Ahmed Omar's Journey: From MMA Fighter to AI Pioneer

Jake Aaron Villarreal: Welcome to our podcast, From the Ground Up, where we interview startup founders, exploring their journeys, their successes, challenges, and lessons learned. We hope you'll be inspired in discovering what it takes to build a thriving startup. I'm your host, Jake Aaron Villarreal, and here with us today we have Ahmed Omar, the co-founder of Sully.ai, a startup with founders from Google and folks that helped build ChatGPT. They're on a mission to help doctors save time and help patients save lives. Ahmed, welcome to the show.

Ahmed Omar: Thanks, Jake, for having me. Appreciate it.

Jake Aaron Villarreal: Great. So, a little bit about Ahmed. Ahmed started from a very young age selling products online at 13 and saw how online commerce solved a vast supply and demand problem, making him eager to start more and learn more. He found a passion in building companies. Uh, he's built a total of six companies today. One of them was an e-commerce brand company in New York City that grew from zero to $55 million in sales over three and a half years. Then he went on and started another company, Odigo, that grew to $14 million in ARR in less than three years. Fascinating experience at such a young age. But there's more to your story than just building technology and businesses. You also are an MMA fighter, have a three-time gold... you're a three-time gold medal champion. And I'm sure some of that has translated into how you run a company or run a business. What led you into MMA?

Ahmed Omar: That's a very good question. I think building companies is a little bit harder than getting punched in the face, I would say in the face. Um, but I think I was really... I really liked the, it was for me like an art. I started, I started actually with boxing. It was really, I was really obsessed with Muhammad Ali and, you know, how passionate he was, how driven he was. It was really... I saw that, and I was always, you know, trying to protect, you know, my friends, protect my family. So, strength was something, you know... Me and my dad, we used to spar when I was a little kid, um, like seven, seven years old. We used to, you know, like, that's how we had fun, we used to fight. And um, and yeah, at the age of 13, 12... um, 12, basically, I started basically learning how to fight. You know, and it made me more disciplined. It made me respect myself more. Uh, it made me, you know, more determined. It made me more, I would say, confident about what I could do. So, I do nothing, basically, to someone that I can know, um, you know, if my strength is better than this person. So, it developed, I think, discipline, a lot of things towards, you know, being a pro athlete. Um, and you could see this, you know... And it's also in building companies, this 1%, you know, daily improvement and this, you know, this final fight. If you can work on yourself and make yourself, you know... this last 5% is what makes you a great founder compared to, you know, an average founder. And um, and and... or a great fighter and an average fighter, you know. Um, and this is... and yeah, it basically was something that, I would say, helps a lot with me building companies. Helps a lot with being persistent. You know, even if things are not working, finding ways because I think something really, uh... really interesting story that I tell a lot of people about that I don't really share, um, with a lot of people, but it's really interesting.

The only person who took the gold Olympic medal in the Middle East, his name was actually Mohamed Aly. So we used to call him the Muhammad Ali of the Arab [world], you know. He's a heavyweight champion in boxing, he took the gold Olympic medal, it was a big deal at this time. And at the age of 14, I remember he came to me and said, um, "Ahmed, I think you really should fight another sport, you know. Um, I don't think that's your game. You've been training here for a year and a half, I don't think that's a good thing. You're not really, um, like you know, these kids have been training since they were seven, eight." And at this point, it started making me think, like, if I take his word and just, you know, go home, from an expert, right? That's the expert person. And go home and do nothing, I'll just do that. You know, I just keep repeating this scenario my whole life anytime someone tells me I can't do something and or I'm not good at something.

So yeah, um, turned out that I actually like boxing. I took the first place and second place in Egypt for boxing, you know, like the second. And then I took three gold medals in Muay Thai, you know. And then I started basically playing, you know, mixed martial arts, which I thought shows more strength than just boxing. Um, and yeah. And it turns out that, yeah, um, I'm a great mixed martial arts player. So I'm a great fighter, not necessarily boxing only, um, but I could still be a great boxer. So, um, I was able to get the championship of boxing and also, you know, mixed martial arts. So, um, it was just this drive that anytime someone tells me I can't do something, I have to try, you know? And I think trying is what counts at the end of the day.

