Founder's Story - He Found $500B Hidden in Healthcare Waste — And Built the AI to Fix It | Ep 283 with Raheel Retiwalla Co-Founder and Chief Product Officer at Boost Health AI

Episode Date: November 18, 2025

In this episode, Daniel sits down with Raheel Retiwalla, Co-Founder & Chief Product Officer of Boost Health AI, the company unlocking the $500B in administrative waste trapped inside healthcare’s rules, guidelines, and policies. Raheel explains how Boost Health AI structures the complex medical rules buried in PDFs so payers and providers can finally access them consistently, accurately, and in real time. He shares the pivotal moment that convinced him this was the problem worth dedicating his life to—why timing with post-COVID financial strain and generative AI made this mission possible—and how Boost Health AI is rewiring healthcare operations rather than simply speeding them up. Key Discussion Points Raheel opens with the moment that shifted his career: a JAMA–McKinsey study revealing $500B in pure administrative waste—not from delivering care but from managing care. He breaks down how the root cause is shockingly simple: healthcare rules trapped inside PDFs, guidelines, and regulations, forcing humans to manually interpret them every time a decision is made. He explains how generative AI allowed Boost Health AI to extract, structure, and validate these rules at scale, giving payers and providers instant, consistent access to the policies that govern every decision. Raheel walks through why timing mattered: post-COVID financial pressure pushed the industry to seek efficiency, and gen AI arrived at exactly the right moment. Daniel dives into the deeper challenge: healthcare cannot use black-box AI. Raheel explains why Boost Health AI is built around transparency, citations, auditability, and an open model where payers own their intelligence instead of renting it from vendors. They discuss how unlocking medical policies speeds up authorizations, reduces friction, and creates room for automation across care delivery. The conversation expands into future impact—rewiring broken processes instead of just accelerating them, shifting from reactive to proactive care, and preparing the system for AI-powered disease detection, drug discovery, and long-term population health. Takeaways Listeners learn that the most transformative AI in healthcare won’t diagnose disease—it will fix the invisible machinery beneath it. Raheel shows how Boost Health AI turns chaotic rule interpretation into structured intelligence, unlocking billions in value and reducing the delays that harm patients. This episode reinforces the importance of explainable AI, operational domain mastery, and building technology that rewires industries rather than automating old problems. Closing Thoughts Raheel’s story shows that the biggest opportunities in innovation often come from problems no one sees. Boost Health AI is proving that healthcare’s future depends on clear rules, transparent infrastructure, and AI systems that empower—not replace—human decision-makers. His journey reminds founders to look beyond the obvious, solve inefficiencies at their root, and build with transparency, courage, and long-term vision. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

So, Rahil, I'm very excited about the future of AI when it comes to health and how technology is impacting the health sphere. I've heard things like AI could solve all of our biggest health problems or health concerns in the future. So how exciting is that? But I'm sure it's intricate and complex going from where we are now to where the future, if we never have a health problem and how long that could be away. So what was the moment in your life? Why was this something personal to you? And what was the moment that you realized how not only big this problem was, but a need so big that you are willing to dedicate your life to building boost health AI?

Yeah, Daniel, I mean, I would say that, you know, I've been in healthcare for over a decade, working with health plans and health systems, trying to figure out how to use digital better, how to improve operations. But it wasn't until in 2020, during COVID, actually a study was released by an organization called JAMA, along with McKinsey, that mentioned $500 billion of waste, administrative waste, just managing how health care runs, not the cost of actually providing health care. It's just managing the way health care runs is $500 billion. And what that article did is essentially articulated exactly what operational areas are driving that.

And it was COVID time. We were as a company thinking about, you know, healthcare from an AI perspective, even before generative AI at that time. And reading that just kind of stuck, that stuck in us. And we said, you know what? If there's nothing else we could do and contribute to the U.S. health care ecosystem, just increasing the efficiency, increasing productivity, and making sure that, you know, we can reduce this administrative burden, that would be huge for us. And we started our journey then. And it wasn't until generative AI became big that it just snowballed and allowed us to kind of

accelerate what we're doing and what we had as a vision at that time. Some of the most successful entrepreneurs have told us that timing was the absolute critical piece in the story for them. So it sounds like timing for you was massive. Like if Gen AI didn't become what it is, we might be having a different conversation today. So can you dive into detailed around what did Gen AI enable you to do within the company? Yeah, I would say, well, specifically to timing, first of all, right? There are timing, at least for us, included in addition to generative AI, it's great that you have a technology, but without a real problem,

it becomes just a technology, just another thing. So for us, another important thing that happened in this period is after COVID, just after COVID, you know, the cost of care became increased significantly.

