In this episode of Health Biz Talk, host Tony Trenkle speaks with Susheel Ladwa, CEO of Onyx Health, about his journey from technologist and coder at IBM, HP, and Wipro to leading a healthcare interoperability startup. They explore how FHIR standards are transforming data exchange, why prior authorization is at a tipping point, and how AI is reshaping healthcare — from risk adjustment to digital HEDIS. Susheel also shares his leadership philosophy of putting the team first, moving at speed, and challenging the status quo before change becomes unavoidable. A wide-ranging conversation on what’s next for connected healthcare.
00:00 Intro
Welcome to Health Biz Talk, the industry’s leading podcast that brings you today’s top innovators and leading voices in healthcare technology, business and policy. And here’s your host, Tony Trenkle, former CMS, CIO and health IT industry leader.
00:00:11 Tony
Hi. I wanted to introduce Susheel Ladwa. Susheel is the CEO of Onyx Health, a leading healthcare interoperability company focused on Firebase solutions for payers, providers and government agencies. With leadership experience spanning IBM, hp, EDS and Onyx, Susheel brings more than two decades of expertise in healthcare technology, digital transformation and interoperability. Susheel was recognised as CxO Magazine’s most strategic CEO to follow in 2025 and Inc Magazine’s 10 most inspiring CEOs to follow in 2024. In he’s a sought after speaker on healthcare and operability, AI and healthcare, CMS regulations and the future of connected healthcare. So Susheel, welcome to the podcast.
00:01:06 Susheel
Thank you so much, Tony. It’s a pleasure to be on this podcast with you and truly appreciate the opportunity. I’m really looking forward to this conversation.
00:01:16 Tony
Thanks Susheel. Well, we have a lot to talk about and of course Susheel and I have been friends and colleagues for many years, so I know a lot about his successful career and it’s a pleasure to have you on the podcast today. So let’s kind of start as I always try to do. Susheel, let’s start off at the beginning. You’ve had a long and successful career. As we mentioned a few moments ago. You’ve been a lot of different companies, starting with HP and wipro. You were leading healthcare and life sciences at IBM and then became the CEO of Onyx. So can you tell us a bit about your journey and what were some of the major events that influenced you during that journey?
00:01:53 Susheel
Yeah, absolutely Tony. At heart, I’m a technologist. I’ve seen technology right from MS DOCS to where we are within the AI world. And oh boy, it has been one amazing ride for every technologist that has been through this journey. And as you go through this journey, it only gets exciting from here on. But just in terms of my experience, I started out more as a technologist that wrote code. So I’m one of those people who wrote a lot of code, Java, Python, Perl and whatnot.
So that gives you a real deep, meaningful understanding of the art of possible with software. And that married with the industry knowledge that I’ve gained over a period of time through some of the experiences that you talked about, be it hp, eds, where I worked, a lot of Medicaid opportunities. IBM was a real pivotal moment for me. IBM was where I think I really understood through the role as to how three streams come together. Technology, industry and finances. Really everybody, most of them understand either technology or the industry or a combination of both.
But adding the financial angle as to why you do what you do and how you produce revenue for the organisation changes the whole dynamic. And that was the pivotal knowledge that I think IBM offered to me. And using that, I was out in Silicon Valley, joined a startup, exited out of that company and for the past five and a half years have been on this Onyx journey and it has been an amazing journey. And Tony, I have to thank you for that because you are the one who introduced me to some of the key folks that eventually become Onyx. So you have a major role to play in the success that we have had so far. So always grateful for that.
00:04:04 Tony
Well, thank you Susheel. And I guess my question, and I know I did help convince you to come to Onyx, but why did you decide to go to Onyx? Because most of your career we talked about IBM, hp, Wipro, I mean these are all big companies and most your career was with big companies. Why did you decide to kind of chaos that aside and go work for a startup like Onyx?
00:04:28 Susheel
Well, in a large company, what I have seen is you are just a cog in the wheel and the processes are all set. There is a lot of large opportunities that come your way. It is to a certain extent a very routine life. You can expect what to expect every day. But once I left the cushy job of IBM and got into the startup world, oh my God, that was a complete roller coaster of a ride. Because there are days where you wake up and you are like, today I can’t make payroll, it’s done, the company’s over and by evening you would have closed a deal and you are on the moon jumping around saying, oh my God, I’ll go to Mars, will conquer moon and so on. So the highs and lows of a startup is definitely not for everybody.