Jake Aaron Villarreal: I love that. And you're not the only CEO that is practicing in martial arts at MMA. I mean, Mark Zuckerberg and many others, you know, on the side are learning and growing, and you never know, you know, what they're getting out of it. But it sounds like the discipline and the fight and the desire to to compete, all those things really are what make up a founder as a startup. Um, I want to talk about your startup here because I think in a lot of ways it's going to resonate with people, not just from a technical perspective but from a humanitarian perspective. What was the inspiration behind the idea of Sully when you first thought about it and what was the main problem it was originally working to solve?

Ahmed Omar: I think when we started building, like when we started looking at the healthcare industry, how a lot of people struggle. Even my co-founder had chronic illness. Chaitanya had two years of chronic illness, you know, our CTO, co-founder. So to see all these struggles and we know what we're capable of as founders—building companies, built tech companies, grew companies in many different industries—we kind of felt that there's not a central place that you could trust and go to in healthcare. And it felt like serendipity at some point that I kept asking myself this question that I never got an answer to until I started building Sully: was why, why was my mom uh diagnosed with polio? She never walked. And the root cause was she got misdiagnosed, which happens today in the US with over 500,000 people every single year. So it started basically, everything started coming together. As you're tech builders, you know how to build technology, you know how to scale technology, why not do something in healthcare that could change and save people's lives? And this impact is personally related to us as, you know, my mom, and the co-founders as well, like themselves, their health. And it just made perfect sense that this should exist.

And we started actually with diagnosis. And then we started saying it doesn't really make sense to start with diagnosis while doctors spend half of their time on, you know, issues unrelated to medical, something that they didn't go to school for. We have to solve this problem first for them, and then we can help them with diagnosis or misdiagnosis. And then maybe we can, you know, um build something for patients to use real-time to help them diagnose them earlier and get them to the right doctor earlier down the line, you know. So this was basically the vision of the company and we just, you know, switched the go-to-market for it to make sure we're not going to solve a futuristic problem while doctors today spend time and make mistakes because of their admin burden and burnout, basically, that they go through every single day.

Jake Aaron Villarreal: So the product itself is an AI-driven application, AI agents essentially for doctors that helps them do administration, scribing, note taking, follow up, basically all-in-one or encompassing platform that doctors can use that's going to automate the menial tasks that they typically do that they hopefully don't need to do much longer, and and helps them become more efficient. Is that kind of the the idea of the platform?

Ahmed Omar: Exactly. I think you articulated it pretty good. And this is exactly how we're building it. And the reason we went with the agent approach is that we saw a lot of doctors are being offered a lot of other tools and it's basically, you know, sold as a tool: "It will help you do X, it will help you automate a certain workflow." But it's not solving the biggest problem for doctors, which is the whole admin part. Misinformation of patients, misinformation and insurance information, declining claims, you know, all these things. They still have to call, you know, insurance companies. They still need the medical assistant to call insurance companies. So we started basically building with an agent way.

