In this episode, Todd Crosslin, Global Industry Principal Healthcare and Life Sciences at Snowflake, sits down with Rajesh Viswanathan, CTO at Inovalon. They discuss Rajesh's journey into healthcare, the complexities of the U.S. healthcare system, the role of data and AI in transforming healthcare, and Inovalon's strategies to improve patient outcomes and healthcare economics.
In this episode, Todd Crosslin, Global Industry Principal Healthcare and Life Sciences at Snowflake, sits down with Rajesh Viswanathan, CTO at Inovalon. They discuss Rajesh's journey into healthcare, the complexities of the U.S. healthcare system, the role of data and AI in transforming healthcare, and Inovalon's strategies to improve patient outcomes and healthcare economics.
---
Calling all developers, business leaders, IT execs, and data scientists! Snowflake World Tour is your chance to learn and network. Discover how Snowflake’s AI Data Cloud can transform your career and company. Experience the future – join us on tour! Learn more here.
[00:00:00] Producer: Hello and welcome to the Data Cloud Podcast. Today's episode features an interview with Rajesh Viswanathan, CTO at Inovalon, hosted by Todd Crosslin. In this episode, they discuss Rajesh's journey into healthcare, the complexities of the U. S. healthcare system, the role of data Forming healthcare and Inovalon strategies to improve patient outcomes and healthcare economics. So please enjoy this interview between Rajesh Viswanathan and your host, Todd Crosslin.
[00:00:33] Todd Crosslin: Welcome to the Snowflake Data Cloud podcast. This is Todd Crosslin. I'll be your host today. And with me will be Rajesh Viswanathan from Inovalon. Rajesh, great to have you on the podcast. Would love to hear about you and your journey with Inovalon and, and ultimately what got you into healthcare.
[00:00:52] Rajesh Viswanathan: Hey Todd. Firstly, thank you for the opportunity. Great, great to meet you. So look, this is a bit of a culmination of both a personal and a professional journey for me. I've spent 25 something years, but who's counting, you know, learning about and sort of running these, these massive scale distributed systems.
So I've kind of been somewhat fortunate that I've kind of got to live on the sort of bleeding edge of cloud computing. I spent almost eight years at AWS, sort of up and down the stack there. I spent about close to five years at Oracle cloud, doing similar things there. So I got the chance to work on some really, really large scale problems, working with some really, really complex technologies.
And I realized I was pretty You know, not to sound immodest, but I was getting pretty good at it. What I also realized I was, I was seeing myself doing well was I was able to kind of step back from the sort of technical detail of it and be able to see the big picture. I decided to challenge myself a little bit.
I was getting what I like to think of as uncomfortably comfortable. So I said, like, let's go, let's go test that hypothesis, right? Can I actually go Sort of lean on my ability to go learn a whole new domain and then take the learnings I've gotten from working on all of this, this bleeding edge tech into that domain and in a way healthcare is about the best possible fit for that sort of a, that sort of a challenge.
Right. Absolutely. It's a obviously massively complex ecosystem, you know, massive impact to society, right? Like it's hard to kind of do something that's more impactful to society than healthcare. So here I am sort of like healthcare seemed like it was, it was really interesting for me to get into. And so when Enoblon came calling, there was a few things that really sort of struck me here.
Firstly, sort of like the sort of broad swath of the healthcare ecosystem that Enoblon touches. We've got businesses across You know, the pair ecosystem across providers, pharmacies, and then life sciences and pharma. So we're really kind of spanning that gamut of all of these various sort of entities in this large ecosystem.
So it just seemed like it was a, like a target rich environment for me to kind of go learn and then sort of like use, use technology to go drive this, this big transformation. So that was one big driver. The other one was sort of the data set that, that Enoblant possesses. There's, you know, almost 400 million unique lives that are represented in our platform, 85 billion unique medical events, 25 of the top 25 providers are in our platform, 23 of the top 25 payers are in our platform.
So really like that notion of, are you going to have a significant impact doing this thing was very readily answered. And then obviously, finally, sort of this, this big mission around, you know, improve the economics of healthcare, improve the outcomes of healthcare, it's kind of hard to sort of like have a challenge that is more impactful than that in terms of just, just value you bring.