But as I got to know the highs and lows of a startup and how to build a startup, how to build a core team, I think the future is all about small teams working really fast at the speed of light, using AI and that’s going to change the paradigm of how industries are going to shape and how large organisations can be disrupted by very small teams. So that’s what motivated me to go towards the startup ladder because I could already see that large companies like the ones that I worked with, they’re very difficult to change and bring up a great idea, which you know is an amazing idea. It just so happens that you have had a great meeting and come out. That’s it. Nothing happens with the idea because it is a consensus.
You need a lot of people’s buy in before you can move on an idea. But in a startup, you have an idea, you execute it by end of the night. That’s what is exciting and keeps you alive. I love the energy of a startup. So I couldn’t have. I wish I could have done this 20 years before, but nonetheless, I don’t regret it. I started about eight, 10 years now and I think I probably will never go back to a corporate world again given the experience that I’ve had with this.
00:06:55 Tony
Right, right. It’s interesting. I see the analogy that you talked about with speed and flexibility looking at warfare today and how these countries with drones are holding their own against much larger countries that have big expensive weapon systems. And it seems like in some ways you’re using a similar analogy about small companies being able to move faster because of AI and some of the other tools versus some of the larger behemoths that have been around for many years.
00:07:25 Susheel
Absolutely, Tony. What I feel is given where tools like Claude and even Codex and a couple of others are, Gemini and so on, I don’t think the industry is yet ready to adopt them to their fullest as to what they can offer. Case in point, the entire marketing department today can be run on AI. You don’t need a videographer, you don’t need a blog writer, you don’t need a social media manager, you don’t need any of those. Those are absolutely automatable today with the technology that we have. But the adoption of how to do it is where the learning is. And similarly, if you look at sales, look at code development, look at product development, look at testing each one of them, I think the tools are already there. It’s as to how fast you can adopt and that adoption is usually possible with a small team. And if you have a large team, there will be a lot of meetings, very little adoption.
00:08:34 Tony
Strikes me also that it brings the technology closer to the end user. Traditionally, technology and technologists were segmented to some extent from the business side of the house. But now you’re giving the business side of the house the technology and the tools to be able to do things that they had to rely on people from the technology side to do before. So that would be, to me, another interesting paradigm when you have people who are experts in the certain areas and now they can basically tailor the technology to their own uses for even greater insights and greater ability to make business decisions.
00:09:17 Susheel
I would agree with you. The biggest change that I see in the software and the business world when they collide is there was a jargonism that came up. It was like, oh, the sprint, oh, the regression test, oh, the performance test. This is like deep in the weeds, this is a loop that doesn’t work, and so on. I think that jargonism is out because code has been simplified, code has been broken down. I mean, pretty much it is going to become a utility. Anybody can code just by talking about they want to build. Once it becomes a utility, I think the jargons are out and that takes away unnecessary complexity in software. Once that complexity is gone, I’m hoping that apps and everything are going to be like any of our appliances, wherein you just plug it in your 110 old plug and it’s ready to go. So that’s, of course, the vision of the future. But to make that happen, I think the start is now. And the companies that can really make that happen are the ones that are small, fast and very nimble to learn and fail and learn and fail and quickly become successful.
00:10:39 Tony
So one of the areas in the healthcare space that ties back to what we were just talking about is the whole area of interoperability. So I’m going to spend the next couple minutes asking you some questions about that, which is, of course, how Onyx made its name was in interoperability. But interoperability is not new. We’ve been talking about interoperability for as long as I’ve been in healthcare, which, you know, is many years now. So what do you think’s changing now? And why do you think we can achieve better interoperability today? And I guess tied to that is what do you. How do you define true interoperability today? Because the definition of interoperability in 2010 or even 2020 is probably a bit different than it is today because we have so many ways to do things differently and so much access to sources of data that we didn’t have, at least not in real time, even a few years ago.
00:11:37 Susheel
No, you’re very right, Tony. The way at least I look at the evolution of healthcare is 1.0 was everything paper, 2.0, take paper and everything else, make it electronic. You know it. There are so many formats for a patient. You say, I Need patient data. You can get the patient data from the Same provider in 10 different formats today. So while the whole world was focused on solving the interoperability problem, I think what really was missing there was the standard. Now if you look at again, I wouldn’t say it was missing, it was not.
00:12:20 Tony
I was going to say there are plenty of standards out there. There was never a lack of standards.
00:12:24 Susheel
Yeah, you’re right. So I take that back. It was never lack of standards, but a maturity of the standards. So think about it this way. When we go to Expedia.com or any of these travel websites, we just go search for a flight and in behind the scenes it’s connecting to all the airlines APIs and getting all that data and crunching it all and giving you a result saying, here are all your flight options. If you look at healthcare, I think to get to that stage, there were a lot of things that had to be done in the back of the scene which were not ready. Number one, you need a standard format that everybody agrees to and says, yes, this is the format. I think with HL7, V1, V2 and so on. Of course, V3 was only for scientists and the guys who wrote V3 agreed to that.