And I worked as a, behind the desk, as a... like in a clinic in Santa Clara for six months. And I started seeing all the scribe tools that are there today in the market. And I started seeing why are they, why are they not really working? Or why, from 30 organizations, 50,000 doctors, healthcare providers that we, like, we spoke to all these organizations holding 50,000 providers, no one's using anything. And we started looking at what is the gap here. And we started saying, the missing point of dealing with a human... you know, they deal with an AI scribe or like they deal with a human scribe that does many things. They don't only scribe. They add, order labs, they add prescription, they do all these things. They can follow up with patients, they can call, you know, patients. Um, they can call insurance companies for certain, you know, MRI, um, approvals or prior auth. So we started seeing and we started saying, let's actually build it the same way as an agent that could help either a human scribe that manages many doctors, or replace a human scribe. Because even humans like people that [are] scribes, they just, you know, training, they're not really interested in scribing. They're just training just to, you know, shadow doctors for a couple months and then there's like a very high turnover in scribes, like it's a three to six month turnover because they're not really interested in scribing. Even one of the scribes, I remember people that were scribing, they were helping us build the product. They were like, "Oh no, you should, the product should could actually do this as well." And they didn't really care if it's going to replace them because it's not the future job that they're looking at. It's just something that they train or, you know, um, get certain, um, you know, internship done or, you know, um, rotation done basically for their, for their medical study. So, um, so yeah, I I think the way we built it in agents, we we see it as like actually helping doctors. And a doctor can actually talk to it just the same way they talk to a scribe, um, to a human scribe, you know? The doctor could ask it, "Can you send this email?" You know, "Can you do it?" It does much more than, you know, an an average, um, tool that is there today in the market. Um, so yeah, this is the, we, so I think you perfectly articulated it and that's why actually we built it in an agent model or an agent-like model.

Jake Aaron Villarreal: So if I'm a doctor, I'm in a small facility, I have my own practice, or I'm in a large institution, it's helping me do a few things. I can assume it might increase my productivity because I have something automating the menial tasks that I have to do but I don't necessarily want to do and it can automate it. Um, I also have to trust that it's going to work because this is information that could change a life. Talk to me about the team that's building this and what's the trust in your company that doctors can look at and say, "Okay, if they've done X Y and Z, I think I can believe in this product or at least try it."

Ahmed Omar: Sure. So I think, I think the team is one of the, I think, the greatest assets that we have. Like, uh, Chaitanya Gharate, CTO and co-founder, built foundational technology that we use today like, um, Wi-Fi routing of Google Maps, you know, Google Assistant, the NLU and NLP part of Google Assistant. Managed, like, teams inside of Google, 30 plus teams inside of Google. And he built like three startups at Google. One of them actually used applied AI. Chaitanya has built products with computer vision 12 years ago, you know? So it's really deep experience in, um, uh, you know RNNs, like um, um, AI and machine learning algorithms, you know, all these technologies requiring neural networks. And this is basically one of the things that we did in the past, basically couple six months, is how can we make it, first secure, second HIPAA compliant, third very high performance, um, agent-like model.

So in terms of let's say the capability of the models that we built, we have also Sully. So Sully, basically, is um one of the people who built GPT-4. Um, one of the people who um actually was working with the co-founders of OpenAI, Ilya and Greg Brockman on, you know, the newest technologies of of of... built the capability evaluation of the GPT-4. Um, and he studied medical and he did it more than, you know, he studied Duke neurosurgery. Been researching medical, like been been an AI researcher for over 10 years, been a medical student for for a couple years. So it has this unique uh perspective of how can we build safe, you know, safe models for doctors to use. One thing that we do is we have internal eval trackers. On every time doctors close a note, we run an evaluation like metric across this note compared to the notes we generated, and we keep improving the models every single day. So there's actually, and the current trackers, the current metrics that we built is actually better than an average physician. So it exceeds the USMLE tests by over 85%. Which is higher than an average physician. So you can consider that any output that the model generates or Sully generates is actually reviewed by a better, by a better than average physician. So it's basically...

And there's a lot of things that we do that we require doctors to approve before we go ahead and do it or automate, which basically keeps um some kind of, you know, um, gateway or some kind of approval from doctors to, you know, get certain information done or get a certain automated action done. So we also keep doctors in the loop for some of the tasks that might affect people's lives or might basically deal with diagnosis, all these things. We actually don't sell any diagnosis, you know, AI agent. Although we do it, we have it, they can turn it on and off, but it's it's very key, basically, to us to see like what, what, how the, how our, how is our diagnosis getting better every single day um compared to, you know, the physicians that are working on it. And how can we actually, once it exceeds those physicians' performance, we can actually share it and give it, give access to the physicians and say, "Hey, this can help you um misdiagnose better or like, decrease the misdiagnosis and diagnose better."

Jake Aaron Villarreal: So just to reiterate, your founders, which by the way, thanks for sharing that. One came from Google...