So that's basically what brought me here and boy, it's been a blast since I came on board.
[00:03:48] Todd Crosslin: No, it's great. And, and I think it, and Ovalon isn't in a bit of a unique situation, especially in the U S and I think you, you said this, there are not many organizations that span all the way from a provider. All the way through to global pharma and really work within each one of those sub industries of healthcare and life sciences.
Sure. If you go outside, I spend time globally, right? As an industry principal around the world. And you, you see these models in Japan, you see it in Europe and you see organizations that really do span all of healthcare and life sciences. But I'd say in the U S. I wouldn't say necessarily you're unique, but it is something that is very unusual to have those different, you know, groups represented within one organization.
And obviously on top, within that, what does that mean? From a data perspective, you've got data everywhere, right? And so I think, you know, when it comes to your decision to have Snowflake be that bedrock. What does that mean for you to have Snowflake across the ecosystem? You know, what does that mean to you as an organization?
[00:04:48] Rajesh Viswanathan: Yeah, no, like you said it, I actually think, you know, healthcare is complex and the U. S. healthcare system is maybe the most complex of them all, just getting all the various pieces involved. And the one thing that is the sort of like the connective tissue really across this ecosystem is, is data. The one thing that we think can actually kind of like, we can kind of connect all of that.
Make that the common thread across sort of like all of the interactions that happen across this ecosystem. We actually kind of like de complexify that, that, that ecosystem really, right? And then that gets us back to our goal of like improving, improving patient outcomes, improving the cost of healthcare.
So that's kind of become our guiding principle as well. Like how do we go take the data that may have been sort of, that may have come into, come into the system through one interaction with. Uh, you know, between entities in the system and then go kind of leverage it in some other place for, for that value.
So there's actually some really interesting examples. One that struck me, especially when I came on board and was trying to learn the system was like, what we can go do to identify gaps in care, right? Like we, we process literally billions of claims to go understand if there's like undetected diagnosis in the, you know, from the claims data and we can then go say, Oh, it looks like, you know, there's a risk of.
Take COPD, you know, there's a risk of COPD that we, that, you know, based on our analytics and all of the expertise that we've built up over the years, we, we, we think there may be a risk. And, and then one, how do you act on that risk, right? Because once you detect the risk, there's, there's the opportunity to go actually improve the patient's outcome, because if they have it, it's obviously better to kind of.
Find it early and prevent them from getting to that intervention end of the, of the care spectrum. Obviously great for, for the payers. It's kind of going to help manage their costs of care. So, you know, taking that data and then kind of like acting upon it in the case of, you know, finding a care gap like this, maybe being able to reach out to the provider for the patient and go say, Hey, can you maybe go check the patient when they come in for the next checkup to see if, you know, You know, they're showing these, these symptoms or maybe going and seeking out medical records for the patient from various doctors they visited and sort of bring them all together and do some analytics around that to go confirm or, or refute that, that possibility.
Or maybe send a nurse to the patient and kind of do an in home assessment. Right. So various ways to kind of like close that care gap and sort of bring that data back to the patient and back to the payer and sort of like close that, close that cycle. So, so really interesting sort of opportunities for us to kind of like connect these, these various sort of interactions to a larger sort of, Value system and, and, and Snowflake's been, been critical for that. Like I talk about data as a connective tissue, Snowflake is our connective tissue for our data platform in Seattle, Oklahoma.
[00:07:31] Todd Crosslin: No, that's great. And, and we could probably spend hours around the quality of healthcare data, but I think that's something that, you know, everyone, our audience needs to appreciate.
Just how difficult that is to bring, right, disparate data sources together and to, and to drive that value, it's a tremendous challenge. Once you've got that, we'll call it that bedrock of data again, so now, you know, something I think people, people are now appreciating is bringing that data to point of care, bringing it to analysts, bringing it to citizen data scientists, so there's a whole application layer that kind of comes along with that.
How does Novoland look at how do you You know, bring the applications to life on top of this massive amount of data that you've brought together.