I think those were not really scalable and they were not designed for the new way of life which the entire world adopts today. The jsons, where we easily transmit, interpret data and talk to data and so on. The FHIR standards are the new standards that actually adhere to all the things that we are used to on the Internet today. So APIs and the standard data formats. So that’s just one stream of thing that is much more mature compared to everything else that the interoperability industry within healthcare has seen before. That’s number one segment, number one, that maturity. Now the second level of maturity is the adoption. The adoption has always been a question because the adoption is very use case centric in healthcare and use case centric adoption does not lead to broad level adoption.
Today I think what CMS has is, which is a great vision I think is they want to drive use case based adoption with a larger picture in mind, wherein health plans and others who just adopt the use case think they’re adopting a use case, but in the back end they’ve actually built a whole infrastructure for it. Let me qualify as to what that is. Today, if you look at 9115, which is Patient Access API. For Patient Access API the requirement is 7 years worth of claims, clinical formulary and provider directory data to be converted into FHIR standards and on a nightly basis, the delta of the claims and clinical has to be again updated. That has to be opened up as an API so that apps can connect to it and be able to read it. Now everybody looks at that app part of it and says, oh my God, we spent all this money, but nobody is connecting using the apps. It was just a waste of effort. The way we look at it is this is the first time in the history of health plans you have seven years worth of clean FHIR enabled claims and clinical data sitting there and it is being updated on a nightly basis. Imagine what you can do with that data. Half of the programmes that you run for risk adjustment, for payment integrity, for hedis, reporting, care gaps, whatever, they all rely on this data and this data itself is a project. But here you have been handed that project, it’s done. You start from there and then build on top of it.
And if you add prior auth and so on to this, I think it really, for the first time becomes real in healthcare. Can already see that happening. Real with many of our clients, though, the adoption of interoperability as a rule, as defined by CMS has been a little lower. The data behind it is the powerhouse and whoever is able to tap that data and be able to do those other use cases, they would have such a bleeding edge over everybody else. They absolutely cannot be even in competition. I mean, they will be way beyond competition if they are able to do it. Short answer, I would say there is a timing aspect to it. The technology is great, but you need the timing. I think timing is there, the technology is there and the will to adopt is there. So when all of those three intersect, I think magic’s going to happen. So the next five, seven years are going to be really magical for the interoperability industry.
00:17:22 Tony
Well, particularly if CMS really puts its weight behind it, because as you know, CMS has the payment rules and they can do a lot of things with those payment rules to kind of force behaviour changes. I mean, they’re reluctant to use it too much because then industry gets upset. But I think if they can continue to prod industry along, then we can reach that nirvana and maybe one day I’ll only have one portal that I need to go into for my health care and now I go, I have like eight portals. So if I can get to the point where I have one portal, I can say, Susheel, they’ve solved the interoperability problem once and for all.
00:17:59 Susheel
Well, Tony, I think to my earlier point, maybe, you know, the eight portals, maybe five or six of them are already API and I enabled, I would imagine given the maturity of the industry. So there should be apps out there which can connect to some of them. So to your point, they may not be able to solve the eight problem, but they may be able to say eight to four is a great start. 50%.
00:18:29 Tony
Well, it’s something we’ve been talking about for a long time and even Microsoft and Google in the past have tried to set this, you know, this kind of super portal up for people with their health care and booths failed miserably. But of course that was the pre fire days. Now we have fire. We have an industry that hopefully is more engaged in looking that way. And certainly I think CMS is pretty consistent in both the Democratic and Republican administrations with pushing itself that can continue. And Congress as well has also been bipartisan, so we can hope.
But one of the areas I think that can make a big difference is prior authorization. And prior authorization is kind of the boogeyman in the healthcare industry. Nobody likes prior authorization. It’s been around for a long time and it’s certainly created problems for providers and patients over the years. Payers in many cases feel like it’s something necessary to create better costs and better outcomes. But now with CMS’s latest regulation, it looks like we may be changing with prior auth. And so what do you think if we were talking about prior auth a couple years from now after the rule was fully implemented, how would it change in. In ways you think would be beneficial for all the players?