Ahmed Omar: Yes.

Jake Aaron Villarreal: 15 years, something like that. Is that correct?

Ahmed Omar: Yes, 25 patents in AI and robotics, um, yeah. And he left to join your company.

Jake Aaron Villarreal: Yes. For the first time in 15 years.

Ahmed Omar: Yes.

Jake Aaron Villarreal: And your other founder helped build ChatGPT specifically for medical questions or just in general the functionality?

Ahmed Omar: The capability evaluation of the, of the model. So any output the model creates, he was building the capability evaluation of it. Um, like um, like yeah, um, the eval trackers, all these things that he actually built for us from OpenAI.

Jake Aaron Villarreal: I mean it's hard to argue why that team couldn't succeed in building something really cool. I love the the purpose of what you're working on. Um, tell me a little bit about, you know, you're an AI-focused application company and we had a chat a few weeks back and you mentioned Transformers and how they help your company or how, how they are supporting what you're building. Define what a Transformer is.

Ahmed Omar: So I think it's a very good question. I think Transformers are, I would say, recurrent neural networks that are working parallel, like in parallel to get to a certain output. And I think they, the way Transformers work, I would say, is very um, like it's a, it's a parallel sequence-to-sequence approach that improves the basic seq2seq, you know, recurrent neural networks breakthroughs that happened before. And it works in parallel to come up with a better output than just a normal machine language, um, recurrent neural network or a normal machine language sequence to sequence. Parallel in terms of they can propagate certain, I would say they can propagate certain inputs and encoding and and decoding or embeddings parallel to each other to come up with the best output. Um, best reasoning, you know, you know, best uh, all these, um, you know, all these um... best reasoning, best output, best performance. So I would say that's my definition of a transformer. Um, most of them are now decoders like ChatGPT, um, so it's only decoders. Um, the embeddings, basically, is simply, you know, the custom embeddings or the, the embeddings of any Transformers is basically giving... It's like giving a human brain a certain piece of information and classifying it and saying this piece of information is X but it also could be Y in other, you know, um, in another, you know, embedding, custom embedding. And and Transformers tend to perform really good with high-quality data rather than large quantity data because they have this aspect of, you know, evaluating which, you know, which input is, uh, which output is much, would perform better than another output. That's why it works in parallel, it works, works in multiple sequence, sequences, um to come up with the best output.

You know, for us, I think it, a lot of people debate like how, how LLMs work, um, or how they should work, but I think it's very similar to human brain specifically with approaches like meta-learning and, you know, you know, the meta-learning approach that said, okay, if we get a baby and teach them how to jump, you don't really need to teach them that they need to jump right, left, front, or back. So meta-learning is basically similar to giving a Transformer, um, how to jump or skip, and then they should be able to jump back or, you know, jump right and left. They tend to do really well with these, um, kinds of tasks because they have some reasoning, they can collect a lot of, you know, a lot of inputs at the same time, a lot of embeddings at the same time. So yeah, I would say that's my simple definition of a transformer.

Jake Aaron Villarreal: And how is it helping your platform to help doctors or help the functionality of it?

Ahmed Omar: Um, that's a very good question. So I think, so think of a transformer like a brain. Like, just to simplify it, it's a brain. And for the past 200 years, there's never been a brain that you could upload the most recent medical researches every single day regardless of the amount of information. So think of the gap between you as someone with chronic illness having certain pains, and there's every single day there's new medical researches about your illness and there's no way that the human can keep up with that. And think of this gap as, Transformers actually and LLMs could help actually bridge that gap which could change many, many lives in healthcare. Um, millions and if not billions of lives in healthcare. Um, so think of this gap that it was never possible to fill before, and now it is. Now we can bridge that gap. Now while you talk, we actually um show you that demo while you talk, you, the the model can actually propose some diagnosis. We have it for doctors now to turn it off and on. It's really amazing to see. You know, if they prescribed something on the call, the AI model can tell it, "Hey, hey, by the way, prescription you gave was a little bit off or could mislead another prescription that he gave two months ago to this patient." So all these gaps, it can now be filled with Transformers with the most recent data. Like daily data. Like minute basis or hour basis medical researches.