[00:08:16] Rajesh Viswanathan: Yeah, you know, one of the challenges we had, especially when I came on board, was just what I call the ponds and the ponds and pools of data that we have.
There's a lot of it, right? Like we're, we're kind of by nature and incredibly sort of heterogeneous, uh, sort of technology platform in a lot of ways. And over time, You know, we ended up kind of creating this incredible heterogeneity where there's, there's super valuable data, but they're kind of like isolated away in pockets that are kind of hard to get to.
So the value here is, is obvious, but it's only if you can kind of achieve that sort of, you know, move from ponds and puddles to like lakes and oceans. Right. Like how do we get all the data into sort of one, one place to kind of go, bring it all together to go manage that data, to sort of govern that data, to provide analytics around that data.
There's just so much to be done in terms of sort of like. Taking a more of a centralized platform approach to the data. So that was the big shift I kind of drove when I came on board, which is like having all of this data is great, but the value of the data is only sort of possible when you kind of bring all of that in some meaningful way, more, more together.
And Snowflake's kind of been a super critical part of that. So the way we've kind of kind of driven a data strategy across the organization is, It's essentially sort of like every application shall contribute its data into, into our data platform and can, and shall sort of use the data off of our data platform for what it needs.
So the data effectively kind of goes through Snowflake in various forms as it kind of navigates its way through our application portfolio. And that's been incredibly sort of valuable and Unlocking new use cases, thing that we wouldn't have thought about as we kind of like, you know, started bringing all of these pieces of data together.
We're like opening up all these use cases. Wow, wouldn't it be cool if I go leverage data from here into there and the two together is just a super valuable use case. So, so that's been, that's been huge for us.
[00:10:05] Todd Crosslin: So Rajesh, we've made it this far in the podcast and we haven't said the word AI. If people follow Snowflake Summit and Snowflake Media, we always joke it's almost like a drinking game.
So for anyone out there in the audience who's been playing along and just sitting there with your drink and not able to, to take a drink, we apologize, but let's go ahead and dive in. So let's, you know, let's just kind of hit it straightforward. You know, it's on everyone's minds. When we talk about AI, what does that mean to you?
What does that mean to a Novolon in today's, in today's world?
[00:10:36] Rajesh Viswanathan: I was waiting for the question.
[00:10:38] Rajesh Viswanathan: It's kind of hard to talk about. Talk about data. In fact, it's hard to talk about healthcare these days without AI, and, and it absolutely is massive. It's a massive opportunity. It's a massive sort of disruptive force in a lot of ways, and in all of the domains that we're involved in, and, and, and, you know, we at Enoblon see a massive opportunity in what we can kind of go do with, with AI, right?
There's obviously a lot of considerations around the use of AI, and we can certainly talk a little bit more about it. Maybe I'll first touch upon some of the areas that we're. That we're already starting to apply AI, and it's a little bit of a, you know, it's a target rich environment, given all the use cases that we interact with.
So sometimes it's all about like, what's the one that we think can deliver the, you know, the quickest time to value. I'll be double down on that. So a few use cases I'll kind of highlight. One of them, and it's a little bit of a category of use cases, it's, it's the ability to kind of extract insights out of medical records in general.
Obviously, huge amounts of information in medical records. One we're particularly focused on, and you know, we've talked about it a few times before, is kind of what we can do to help the process of medical coding. If there's, there's one thing about, about healthcare, it is the amount of manual processes that are involved in the ecosystem in general.
And, you know, one, one real thrust of our AI strategy is really, how do we actually kind of like drive down just the amount of manual work that's involved and it's kind of goes back to our mission around to produce or improving the economics of healthcare, right? The more you're kind of doing stuff manually, driving the economics higher, but more importantly, it's actually driving.
It's probably an increasing number of errors because humans are fallible. So a good example of this is a focus around around medical coding, right? There's a process around kind of being able to look at an EHR record and be able to extract ICD codes from it to be able to go for the insurer to be able to go determine what's what's a risk risk codable condition.