00:19:52 Susheel
So you’re, you’re right. Prior auth is one of the biggest time sucks and the burden producers in the industry, both for payers as well as the providers. Right. And especially it impacts the patients because they are stuck between this process. Now if you look at it from a provider standpoint, the biggest challenge for them is if they need to just input a prior auth, they need to go to a payer portal, they need to update everything and they need to submit or the good old way you will e fax them. And once that data is sent, chances are if you are a gold carded provider, you don’t have to have prior auths on certain things but majority of them will still have to wait for a prior auth.
And when we looked at the prior auth data that the health plans had, what we saw was almost 40 to 50% of the prior auths that were coming in from the providers were just in case prior odds they did not need prior auth they were just put in just in case so that, hey, who knows, this might need a prior auth. So there is a lot of waste because of that. And secondly, I think the burden that is there on the providers as well as the payers because now, because of those additional prior auths, payers will have to staff up to look at all the prior auths and approve, deny and so on. Well, there are standard rule sets that each one of them have.
I think the big change that is coming out as part of the CMS Interop 0057 regulation is the fact that there is a new way to submit a prior auth and that new way is directly through the EMR workflow of the physician and click a button, they will be able to call an API call from the payer and get an instant result saying if prior auth is required or not and can eliminate 30, 40% of the prior auths. Imagine the amount of bandwidth that frees up for the physicians and the physician practises. But for the ones that do need a prior auth, there is a questionnaire that comes back to them and which they can fill out and it becomes much faster. And now the Turnaround time is 72 hours for an urgent prior auth, seven days for a normal prior auth. So those are stringent conditions that the CMS is laid. Well, what I feel is the prior auth industry is at a tipping point. The tipping point is once we go live on Prior authorization using APIs and the Provider uses it once, they will never go back. The simple fact that they didn’t have to spend any time on it and the overhead is gone,
I think the adoption is going to increase drastically. So getting it right is going to be the difficult challenge for the industry. I don’t see an issue with provider adoption because the moment providers see there is a new way, there is a resistance to adopt. But once they click a button and say that was it, that was it, after that they’ll never go back to the other ones. But now all the prior auth systems have become very complicated in the backend, right? So every UM vendor has their own rule set, has their own way of getting data and whatnot, and they are not all FHIR enabled today, right? So for FHIR enabled prior auth to happen, it’s not only that the payer systems are FHIR AUTH enabled, but also the UM systems need to be FHIR enabled.
So that’s going to be a long tail in the long pole in the tent of prior auth. But I feel like majority of the UM vendors are really making good progress there and we are testing connectivity with all these UM vendors on a daily basis. So we feel like towards end of 2027 is when we will actually see majority of the prior auth vendors being fhir enabled and fully functional. But there are definitely leaders already that are FHIR enabled and whom we are connected to. But I’m very bullish about this 1-1-2027 rule for prior authorization. I strongly feel that rule is really going to move the needle on prior authorization and it’s absolutely going to change that industry.
00:24:35 Tony
I think the, the rule certainly will move industry in the right direction. I think it’s also got to be followed up by making sure that CMS is, you know, exercising oversight to make sure that the rule is being followed. I think there’s also a rule, a role for the vendors and for organisations like Onyx, because you’re on both sides of the equation. You’re tied in with the vendors, you’re tied in with the payers, you’re tied in with providers, you’re tied in with the government. So you’re really on all sides. So how do you see what you all can do to play a leadership role here?
00:25:16 Susheel
So firstly, I think as Onyx number one, we need to make sure that the clients that we work with, majority, that is the payers, the first part of the problem is solved for them right away, which is 30 to 40% of prior auths that are placed just in case because there is an immediate ROI to show for the health plan saying that just doing this, look, you reduced your overall prior auth volume by 30%, 20%, 40% in some cases. That’s like magic, right? And that magic is not only on the payer side but also on the provider side because providers don’t have to put those prior auths anymore. That itself is going to propel the next wave of adoption.
So what we are really focusing on is immediate when let’s do crd, which is coverage requirements discovery, where it is the first API which essentially says is a prior auth required or not, then let’s tackle the other APIs as and when UM vendors are ready. If two UM vendors are ready, let’s go with two UM vendors that are fhir ready and then keep going. So the integration and the implementation approach is focus on a low hanging fruit, show the value and that value will drive the next level of adoption and the next level of adoption and so on. And from a provider standpoint, education is going to Be very important Tony because the industry has been doing prior authorization in a specific way for the past 40 years, for 30 years, whatever it is. It is very hard to go to somebody and say you have to change. They’ll be like why? I’m comfortable doing this, I’m comfortable sending e fax. I’m never going to try anything different. So to give them enough training, enough training material to show that there is a new way of doing things and that’s going to save you a lot of time, effort and money. Be the big, big bump that we will have to get through. But if we get through it once, like I said, it’s a tipping point. There is no coming back from that.