Jake Aaron Villarreal: Great. You know, um, as you're building your products, what have you learned from doctors that have told you about your solution or what is it that doctors like that they don't have today in the market? What are some things they they just don't like about technology? Because there's so much technology out there that's archaic, that's been in healthcare for years, it's been trying to be disrupted for a long, long time. And you know, we've got a new generation of doctors, you know, they expect to have modern technology, a modern UI, better functionality, mobile integration. Like what, what, what are they teaching you as you build your product?

Ahmed Omar: One of the most feedback that we get from doctors, other than metrics of course... So one of the doctors actually are, are using us more than 10 times per day now. So they're using us with every single visit. So they... and it's a new doctor. So it's someone that wasn't used to do that and now they started doing that every single day. So other than the feedback we get from doctors, I think one of the biggest feedback that we get is the user experience. I think, and I and I mean this with really, like, stressing on user experience. Anything they try to do in their current EMRs takes like 10, 15 clicks, 20 clicks, 50 clicks, and they basically have certain workflows that they don't want to change. And with let's say older generation doctors, being in the flow of their workflow was something that, like the first thing we started with, decreasing any amount of clicks, removing any signups, any paywalls, anything, was the first thing we wanted to do. In newer generation, we could see that there's no way for them they're going to use Epic or, you know, all these old systems. They're like, "No way, that's like really old." But if they have to work for an organization that is using this system, they need another way to practice or they need another tool that they could use to practice. So I think the biggest feedback that we see is "easy to use." It's really easy to use, it's really easy to use.

And and I think that part was really missing from healthcare. And the excuse from those, all those startups was, "Oh, the regulations" or "the HIPAA compliant" or or or... And and we were, we found ways beyond that. And I think that's one of the things that we like, pride ourselves to have, like user experience, user experience, user experience is number one thing. And then we see what kind of regulations we need to go through, what kind of you know hurdles we need to go through, but the first thing was user experience. Now you activate Sully by saying, "Hey Sully, can you send an email to the patient?" and it will do it. So, so less than one second, less than like three seconds voice prompt with no friction was something we wanted to do. So it will create a JSON file, build an email template, add the email of the patient and just wait for you to hit send, or you could just say "send it." So it's, it's really I think it's in the healthcare industry, the the user experience part was really lacking for the past years. We were shocked as well to see like why do they have to go through all these clicks to add a lab order or send a referral to another doctor or add a prescription? Um, it was it was really funny to see after all these years in the Bay Area we thought there's going to be like flying cars and robots everywhere but yeah.

Jake Aaron Villarreal: You know the the tactical side of it sounds simple. You have a platform, you integrate on top of technologies that are already there, it's better user experience, it helps take away administration time and you know give doctors an area to focus on things that are maybe more important. What's the, what's the end goal for the, the vision of the product? Where does it save a life? Where does it improve someone's health? Like where is the the story to you know, helping humanity?

Ahmed Omar: That's a very good question. So we started seeing like, by building the AI medical like assistant, we started seeing a very clear road to having a real-time diagnosis or having, having patients who diagnose real-time and then they can be sent to see an expert or they can initially get their labs done or lab tests sent to certain like LabCorp or you know, um um, or prescription sent to a pharmacy. So by decreasing there, there's a huge time-to-diagnosis gap. It's almost 22 days on average to diagnose a patient today in the US. And there's going to be a shortage of physicians, of 100,000 physicians in, by 2030. So the way they practice, the way healthcare is currently built or designed should completely change. So the only way to do this is to see where, where's the biggest gap. Now you see websites coming up and say, "We can help you see a physician in less than four weeks." So that, that's a marketing approach. So, and we think it should be real-time. So bridging this gap of getting a version, maybe an FDA version approved that could help diagnose certain pains or certain reasons for visits at least at the beginning. If you diagnose earlier, you can save you know the problems earlier, that you can basically have a treatment assessment plan earlier, you could save lives at the end. Um, if you have all the medical information of this patient, all the medication, all this data is in one place and in one model, you could give real-time feedback to doctors to say, "Hey, this is a risk. This is something that they could hurt, like this is uh food that they are allergic to," you know, in real time.