And that even today has healthcare systems and health plans, essentially they've shaken up. Your financials are not the way they used to be. So what has happened is that the need to actually create efficiencies has multiplied and become more urgent. So that's just keep that in the backdrop because that is one of the timing parts.

And then generative AI as a solution, potential solution, came about. And what we did is we looked at the power of generative AI and said, where can we apply it? You know, there's so many different areas. And you're seeing and hearing about AI, as you said earlier, lots of different places in health care. But where are those opportunities when we think about the administrative waste, as I mentioned to you earlier? And what it turns out is that a lot of the decisions that are made in healthcare, so whether I should approve someone's authorization for physical therapy, how much therapy should they get, where should they get it, what should it cost?

All of those things are essentially rules. rules that are stuck inside PDFs in guidelines, in policies, in regulations, in benefit statements. So regardless how much investments, healthcare organizations have made to make things more digital and portals, this particular act of validating the rule by some human, somewhere, to say, should I do this? What should I do here? What policy applies here? It just is what drove that $500 billion or additional. administrative based in the U.S. So our goal was to use and think about application of generative AI

in unlocking those rules. What if we just unlock those rules and made them available to every workflow, every person involved in the healthcare continuum, so they have access to the rules in a manner they can use consistently, accurately, and, you know, do what they need to do faster. And that's kind of where generative AI became the sort of the technology platform for us to essentially validate first, test it, and then not make it available through boost help. I mean, I can't live without generative AI. I use it every five minutes of the day. I wasn't feeling good yesterday. And I was asking chat CBT, what should I do?

And so for 48 hours, I basically followed along. I don't know if that's a good thing or not, but everything it told me has been factually correct. So I know you have a bold idea of payers should own their intelligence instead of renting it from vendors. Why is this critical for the future of AI and healthcare? And what have you seen when this is capable? Yeah. The biggest reason for us is this idea of a black box. And it goes specifically to the point you made earlier about, you know, bias and, you know, just the trust in AI.

at the same time, there's a ton of innovation happening, right? So there's a new point solution to solve problem X, problem Y, problem A, with its own AI something. So if I'm a payer or a provider, and I'm starting to just kind of invest in these tools as they're coming along, I have no idea under the hood what is actually happening. And the big thing about healthcare, that it has to be explainable. Everything that happens in healthcare has to be able to say exactly where it got that from,

cite the facts. In our case, from a boost perspective, means exactly which policy statement, which policy criteria are you referring to? What document did it come from? What specific benefits, specific state clause, which regulation is limiting this? So you have to be able to explain that

and you have to cite the criteria and you have to audit exactly in situations where you're providing a response to somebody exactly what the situation was and how you responded, how the user actually benefited from that. Because we want the insights from, it's a partnership. It's an augmentation collaborative between AI assisting, whether it's a person making a decision on an authorizing something

or a person deciding on the benefits you can apply. All of those are decisions that people make. And AI is just there to help them accelerate it. So they're more confident in the decisions they're making, rather than replacing the human and making the decisions. So for us, it's how can then AI become more explainable and auditable and observable? That's the really important thing.

If you're buying point solutions, what happens is you can't guarantee you're going to get that level of explainability or observability. So for us, what we're trying to do is we're flipping that model. We're in fact giving away our boost health IP so that our clients can actually see the code and use them. it in other ways beyond the initial use case that we may work with them. This allows them to have control over their AI. And as technology, obviously, as we know, enhancing super fast right now, they can actually make inroads and additions and improve that, you know, with our help or

independently versus getting stuck and locked into something that they can't understand well. So are you, do you see the future of, let's say, healthcare and AI? Do you think a lot of companies, are going to be sharing abilities, having the ability to open their APIs or share, like, let's say you're working on some piece and then another AI is working on another piece. Do you think companies will be combining or organizations will be working together more in a collaborative versus competitive nature? Yeah. I, you know, first of all, just generally speaking, without even in AI right even if you don't take AI into consideration the need for interoperability and data sharing is just very important healthcare just generally speaking right because we all are part

of an ecosystem and a machine that works together so if one hand doesn't know what the left hand is doing it doesn't work which is where a lot of the inefficiencies to be you know generally are so the fact that we need to do that great but when I think about the application of AI today from a maturity standpoint, health plans and health systems have so much inefficiencies just in the way they do things today, themselves and how they run their business. It's like saying how you ran your business,

in the sense, how you run this podcast. If it was so inefficient, like, what's the point of sharing stuff with other people? You can't even get your own house running correctly, inefficiently first. So what I'm saying is that there's an opportunity to clean up stuff, make yourself more efficient, while there are regulations already in place demanding the level of interoperability.