00:27:37 Tony
And I think what it’ll force these companies to do, whether the payers or providers, is to really start thinking about their whole workflow and thinking about how they can really change to meet the not only the requirements of the role but also from a business standpoint. This now gives them opportunities they didn’t have before.
00:28:00 Susheel
Clearly, like I said, interoperability with prior auth is just the beginning. What we are seeing some of our clients nowadays do is the whole chart chasing industry as you know it, right? You need data for risk adjustment, you need data for digital. I mean HEDIS to get that clinical data itself has become an industry where there are PDFs still being and CDCDs being sent and the same data being asked three times in a in different formats and so on. FHIR interoperability has a big role to play there. What happened was there is a rule called G10 that went live with all the EMRs 2023 Jan with that rule essentially 2024 I think Jan what essentially happened was all EMRs are required to provide a FHIR interface for the data and the payers or anybody who is authorised to have that data should be able to connect to the FHIR APIs and get that data back.
So think about it, what’s happening in the industry now? You have seven years worth of claims clinical data coming in from the payers and sitting in the payer fhir repository for CMS 9115 and now you have the ability to directly go give a roster of patients and get FHIR data from providers and enrich this data. And once you enrich this data and create a longitudinal health record, that’s the magic. Once you have the longitudinal health record that is constantly updated you can use that record for hedis, you can use it for risk adjustment, you can use it for payment integrity, you can use it for so many different care gap analysis. That’s where the secret sauce is. And that’s the magic that we are working on as an organisation right now.
00:30:09 Tony
And I think industry is certainly the leaders are understanding that, but others would. As you know, there’s a lot of risk adverse companies, they’ve made a lot of money by being risk adverse over the years and now they’re saying, well, we’ve done the interoperability, we had the patient access, we’re not seeing the usage. So why, why are we not seeing any kind of advancements here? And now you’re asking us to do a much larger thing with prior auth. And so what do you think is the reason why the patient access did not succeed the way it was touted, you know, five years ago?
00:30:50 Susheel
I think, I think patient as the centre of the universe in healthcare in the US ecosystem has a problem.
00:31:01 Tony
All right.
00:31:02 Susheel
If you look at any other global healthcare system that is out there, the patient is really the epicentre of everything because the patient pays for it.
00:31:12 Tony
Right.
00:31:13 Susheel
And once a consumer pays for something, every everything around them becomes accountable for it. But US is the only system where a patient can go to a doctor not knowing what they are going to pay till 30 days, 60 days, 90 days, whatever it is. Right. So I think it’s not that the patient access API did not work or the adoption of that did not. I mean, obviously the adoption did not happen. The issue is with the ecosystem where patients access is not considered as critical by the patient because they are not financially liable for it.
00:31:58 Tony
So have you seen the FHIR API for patient access work in other countries?
00:32:03 Susheel
Well, let’s take examples of say Europe or even India and others. Right, Europe. I think if you look at some of the FHIR providers that come from Europe and many of those servers come from Europe. I think there is a lot more accountability and a lot more consumer centric use cases there that I have seen. So is the case with Canada Health Infoway in Canada, certain provinces need that data to be shared because the patient, as they go to the doctor, will have to have that data with them, otherwise they will not be seen. So those forcing functions help. But have we seen that work? I think it works much better outside of the US than in the US currently. Primarily because the financial burden of that lies with the patient and not with the health plan or the ecosystem the way it is in the us I.
00:33:04 Tony
Think also because we have much more fragmented health care system in the US as well, and many different payers, many different providers, some areas Government is heavily engaged. Some areas, government is still engaged but not as heavy. Whereas it seems to me a lot of these other countries it’s much more uniform the way they do things and that makes it harder to adopt certain types of regulations and see the success in them because the patient is not engaged, because they feel like they don’t, they aren’t the ones in control.
00:33:41 Susheel
That’s, that’s very true and I think, I think that’s going to change slowly over here. As you can see, plans are also changing. Right, so now you have HSA accounts where you can pay a lot more. So slowly the burden is being shifted over to the patients. I think the more burden that gets moved to the patients, the more healthcare system becomes accountable for it. Because it’s insane that you go have a small procedure like an ACL surgery or something small and the charges are like six times what you would have paid in another country. So level of accountability is just not there financially here. Unfortunate that we are in a place where we are from an ecosystem standpoint, but I think we as us should stop out of smaller experiments and just there needs to be a broad level push from the, from the government saying here is a new way of doing things wherein the consumer is much more responsible for paying for it than the payers. I think that’s going to change it. It’s a very hard thing to do, but I feel like we are on that path out here given all the regulation and everything else that’s happening.