So doing all these small, you know, building these small um puzzles, pieces of puzzles will at the end save a lot of lives. Like today you have 500,000, 450,000 people... 46 people die every day because of misleading prescription or misdiagnosis. Um, there's also a reason for the death of late diagnosis as well. Removing this gap to getting real-time diagnosis that is FDA approved, that is fast, that is linked with the local infrastructure, you know, the pharmacy next to your home um, or or a lab next to your home, you know, that is linked in all in a single place can help, you know, save a lot of lives if we are able to get this data earlier. For example, something we do now, an abnormality coming into your EMR when an ab-, like let's say a lab result coming back. So what happens is doctors send, add the lab results on the EMR, lab orders on the EMR. The results come into the EMR. The results came in Friday night and there's one abnormality that if it's a certain, if it's passing a certain point, a doctor told the model that "Hey, if it's this, if the WBC is passing a certain point, please send patient to ER right away." So before Sully what happens is the nurse comes back on Monday, tries to reach the patient, probably doesn't see the the the the lab result coming in by Monday or Tuesday. Can basically reach out to the patient on Wednesday. Call, the patient does not answer, fine. Finally connect with the patient on you know Thursday and tell them "Hey, um you have to go see the ER." And that's what, Friday to Thursday, that's seven days. And imagine what could happen in these seven days. If Sully now is in your EMR can easily say, "Hey, it's Friday night. Once the lab results come in from the labs, it could say hey, go see the ER right away," you know? So at Friday night instead of spending all the six days to get to the patient. And that alone could save a lot of lives if it's a very serious U-point or an abnormality in the lab results. It's amazing. What's even more amazing is 100,000 physicians or lack thereof in the next 10, 15 years, um hopefully AI can help solve that problem.

Jake Aaron Villarreal: It sounds like, you know, you're starting with administration which will kind of gain more traction as it gets deeper into the process of of healthcare. Um, you started the company and you got funding, you've got a great team. Uh, um, who are some of the backers of the company? Who are some of the people that are funding the company?

Ahmed Omar: We have really good names like 500 Startups, um YC, Goodwater Capital, TOVC in in Menlo Park. Um, we have very good names. We have really good angels. We have, like, people like Haroon Mokhtarzada, sold his company for 1.27 billion to... sold Truebill to Rocket Money for 1.27 billion. We have Brendan Driscoll, sold sold his company to Spotify. Um, Sequoia actually scouted us for for this round. Um, so we do have amazing I think people on board in terms of angels. A few Saudi angels, healthcare Saudi angels, Act3. So angels that funded like, or basically own hospitals in Saudi and all these things. So I think, I think there's a lot of good names on board and I think yeah, and I think there's there's amazing basically people backing this company and there's amazing people that see vision of this company and how this could be big um in many, many ways. Um, so yeah, I think we, we have pretty good names um on board and I think this round too I think there's a few good names that are joining as well. We have a new round opened up um for our Seed Plus. So yeah, um I think I'm yeah, really proud of the team. The the team and the investment team as well. Like the investors as well. Really, uh, we really like everyone on board now.

Jake Aaron Villarreal: That's really great to hear. Sounds like there's a lot of upside of not just getting people on board, but you know, growth in the company. As a company today, how big are you?

Ahmed Omar: So we're about four. So we're three three co-founders, one founding researcher. And I think yeah, and we have a few like advisor slash, you know, um, partners. So our our first like, one of one of the investors was our first user actually, um, that helped build the product. Um, he's also, you know, building the product with us. Um, we have people from Kaiser actually joining on Friday. Like one of the CMIOs joining on Friday the team as a, you know, um, part of the team. So I would say currently we're about five to six people, five to six team members.