So that's going to happen anyway. Let's start cleaning our house. So we are able to benefit from that interoperability and data sharing when it becomes a mandate, when people are seriously doing it. And it's happening and it's going to happen soon as well. Well, I think it will be in a better world, right? If interoperability and people are really sharing and I hope so. I don't know if we've been in a world where we've been so open, right,

with the ability to share. It's almost the gatekeeping of information, right, has become why I think some organizations became what they are. But maybe we're going to a world where we don't think that way. It's almost like more of an abundance mindset. What changes in these payers organizations when they start using the foundry and factory? And what do you hope that other organizations, who are not using it, what do you hope they know so you can really even scale this more?

Yeah, I think the fundamental vision that we share with our health plan clients is very simple. You've got rules that are blocked, drives inconsistencies and how people make decisions, which drives inefficiencies, costs, and that's what we're trying to reduce. So what we work with them on is identifying specifically the document types that are, you know, that are locking up those rules. medical policies is a really good example. You unlock a medical, what is in a medical policy?

Very simple. When will the plan, you know, pay for a procedure? When will it not? One of those rules that will, that somebody has to follow. One of the things that we do when you unlock a medical policy is you open up tons of opportunity for automation.

So for example, imagine if we were to be able to tell your doctor when they're submitting a, you know, a authorization for you to get an MRI with the right information first time versus what happens today is submit an authorization, somebody of the payer size reviews it for a week. If it is a complex scenario, then somebody more clinically inclined to doctor or whatever has to review that. It can take a cycle. If they don't have all the information, then it goes back to the doctor to submit again. And that cycle can take time

from someone who needs care, not getting care at that moment in time. And we want to avoid that. So Unlocking the medical policy allows the doctor to be able to validate your submission against the policy real time. So even before they're getting the information. Like, wait a minute, your document isn't, your submission isn't complete. Here's why. Provide this level of information. Go do this, go do the why. And it just reduces that cycle.

That's just one example of many, many unlock medical policies. And if you think about the other different document types, you kind of unlock value in a wide variety of ways. So for us, it seems really about talking to our health plan clients about that approach. Many are obviously working with us and thinking about this and working on early implementations of their workflows. And that's what you want to tell people is that, look, you know, you've spent a lot of money already in point applications, in digital transformation, focus on what has always been the problem, which are these rules. So it sounds like if you succeed at full scale that I think you had said there's $500 billion in inefficiencies or I don't know if I got that. Is that correct? Yes, that's right. That's right. And I would imagine there could be trillions of dollars when you add in

a lot of other inefficiencies or things that happen, not just in what you're looking at, but all different aspects of health care. What do you think this impact will have on the health care in the U.S. or maybe the world as a whole, just the whole ecosystem in the future, if these inefficiencies can be removed, can be solved, and potentially could be trillions of dollars, I'm thinking, reduced. What do you think will happen to health care for everyone in 10 years from now? Yeah, it's an amazing question, because, I mean, you know, the lens we have right now at the moment for boost health is focused on

payer efficiency, right? Even within the payer efficiency, our goal, and our goal generally as a whole, as everybody involved in AI and systems, is not to just do what is already being done just faster. It's to rewire things because it's, in many ways, the processes that were built were done because of how the situations were at those moments in time, in the past.

We're not shackled by those anymore. We have lots of capability, lots of ways to kind of think differently. So our goal is to essentially not say, well, here's exactly how you did it. Now AI is just going to make it, you know, little faster. No, what we want to do is you want to completely rethink the process, rewire that operation so that you're gaining 50% improvement, not 10% or 15%. That's where you really get into the value of AI. So if I think about, you know, administrative burden, even just on the payer side,

and if you start unlocking, just even taking the documents as a way to kind of unlock, value, it's in tens of billions of dollars or close to 100 plus billions of dollars aggregate across the pairs. It's a massive improvement. Now, when you add on the other side, you know, and the bigger thing is one of the places we want to get to is we want people not to get sick, you know, when we could have prevented it, being more proactive about working with people on their health. The problem is we are so reactive today in healthcare.