00:35:08 Tony
Well, I think it’s still very fragmented and I don’t know if we’re on the right path or not, but we can agree to disagree on that. But anyhow, let’s talk about one of the saviours that’s supposed to help us in healthcare, which is AI. And I know you’ve been very engaged with AI for quite a while, even back to the old IBM days with Watson health and, and related activities. So what do you think is the current and future use of AI in the healthcare industry and where is it most valuable now and where do you think about the five to 10 years? Because you know, what’s the old low hanging fruit versus solving the more complex problems. But it seems to me from what I’m seeing is that a lot of it is a lot of what we talked about with IBM 10 years ago, but it’s just in a different stage today than it was then. Obviously there’s, there’s greater tools than there were then, but still it, it strikes me that the business problems that AI is looking to be used for still are pretty similar in a lot of ways.
00:36:19 Susheel
Well, you’re right Tony. So the AI is the buzzword right now. Everybody is like AI is going to solve everything, including world hunger. It potentially can. But specific to healthcare, I feel like there are two kinds of AI companies. There is a AI native company and an AI adaptive company. Look at the AI native companies, these are going to be very small organisations that don’t know anything. But AI as the first way to do things and then grow. Case in point, if you, I don’t know if you heard about this company called medivh, started by two brothers, $400 million in revenue, so just two of them and they built a massive healthcare company. Same with Harrison AI, there are quite a few examples there. But there is an AI first way of disrupting a large organisation or a large segment and starting something new and doing it. So I feel like you will see a lot of disruption there in products that AI uses today.
00:37:30 Tony
Right.
00:37:31 Susheel
Companies use today because product is easy to disrupt. Something that would take a year to build can be built overnight today with cloud, with the right prompting. So there are AI native companies that you should always be looking out for that can disrupt things. But largely healthcare is not an AI driven organisation. They are usually the laggards when it comes to adoption of technology. So the larger AI play in health plans, in providers and PBMs is going to be very use case centric unfortunately and it’s going to have a cycle where they adopt it. Also what I have seen practically on the ground is when we sign contracts with clients, there will always be an AI clause that says you cannot use our data to train any of your and whatnot. Right. So for rightfully so because of phi reasons, because of so many other reasons.
So given the regulatory nature of it, I think AI adoption within healthcare is going to be a little slower compared to every other industry. That’s the reality of every other technology that has been adopted by healthcare. Now where exactly can AI play a role? I feel like where AI can really play a role is data insights. There is a lot of data out there. It is not about that. It needs a lot of human intelligence. But case in point, if you want to do digital hedis, I think digital hedis is something that can be completely done by AI because it has the ability to look around data and do something with it. Risk adjustment is already heavily AI driven, so there are pockets where AI is going to mature. But I feel like the AI that kind of matures and that gets really used in the healthcare ecosystem will be developed by non by AI native companies that come up with products that get.
00:39:41 Tony
Adopted here, which is kind of similar to what we saw in the digital world. The companies that went digital first were the ones who eventually ended up being the most successful in many cases. I’m thinking of the Amazons, the Googles and some of the other ones. So I think the same thing here, the ones who can figure out how to use the technology the best are going to be the ones who can leverage it to solve a lot of these business problems and clinical problems for that matter.
00:40:10 Susheel
Absolutely. So I highly doubt that AI adoption within a large healthcare organisation is going to be something that AI is developed by the organisation, maybe a handful of them can do it, but majority it’s going to be the AI native products that have completely disrupted old way of doing things that will come in and say here is a new way to do things and they can show results in two weeks.
00:40:40 Tony
Yeah, right. It just seems things like risk adjustment, I mean, or even something like coding that’s been a very manual process for a long time. I mean it just seems like something like that’s made for AI to just come in and take it over.
00:40:55 Susheel
Absolutely, absolutely. But that AI infusion really needs to happen as products that healthcare organisations can buy with confidence.
00:41:06 Tony
Right.
00:41:07 Susheel
Where the AI adoption I think is real for healthcare.
00:41:11 Tony
So the other challenge with AI is the cost and the tokens and the adding up of. We’re starting to see a lot of businesses are starting to get very concerned about that impact and I could easily see in the healthcare space that could really add up pretty quickly. What’s your thoughts about that?
00:41:33 Susheel
You know, the measure of AI effectiveness as a part of how many tokens you use has gained a lot of attention lately.
00:41:45 Tony
Right.