That usually takes hours and days of somewhat manual effort. It's a really great opportunity for AI to kind of help make a difference. And that's an area that we spend a ton of time focusing on. You know, not only are we involved in sort of like extracting the ICD codes, we can actually help with the process of auditing that process and so forth.
So you really have multiple ways to kind of like apply AI to that domain. And you can think of that as kind of one example of a broad category of use cases around medical record review. You can think of medical records being useful for, you know, actually confirming diagnosis that were done elsewhere in the system and kind of using those two together to prove an audit, to prove a claim for audit reasons.
You can think of it as helping provide a set of governance structures around decisions that are being made based on the data in your medical record, et cetera. Another one is sort of like, you know, both the front end and the back end of revenue cycle management. And again, this is. You know, the, the somewhat labor intensive parts of healthcare that don't get talked about, but are a huge driver of costs of healthcare, right?
Like, so the process, for example, around a patient walking into an office and, you know, getting verified for insurance eligibility. That's a process that, you know, involves dealing with what could be effectively dirty data or data, data that is not complete. And then providing that to a payer for verifying the patient's eligibility for, for insurance for care.
And the slower that process is, the longer it takes for that patient to get to, get to care, which is really our goal, right? How do we kind of simplify that process? So, so we've got a bunch of things that we're doing around kind of speeding up that process of eligibility verification. So can we actually go use our data and our, and our AI models to enrich the data?
That is being provided to a payer to kind of speed up that, that path to verification, get them to care quickly. So that's the front end of the revenue cycle process. The similar thing exists on the back end, right? Post the care, being able to go make your claim and then have a reasonable expectation that that claim is going to get to being processed quickly enough, right?
Some of our, our claim systems are, Relatively, shall we say, monolithic.
[00:14:35] Todd Crosslin: I was going to say, you can be, you can be kind here. Or, yeah, be, try to be kind.
[00:14:39] Rajesh Viswanathan: Uh, I'll just say not modern. Uh, and the results of those are, these things can take up to 14 days to even identify that you actually had something wrong with the claim.
So, and then you sort of rinse and repeat the process again. You go fix something, resend it, and wait, and so forth. Which, you know, for a provider is, is basically a cash flow problem, right? Like my cash is not coming in when I need it. So how do we actually kind of like help with the process of inspecting your claim, predictably helping with edits on your claims, based on rules that we've kind of learned over the years for different payers having different rules and helping you with making sure that the data you're providing is as comprehensive as possible to kind of speed up that that adjudication decision around it.
So a really great example of where like data and AI models, good AI models are really helping the The, the entities in the ecosystem, uh, and, and it's the parts of the ecosystem that don't often get talked about enough. Like people think of care as like, Oh, I'm going to type in my symptoms and someone's going to pop out a disease, never as simple as that, right?
Another interesting example that we've been looking at quite a bit these days is medication adherence programs. And this, you know, we've got this pharmacy application Effectively an ERP system for pharmacies. I've got to learn a lot about the pharmacies and it's, it's, it's fascinating the number of entities involved, the doctors sending the prescription to the pharmacy, you know, the involvement of drug manufacturers, distributors, nurse practitioners, there's just a number of entities in the ecosystem.
And being able to use that data to be able to predict the likelihood of a patient adhering to their medication regimen. It's actually really valuable. Can we actually go tell, right, Rajesh tends to not take his medication often enough. Can we then drive a program that kind of drives, drives better endurance?
That's a super interesting opportunity that we've been uncovering and kind of like focusing on. So they're just a few examples, but you can see how they're kind of like representative of like this really broad spectrum of opportunity we have.
[00:16:32] Todd Crosslin: Yeah, it's almost gotten to be a little bit, a little bit crazy.
I think the, I was speaking with a group of payers and I was speaking with a group of providers and you were talking about revenue cycle management. We've almost gotten to the point now where it's almost a, what I call a battle of the bots, right? Where you're, you're using generative AI. To look at the appeals process and someone generates an appeal, right?
And that appeal is a, is a large document. And so you take that document and you use gen AI to summarize it and to assess it, right? So that's, that's the, the appeal that is being sent over, like from a provider to a payer for myself, getting to sit right in between all of you guys and you represent right, both sides of it.