Jake Aaron Villarreal: That's great. You know, as you look in the future here, I'm sure you're going to be growing. What, what type of culture you trying to build? What would make it exciting for someone to join if they heard about you or you know maybe you're at Google and want to leave to a a company that's building a startup?

Ahmed Omar: Yeah, I think the I think one of the most things that I value in the company culture is there's no limits. Like there, there's there's no limit, like there's no there's no ceiling of things that could be possible. And that's one of the things that we made sure to maintain in the culture. Anyone saying "impossible" or something like that, we we don't have this in the culture. Um, we we don't have like uh... yeah, because we've done things previously in our lives that was completely impossible or kind of, and many, many people over and over told us it's impossible. And we still achieved it, sometimes broke the numbers that we provided to investors as well. So the culture is built that everyone, everyone is basically, has no ceiling, which is the gap of learning is extremely high. Everyone in the team is extremely exceptional. So for someone to pass through, you know, all of us, it's really hard. But if they did, it's like, um, a really big, you know, learning experience. They can, um... every single Saturday we run through research papers, see what we could actually do from research papers in our company and what could actually improve or solve some problems that we have today or challenges that we have as a company. So the company basically, the the people in the team, even we were, I'll I'll be honest with you, we were kind of shocked when people like Chaitanya working 15 years with with leadership at Google, um, Sully Chen working with uh, you know, Ilya and Sam Altman, coming to us and saying, "You guys move faster than like companies I worked at" or "I learned a lot from you guys." And I think it was just, you know, the culture of learning. And you know, nothing is impossible, if you want to try an experiment, you should try it. It shouldn't be a thought. And and the way we build things, you know, from putting the customers and the experience first and then working backwards from there. You know, um, taking like responsibility of anything. We we're all basically as well technical. So we all did coding, did engineering at some point of time in our lives. So we all understand each other's pains. Um, and yeah, we we trust, I think before getting someone on board, we trust, like we try to make sure we can trust them with what they do. So we basically give them the freedom to do certain things. Everyone in the company is self-managed. Everyone has a mission to achieve, either in LLM performance or in application, you know, um, speed, or in user experience. Everyone has, you know, a core direct responsibility, like a DRI. They're self-managed. Everyone is driven by the mission as well and the vision of bringing, like, you know, one, one doctor to every human on the planet. That's our ultimate vision. And and I think yeah, and I think that's the kind of culture we have. If someone is not D... someone is looking for a full-time job, um, or like a typical full-time job 9-to-5, I don't think that's, yeah, I don't think they, they'll fit in in Sully's culture.

Jake Aaron Villarreal: Sounds amazing. You're, you have an amazing journey here and I know you're kind of in the middle of it getting started, but actually have product in the market, you're getting revenue from... what's next? What's 2024 look like for you? What's on the horizon in the road map?

Ahmed Omar: Yeah, and 2024 is really exciting. First thing is removing hallucinations from LLMs for good. So I think in 2024 we want to reach to an enterprise level technology company where it can scale to hundreds of thousands of concurrent events a minute or an hour. So I think 2024 should be... so I think we're very clo- like we're early seed but we still are very close to product market fit. We can see very good early signs of, you know, 50% of the people are using it every single day, um, 80% of the people are using it every single week. So, um, we could, there's very good signs of product market fit. I think reaching, hitting product market fit, reaching profitability, default alive mode is the goal in 2024. Um, we're raising this round, we want to make it the last round that we raised, that we need to raise. Other than that, like, beyond this round, we shouldn't be raising that we need to raise. Um, it should be a default alive profitable company, at least close to profitability. Um, but it should be a default alive company. And that's the goal for 2024.

Jake Aaron Villarreal: That's great. What's the uh, market opportunity? I know as a company you go out, you raise funds, you got to provide some numbers and some opportunity to to scale and and show what that looks like. From your view, from your research, what's what, what's that look like?