00:41:46 Susheel
I think the industry is slowly coming to a consensus saying consumptions of tokens is not the right way to measure the AI because some of them have learned it the hard way, wherein they forgot to switch off a credit card or a P card, corporate card and suddenly they have like $500,000 of bill.
00:42:07 Tony
Exactly. Yeah. It’s a lot like cloud, only worse.
00:42:12 Susheel
Yeah, it is worse. It is worse and it has a mind of its own so it can whatever it wants. So my, my view is, I think AI is, AI is currently expensive, it’s only going to get cheaper. But the way you adopt AI, the way you prompt for AI changes the amount of tokens that you use in a query. As organisations become smarter, that something is AI token optimised. And you use that, I think it’s going to lower it.
And all the organisations that are using AI today, that will become second nature for them to actually optimise it like the way they did it for cloud. And that’s something that we are doing at Onix as well. So whatever AI adoption we do, we have now best practises around how to write those prompts so that the tokens are minimally exhausted and so on. So a small company like us, when we are already doing it, that means that the industry is also definitely looking into it. So going forward, I think it’s going to be an ongoing challenge, but I think that’s a solvable problem.
00:43:38 Tony
Right, right. Yeah, it’s. Presumably it will level itself out, that some companies will take advantage if the larger companies start getting too greedy and some smaller companies can come in and disintermediate that and offer a cheaper way for a lot of these organisations to, to utilise AI.
00:43:58 Susheel
True, true.
00:44:01 Tony
So we’re going to switch a little bit from technology now to leadership. Susheel. And we talked a lot about, you know, some of the stuff that you’ve done over your career, but when you got into Onyx, you had to adopt a whole different style. He became a CEO. You had a company that started from nothing and you learned a lot of lessons along the way. But what leadership characteristics do you think helped you the most to help Onyx succeed? And then what lessons have you learned along the way?
00:44:37 Susheel
Well, that’s a very profound question because there are so many books written about leadership and so on. Right. The core of what I realised about leadership is putting your team first and making sure that every single individual in your team, you are personally accountable for their success. If you can make every single person in your immediate leadership team super successful and once they realise that you are really thinking well of them and your success is really tied to their success, just magic happens, I think. And it’s, it’s easier to say very hard to do it.
00:45:27 Tony
Right, right.
00:45:28 Susheel
Because the trust, the trust develops when somebody knows that, you know what, whatever I do, I know that I have the backing of my immediate leader if something goes wrong. So how do you eliminate the option of risk wherein they can take independent decisions and be able to drive through? So that becomes a challenge. But again, leadership from my standpoint, I think it has various styles. A leader that is great for a small startup where there are five people and they need technical superiority or everybody else.
That leadership is going to be very different from a 50 people company where the process are a little more established, but the CEO and everybody is still hands on, very different from 200 people where you need to start developing processes and so on. So I think a true leader is somebody who can identify where they are and quickly adopt to that particular leadership style and be able to drive that style. But at the same time there are circumstances that change your leadership. There are issues that need instant attention and you cannot just wait and let the team figure out because there are certain, you know, that you better figure it out along with the team otherwise it’s going to blow up in everybody’s face.
So getting ahead of those and sometimes just taking a backseat and saying let the people run with it, let them make some mistakes, let them learn from it. Even though those mistakes can be avoided in the first sense, but you still want those people to make those mistakes because you know they are going to learn a lot and they are going to become better leaders. So there are so many nuances of that. I used to have one leader at IBM and he used to say the reason you win a long battle is you have to do a lot of things right.
So I think the leader is somebody who does majority of the whoever gets the maximum amount of things right. Be it one on one conversation, be it an email that you send out, be it just here listening to somebody and hearing them out, being somebody who’s going to say this is not going to work, back off, let me lead this and you guys follow. This is the only way it’s going to work on this. Taking all those avatars based on the situation and the scenario is a very, very difficult skill. But the one that can develop that skill can be a great leader. And through this journey of onyx plus everything else, I think I’ve learned a bit or two of that and it seems to be showing up in the results so far.
00:48:22 Tony
Well, I think one of the things I’ve seen with you is you’re a very engaged leader. You’re not a micromanager, but you’re very engaged. And I think one of the things you’ve done is you’ve set very, I’d say, very easy to understand strategic goals and you really focus the team on achieving those strategic goals. A lot of times people get very ambiguous with their goals and that creates ambiguity with the team. And one of the things I’ve noticed with you is you’re very strategic in how you set the goals and then you’re very engaged in how you make sure the goals get met. But you’re also flexible Enough to see that if you need to modify it along the way, you do.