And all you're trying to do is try to make the process better, faster, easier. Right. And more efficient for the patients and for all parties involved. It's crazy times, right? When it comes to just how it's being used. So having said that, where does it go from here? Right? Like, so what gets you excited about what is the potential?
for what can happen with AI because it's, it's happening so quickly.
[00:17:35] Rajesh Viswanathan: Yeah, look, the amount of stuff that's been written about what AI can change is, is, there's more, there's, there's probably not enough cells in my body compared to the number of words about AI out there. So, I'm not going to sort of, you know, sit here and say like AI is going to be amazing because it is.
But there's a couple of things that I think observationally sitting where I sit, I've kind of, I've sort of started internalizing. One is, I think there's kind of going to be a little bit of a clustering of like the real value from AI going to providers that have access to that primary source data, because in some sense, the capability of models is It's kind of moving towards a commoditization.
Like there will be much better models and so forth, but they're all going to start looking, it's kind of going to be this somewhat of an incremental change over time as we get better and better at these things, but it's the data that's eventually going to go matter, right? It's the data that's actually going to help with the, the quality of the, what the models are producing.
So I think there's going to be a little bit of a clustering of like real AI value from the providers that actually have that primary source data. And we're fortunate to kind of be one of those things. So that's one observation. I think the other one is like healthcare in particular, I think we'll continue to have this, I think of it as like co pilot rather than autopilot mode, you know, as much as you hear about the magic of open AI, ECDB, all of the various sort of benchmarks that you see now about, you know, they pass the MLE and like they beat every other student there and so forth.
Like there is still that human in the loop element in healthcare that I think is, is Hard, nay, almost never going to kind of get completely eliminated. So I think there's this notion of how much of that do you incorporate into the way you're thinking about the application, that there's always a human in the loop kind of like providing some level of oversight, you know, into the decision making coming out of, out of the AI model.
So I think like you'll see a lot of the, in my mind, you'll, you'll see a lot of the success coming out of the providers that kind of build their application, their AI applications to not be in this. Pure autopilot, I'm just going to tell you what problems you have, but kind of like set it up to kind of expect a, a practitioner in the loop as it were.
So that's kind of like where I see these things going trend wise and I think we're well positioned because we kind of got, we've got the data, we've got the expertise in the domain. So I think that sets us up to do well. to do really well in that direction.
[00:19:51] Todd Crosslin: Yeah, I think you hit the key point, the expertise in the domain.
If you're in the engineering world, software engineering world, there's all the videos out there and the people are like, you know, these auto coding agents are building apps. But the key is, is that there's someone asking the right questions. And it is still an iterative process. And to your point, subject matter expertise is huge.
I look at how, how complex, to your point, and we've talked about how complex healthcare is in the U. S. So when we see that, you know, what are the challenges? Beyond the complexity and the quality of healthcare data. So how do you look at these challenges and, and, and what, what's your, you know, leadership style, what are you doing to kind of prepare your teams for that?
[00:20:35] Rajesh Viswanathan: Yeah, look, if I were to kind of like list out the top three things that make like healthcare data, Somewhat unique, I think, and this applies kind of to AI as well. It's, it's, it's like the volume of the data.
[00:20:46] Todd Crosslin: Right.
[00:20:47] Rajesh Viswanathan: That’s one, right? Like, I think the sheer amount of data coming out of my aura ring here in like a minute is probably like more than, I don't know, a year's worth of data that came out like, just maybe a couple of years ago, right?
Like, I think the, the sort of curve at, of, of, you know, how much data, how much healthcare data there is that's flowing in is just exponential. So in some sense, having the, the systems and the scale to kind of deal with that volume of data is kind of like fundamental. That's super critical for us. Like just, just being sort of scale ready.
I think the second point you mentioned it, but it's hard to put, to find a point on it is the quality of data. There's, I think healthcare data by nature is dirty, meaning like there's just so many manual hands that get into the data at some point in its life cycle that You're going to get like the mistyped addresses and the wrong zip codes and all of that stuff, right?