Ahmed Omar: So now the research says $4 trillion, uh, 90% administrative workers. The billing is at a $13 trillion industry. So together it's like a $15 trillion industry. I believe it's more than $20 trillion. The reason is the possibilities that the AI will bring to this industry will open up even more ways of not only monetization but more solutions that will basically be core features to doctors, patients, healthcare providers, you know, insurance companies and any healthcare basically organization. And I think now with all this admin work and paperwork, it's a $15 trillion industry, or let's say worst case $4 trillion dollar industry. I think it's actually way more. Um, and it will be, if it's not way more today, it will be much, much more. And the reason why I think so is it's directly related to people's lives and it's basically a, a need. It's not a want. Healthcare, 8 billion people. And, and yeah, I think it's, I think it's not really, no one has built something that is as big as, as what we envisioned Sully could be. Um, and yeah, I think with, with, with all these technologies, I think it'll open up a lot of bigger markets as well and it will improve lives as well.

Jake Aaron Villarreal: Yeah, 100%. Yeah. You talked about a billion people. You're focused on the US. What's the opportunity internationally?

Ahmed Omar: So that's, that's I told you it's more than $2 trillion. And and and I think the, the, the problem worldwide basically it's hard to track some countries like Egypt, Africa, you know, um, very, very poor countries that have no access to a lot of other things. Most of the records are not even electronic, so it's really hard to track in other countries than the US. The global opportunity, I think it's more than 20 trillion. And that's why I said it's the whole global, the whole opportunity is more than $2 trillion. Either we can track countries like Africa, in Africa, or you know, um, countries in Asia. We can't track those. But they do need the product. Physicians actually, their physicians need the product more than the physicians here. Um, they have very few access to education as well. So that can bridge a bigger gap. And and yeah, we think we, we scaled the company before globally so in three plus countries, four countries. So we know how to scale companies beyond, you know, just one country. We know how, what it takes. We know, you know, what, how to structure this. We know how to build um, a basically a, a copy and paste model basically in any country we launch. So, um, so yeah, and now Transformers tend to do... Transformers were built for language translations first, right? And uh, they, they tend to do really well with different languages. So they actually understand like most of the Transformers now work with 26 plus languages. Um, some of them work with 90 plus languages and dialects. So the the barrier of you know, languages and translations and all these things is also really um, became really simple after all these innovations.

Jake Aaron Villarreal: Ahmed, if I'm a physician out there or a health care provider or a a hospital or a network of hospitals, they want to try your product, they want to know more about Sully, where do they go? How do they find you?

Ahmed Omar: They can email me anytime at ahmed@sully.ai. So A-H-M-E-D at sully.ai, or they can um, find us at sully.ai. They can schedule a demo with us. They can email us if they have any question questions. Uh, yeah, I think they, they can find me on LinkedIn. Um, and yeah, I'm really happy, I talk to customers every single day. I don't think I, I don't think anything makes my day than talking to users, either potential ones or current ones. Um, I go to the clinic maybe four times a week now to one of the client clinics, make sure everything is working fine and the new features are actually helping and it's not a nice-to-have features, it's a must-to-have features. Um, so yeah, I uh, they can find me basically anywhere. Yeah. Um, either email, LinkedIn, or um, use the platform, uh, the, the, the website.

Jake Aaron Villarreal: That's great. All right. Well, thanks so much for joining here. A big shout out to you for uh, having the courage to come on and share your story. As well as, I want to thank the listeners for listening. Means the world to me. Uh, my name is Jake Aaron Villarreal. Uh, I'm your host and look forward to catching up with you all on the next episode. Until then, take care.

[Music]

Before we wrap up, I want to give a big shout out to all the entrepreneurs that have joined to make this podcast possible, and for all the listeners for listening. It means the world to me that you chose to spend your time with us today. I'm your host Jake Aaron Villarreal, signing off for now. We can't wait to connect with you all soon on the next episode. Take care.  

This show is sponsored by Match Relevant, a company that helps venture-backed startups find the best people in the market and they do it in three simple steps. First, they sit down with founders to understand their story. Second, they tell their story into multiple candidate channels. And third, they schedule interviews within 48 hours. Find us at matchrelevant.com to learn more about how we do it.  

‍