is now AI scribe.

Nathan C:

Mm-hmm.

Pj:

I want to see something'cause I, I really love AI scribe'cause this is like first AI solution people ended up with loving it. Most of the previous solutions are mostly driven by healthcare administrators. Concept, enforce certain actions. You should do it. You should do it. Click, click, or like pop up. Don't ignore it. Kind of in a way, build with the distrusting clinicians that they should follow so that people don't like it. Not all adopt, so you don't have to change the world. But AI experts probably one of. Few, all these solutions is actually working, but I think it's still covering on the very small portion of care because surprisingly, there are a lot of clinical care area that patient encounter is not really happening. For instance, if you're discharging patient as hospitalist, you're talking to patient for five minutes. And therefore next two hours is there not just documentation, you just need to connect all the dots within guidelines, like hospital protocols and what's the best thing to do to send out prescription and then make the right documentation? The billing. And then even deciding all of, basically this like evidence is the patient even read for discharge, like kind the way competi making is there and AI scrap cannot really tell you anything because there's no recording. As private life kind of make an assumption that what you spoke during the conversation is the ground trust, but here, doctors trying to make a ground trust right thing. So how can you help those kind of areas is one thing. second part is I want to see happening more beyond EHR companies, such as like, eh, like epic because like when we are doing this kind of things, everybody asked. Or what if Epic is doing a, B, C? Mm-hmm. that's a really good question every single investors ask us, but I think it's somewhat misleading questions because Epic Technic can do it'cause they own initial user. I terms layer, they have all the technical like stack and then pretty much operating system. But on the same way. Like Microsoft build windows. Okay. Like let's say somebody build amazing game like Roblox, nobody asking people to build Roblox. Like, okay, you're making billion dollar business. What if Microsoft is taking over Roblox and building on their own, their own Roblox? Mm-hmm. like what if Microsoft building EHL, I mean they're clearly at advantage. What if they have like default application roadblocks and EHL. It doesn't make sense because first it's not gonna be happening that any single company can do, covering every single area in the world. Okay. It's impossible because that's why startup didn't exist. Yep. Yeah. people have a different culture, different style of teams. having money doesn't really guarantee anything. That's why I started with individuals. Second delay, they shouldn't do it actually. like same way. Not necessarily Apple, iOS. They, they don't need to build it every single app. Okay.

Nathan C:

needs to be a platform level tool. Sometimes, other tools, building on can actually make a more robust system, right? Sure. If, if that's

Pj:

happening, the, like, iOS, apple will be almost like Soviet Union, like trying to like govern every single aspect. Security. It's not gonna be working, it's gonna be inefficient. That's, that's why they don't do it. It's clever. And then even Epic and many Edge company, as I talking to, they don't see them as. They're doing everything They know, they want to retain some core elements mm-hmm. As a fundamental experience of e hhl, but all the areas, they don't want to touch it even. They want to touch it and see that very optimum success.

Nathan C:

you make a very interesting point about where a platform is interested and invested versus where a more focused solution can provide deeper value. Oftentimes there's this assumption that platform play is the best way to go because it serves so many needs. But if you're serving so many needs, you're maybe not serving them as deeply with as many insights with. the cultural, workforce specific nuance, that a targeted, solution can provide. I really love it. so let's move on and actually, turn, towards your company specifically. What's the glow up, that you're working on for AVO MD? What are your big goals, for the next six months?

Pj:

next six months, not six years, right? Yep.

Nathan C:

I mean, you can give a six year answer

Pj:

as well, like taking over their medicine and we'll just

Nathan C:

to podcast a few more times on the way.

Pj:

so I told you that we are a platform which allow us to build like any types of pil. Whether it's admission or you may with anything scale away. So that technically speaking, I'm want take over every single workflow in medicine, which is not practical. We cannot do that now. Not yet. So we just, we are hyper focusing on a few main workflows such as admission, discharge or rounding. And then like charting different patients visiting combined with AI Prime. and the proving ROI on it. for instance, let's say we are solving like discharge, so there will be, like there are some patient doctor conversation elements coming in that proven from the EHL, like talking and then eat this. Our solution is talking to hospital guideline and then like Universal United States guidelines, making good, recommendations and doctors coming in, and then kind of make discussion with doctors, Hey, this is an idea. What do you think? This and ai, and then I'm trying to make this like meaningful doctor AI relationship, as I said, with the right expert than just asking for clicks. This is an idea. This is why. So the users will be able to acknowledge and agree with this disagree so that they both make better outcome, working together faster and better quality and hospital making more money. I'd like to say our kind of like long-term goal is how we can make patients get better. Doctors, go home early hospital, make more money.

Nathan C:

Amazing. That's kind of a new iron triangle for, healthcare products. Right. Like, for healthcare outcomes, I guess. I love it. as you're thinking about this glow up, what's your approach to partnerships? And, your overall go to market. who are you here at Health looking to meet and, how are you growing?