00:49:06 Susheel
Absolutely, Tony. And that is something that I learn every Thursday for an hour on a call with somebody that I’ve always looked up to, which is you. So you have a big role to play in whatever I’ve learned about leadership.
00:49:22 Tony
Well, thanks, Susheel. I appreciate that. But one of the other things I think is important too is to move from problem to solution, as I’ve always said. And part of the challenge with a team is they can get too focused on the problem and complaining about the problem and then flipping them to focus on the solution. And I think you’ve been very good at doing that to keep them, you know, away from sitting on the problem side of the equation and really focusing on, okay, how do we solve this problem? Not just, okay, we have this problem. And I think that’s another reason you’ve been able to be successful as you’ve moved along.
00:50:02 Susheel
No, thanks for that, Tony. And that’s something that I learned from you. Right. Are you on the problem side? Are you on the solution side?
00:50:09 Tony
Right.
00:50:09 Susheel
Okay. And you can also judge people as to where they are sitting right now. They’re still noodling on the problem or have they moved?
00:50:16 Tony
Exactly, exactly. And the problem with that with business today is if you spend too much time on the problem side, somebody else is already ahead of you on the solution side, and then you’re never going to get caught up and you end up losing the race very early.
00:50:32 Susheel
Absolutely. Absolutely. Can’t agree with you more.
00:50:35 Tony
All right, Susheel. Well, we got a couple more quick questions before we end. I know you spend a lot of time building and putting your heart into this company, but you also have a personal life. So can you just tell people a little bit what you do beyond that? I certainly have heard many of your different hobbies and things you get into in addition to your family, but I think the. The audience would be interested as well.
00:50:59 Susheel
No. You know, running a startup like Onyx, growing it from almost zero to where we are today, it’s a very, very consuming task. You pretty much wake up in the morning, roll off the bed, you are thinking about Onyx and you are working through Onyx. And then when you are not working, you’re thinking about Onyx and then you get to sleep and you’re sleeping, you are thinking about some of these problems or opportunities. So it’s hard to get a break every now and then. But I think those breaks are important. So I have a few things that I Do. One is I’m a big lover of motorcycles.
So that is the one place where all senses are on their toes. Because one mistake and you got to pay a price for a long time. So you get to get away from everything. And just the wind in your face gives you a fresh outlook to life. So that’s something that I do on a weekend basis based on the weather. So outside of that, just busy with family. I have two kids. One’s going to college, the other is still playing soccer in middle school. So driving around, just enjoying those soccer teams and matches, that’s what keeps me busy and alive. And by the way, till date, I think I’ve read about 45 books this year, so audio, audiobooks. So I just cycle through those books very, very quickly nowadays.
00:52:30 Tony
Well, good, good. Well, I think I’ll end with asking. You’ve been in the industry a long time now, so I just want you to kind of take a long view as a thought leader. And what do you think will be the biggest changes in the industry in the next five to 10 years? And if you can say that without saying AI, you get extra points.
00:52:55 Susheel
That’s a great challenge there. I would say, I think challenging status quo, that’s going to be the biggest thing for all the executives that are running these large corporations. If they don’t challenge the status quo with the way and the pace at which things are coming at them, they will not have a shot at even recouping from their decisions. So just move before just brings me up this Jack Welch’s quote, maybe it’s a great way to close change before you have to. So I would say to all these executives, change before you have to. Because the change that’s coming our way is 10x more than everything else the industry has ever seen.
00:53:44 Tony
Right, Right. And I think that would be the same response you’d give for vendors as well, is not just the company themselves, but vendors. As you said, if you’re not flexible and you don’t move fast, and I think the old thing about moving fast and break things is not the answer. I think it’s move fast and fix things, not break things or make things better.
00:54:04 Susheel
But yeah, move faster and disrupt, that’s the way the world is going to go now. Because it’s, it’s just anything that you want to release today, if somebody says it’s going to take three months, you should question the team. Either you have a team, you have a problem that you nobody understands, or it is a problem that you should understand deeper and try to solve it in two weeks. Right?
00:54:33 Tony
Right.
00:54:34 Susheel
Speed is the game of the. Speed is the name of the game. And our new mantra at Onyx is speed of light. Whatever we do, do it at the speed of light.
00:54:43 Tony
That sounds good. Well, that’s a great way to end this. Susheel, thank you very much. I appreciate the time you spent and looking forward to additional conversations on this in the future.
00:54:56 Susheel
Sounds good. Tony, thank you so much for having me on the on the podcast and all the support that you have been and the advice that you give me every now and then for. I owe it to you on where I am today. So thank you.
00:55:12 Tony
Thank you, Susheel.
00:55:13 Susheel
Thank you.