Most of our care managers, people who are working with people that need care, are doing so with the top 10% of the population that is highly sick or has multiple diseases. The rest 80% have no visibility. We want to direct the attention to those folks that are rising riskers that kind of we have no clue about until they become really sick. So it's those kinds of things that when you start thinking about what impact that could have, when people leading better, more improved health, healthy lives. And then you add what's happening on the health system side, how AI is changing the way disease are detected. Medicines are being created. The change and shift is, I can't imagine it could be completely honest. It's just, it just boggles my mind as much as yours, I'm sure. But I think there's meaningful steps that you can take along the way. Reduce your operational burden. Get ready for

that change that can come and help, you know, that you can really become more proactive versus reactive. And that's what our foundry does. It has these components. You can start plugging into these rules and into your workflows and start making those meaningful impact today. It's, you don't want to get into analysis paralysis. There's a lot we can do today. And we try to get our clients get moving versus getting stuck in like, where can I go?

Where should I start? We had a guest on a few years ago. They had raised a couple hundred million dollars to build out these preventative centers that were basically no humans. unfortunately they went out of business. But I think the idea that they had was great around this huge thing of preventative care and using a lot of wearables and all these devices to detect things in advance and then having that data then spread to doctors and stuff.

But I think they're maybe ahead of their time. But the things they told me, though, I was like, wow, you know what? I wonder how much further could we live or how much further can we live healthy versus, you know, living to 100, 130, but I'm not healthy because I wouldn't even want to live at that point, right? But I'm excited, like you said, of how many diseases can be eliminated, how many medicines can be created in fast times. But for other entrepreneurs like yourself or other entrepreneurs that want to be like you in the sense that they want to create something that's going to change the world, they want to impact something. They're going into a space that probably is not so up to speed with technology, and they have a huge problem to solve.

What advice do you give to them to really get started? Yeah, I mean, I think the biggest advice is get to know a domain really well, like a domain, an area, whether it's an industry and a specific process in the industry. I think there's, you know, as we get started with application of AI, the way we're thinking about it right now, it's about how. how people do what they do and how do they run their businesses, right? If you start, you know, if you start focusing on those inefficiencies, you're going to make businesses run better,

which just allows them more capital, more abilities to invest in different ways, different things in enhancing your own products and services or reach to market. It just changes the game for them to be able to unshackle from just being able to be constricted by the way they're running things. So that's what I would say. I mean, you know, I have kids who are in high school. I tell them even now, I said, look, you know, learn industry, understand processes, think about, you know, why is this happening the way it is? And then think about if I had AI assisting me here, what could it take from my plate? Where could it help me do more the mundane things that I can then focus on more value things and what those things could be.

So I think that's what I would say. And if you look at every industry that's going through in AI Renaissance, you know, many of those use cases are very much in those areas. Is that call center, why is call center automation call center is so big? It's just a massive cost. It's just a massive cost. Right. So let's improve that, you know. So. Well, I've been I've been vibe coding. Right. I've been making apps. I'm not. doing anything with them because I'm too I'm too much of an idea guy so I'm like oh I'm going to solve this problem let me make this app and then the next day I'm like I'm going to solve this problem let me make this app it's it's almost scary so I've had to cut myself off from making

apps but it it does make me excited that people it's not like before I think you had to raise 20 million 50 million 100 million dollars right you could literally start something tomorrow create like an idea get it like workable And then you could start pitching it, start gaining some traction, which is, I think, incredible. Like, you could have people that are on an island somewhere that have internet access that can do something that might be able to solve a problem globally. So I'm very excited for that.

But thank you so much for joining us today. If people want to get in touch with you, they want to find out more information. They need to get involved because this is going to help their organization. How can they do so? Yeah. reach us at BoostHealth AI, BoostHealth.A.I. That's our website. As a contact us page, it will lead you directly to me and my team.

Oh, Rahil, this has been great. I'm very excited. If I'm still doing this show in five years from now, and you're a billionaire in five years from now, come back on and let's talk about how the next five years is going to be from that point. This is amazing. Boost Health AI. I hope that you solve the AI, the healthcare crisis,

that many, many people are in, and I hope that this will help those people. Maybe it'll help affordable health care. Like, I think it could do a lot of things with the amount of money that will be saved and the efficiencies and the time and everything. So I foresee the impact is going to be massive. But thank you so much for all that you do in joining us today and Founder's Story. Thanks, Daniel. Thanks for having me. It's a little pleasure.