Pj:

one way to grow, now we are exploring series A. one of the things we are hearing is that We found product market fit. People love our solutions, but we also need to prove that we want to get loved by end users. for don't love it. We are not going anywhere, but hospitals should pay more money because we have ROI. So we are trying to prove our ROI and that we are actively working on our partners. when we say partner, we say content and IP. AI needs data, AI and IP doesn't go well in general. You know, that open AI releasing new video generation, like people are getting crazy, like studio, anime style, whatever. But in medicine, things are different. we talked about discharge a bit like hospital system, like doctors trying to make a safer decision for discharge so that they can discharge faster and safer. it's not really about ask ai, Hey, can you discharge patient today? Mm-hmm. there are series of like evidence-based guideline pieces. Validated and proprietary and then AI cannot own it. Yeah. They can't realize it if be given. So that, what we are doing is that, that this proprietary IP has a lot of, like ROI if proven ROI, so that, but what we notice is that they. Have consistent fear with ai, they're gonna be less and less like meaningful.'cause people cares less about like opening PDF guidelines. Their, their doctors list of doctors build a PDF guidelines to hundred pages and people are less tolerant on those kind of things. Mm-hmm. But AI is really good at consuming those to make a better outcome. And then we made really good fit with this. this, the partners, because they're not tech company at all. We have a platform of like building user, friendly, clinician friendly interface like end to end workflow tool for better discharge of mission, whatever. Our turnaround time can be 10 days within 10 days. Meanwhile, like big companies taking half year, one year. So that's where we are moving to. For instance, we partner with a company. Called MCG, which is the building guidelines for mission and discharge. And we are the only partner for them, to build a discharge like assistant, helping doctors. So to solve everything but discharge make you vital documentation, better planning. Incorporate inclusion. our inputs, AI scribe if necessary, incorporate If they don't to the patient, that's fine. We handle that, those kind of things. So that we make more money, they make more money, and then we take some of advantages and then'cause also EI companies don't do this content play at all.

Nathan C:

Yeah. I love this. this is like a very good example of one of my favorite applications of AI where you use AI To take care of the technical things and the checklist kind of things, right? What is the right process? What are the indicators? the AI is focused on a very targeted, part of the process. So you let the computers do the computer parts, and you let the humans do the processing and the thinking and the cognitive, it's just a perfect, alignment of like asking the tools to do the right thing. For training and for, right. Like the humans in that interaction can change. But because the platform has the right context and instructions, it can keep the standards in a way that's like hard to enforce otherwise. So pj, I'm so enjoying this conversation, but we have one last question. the theme of the HLTH show this year is Heroes and Legends, and I know that every entrepreneur has at least one mentor Or somebody who has guided and encouraged them along the way. That's really made a difference in how they got to where they are today. So I'm curious, how has a mentor, or a leader impacted and encouraged you to be where you are?

Pj:

I haven't. So many mentors, but like, I probably don't want hear the oath. I thank you. My parents or things like that. Yeah. Not quite time for an Oscar speech. it's interesting somebody asks me little questions and I had no idea what to say. because maybe this is sort of especi in medicine because if you ask me who's the legend in physics. Einstein, like who's a legend in cooking? Like, okay, Gordon, or kind of things, but who's the legend to a doctor? Like, I'm in medicine, I'm a doctor, I have no idea who the person is. Maybe we have some kind of culture like suppressing, like individual cult kind of thing. like El Musk and or Sam Ultimate kind of things. I mean, I don't mean that there are no heroes. A lot of heroes working in a team. secondly, if you focus on not just being a doctor, anything like the clinical, there are usually innovations coming from industry first, nonmedical, large language, everything. And then it takes so many years connecting all the dots and complexity. The big tech is not really able to solve. Even health company healthcare company cannot really solve. So there's a lot of struggle. Trial fail, people try Google. Glass not working is dead. Those kind of things continue still happening. so that usually for me, there legend, now it's almost, nobody really talks about Apple nowadays because of ai, everything. but when I first starting it, my. Legend was, I feel a bit shy to say this, but it was Steve Jobs actually. Awesome. I love it. not many people talk about him any longer. more people might talk about the same ultimate nowadays. But because the reason is that, that's when I was working the ICU watching his keynote, and then what I really enjoyed is that he's like deep, hyper-focused on. Session on and part of the desirability. Yes. Like pj, I love

Nathan C:

it so much, but we have to wrap it for our time. That's such a fantastic place to end. I love the accidental entrepreneur, and wrapping with Steve Shep. PJ AVO MD, thank you so much for joining me. I see you. I'm the tech, HLTH Tech Glow Up Thank you. Amazing. Thank, thank you.