And you're really kind of trying to make life and death decisions on that data at some point in these things. So like the quality of data and the verification of that is, is massively important. We just can't not do it. It's a little bit like, it's a little bit like security, which is my third point. It's one of those things that you're never going to stop doing.
The bar is always going to keep getting raised on you. And you're always kind of going to like have to have to play both defense and offense on what your security posture is with data. That's, that's another area that we obsessively think about. Right. And these are in some sense secular, like it's regardless of the use case, regardless of like, you know, which product, like these three things are true no matter what the quality of the data, the volume of the data and the security.
So, so that's fundamental. And the way we're thinking about it organizationally is like, therefore, there's, there's a need to kind of organize around those core principles. There's a, an, a, a data platform team that kind of owns and thinks about this thing horizontally across all of our use cases. They are the team that kind of, for example, manage snowflake deployments and, you know, they're the team that basically owns the data and they obsess over it the way you truly do in a, you know, in a multiscale or a hyperscale where you've got like, you obsess over metrics.
Anytime you see a metric flapping, someone's getting paged and so forth, kind of just goes back to how I run systems at AWS. From a product perspective, literally the first question we ask in any, in any product What is the opportunity to differentiate with data? You know, how does our data make a difference to that use case?
So really sort of like moving to more of a, the data is, is the point sort of approach. All of the applications just happen to be various incarnations of that data applying to real life is kind of like how we're looking at it.
[00:23:28] Todd Crosslin: Going back to that quality of data, one of the things that, that we as an organization to your point, horizontally, you know, not just across all of the industries of healthcare and life sciences.
But across all industries is, okay, I've used Gen AI. With JNAII, I created summary data. With that summary data, I ran a traditional ML model, and then I joined it with all this other proprietary data, and I've generated some type of an insight. Right? It could be a model, it could be just an advanced analytic or what have you.
And then the human, the wonderful human with the subject matter expertise, says what? I have a question. Where did that data come from? How did, like, how did you come up with this answer? Oh, great. Gen AI, you know, large language model. And what a big focus for us has been data provenance and data lineage.
Because in the end, what your subject matter expert wants to do, whether it's you, one of your engineers, or a clinician, or whomever. And I think some of the, you know, the perplexities of the world or whatever got a lot better at this is notation, right? It's saying, okay, I've given you this answer, but this is why.
And so that's become huge for us is to have an ability to dive all the way back to that source data, no matter how dirty it is. And you are correct. It can be some of the clinical notes that I've seen in my 30 plus years blows our mind all the day, you know, all day, every day. It's so important to have that traceability to be able to get back to those sources.
So I think that's, that's a huge deal. So for, for other leaders, right, as, as you're going around and being asked, and I get asked this all the time. You know, what's the advice? What, what, what is it that, you know, when you get stopped at a conference and, and, and they see what you're doing, what's the advice that you give to other leaders, right, in this space on how they should approach AI and ML? But I think in general, it's strategy, data strategy.
[00:25:18] Rajesh Viswanathan: Yes, I do get that question often. And you know, the one thing that has worked for us a lot is this sort of platformization of data approach. It's sort of data at its core has to be connected and unified for it to really be valuable. And really, organizations need to be transformed around that core principle.
If you have one sort of unified way to kind of Bring your data together, manage your data, govern your data, then your organization needs to align with that structure, right? Like there's always the push pull of, and we certainly have it as well, of the sort of classic platform versus application sort of struggle, like how much are we investing in kind of building, building it and managing it to how much of it is do we go focus on a particular use case?
And, and we have tilted the scales very much towards like the data is the, is the asset. The applications happen to be views on that data that, that can change and so forth. But the data is really where the differentiating opportunity is. So if there's one piece of advice I give people that are asking me that question, it's really sort of like, think of the data as your core platform asset and kind of organize.
To align with that, right? Like, once you have that, like, you may, you may do a couple of applications that, you know, aren't necessarily as successful as you'd like it to be, but like, you actually have the data to kind of just broaden that much quicker than actually the challenge of like, if your data were not managed together, if your data is not in one place, or if you don't have all the data that you need for your application.
You're not going to be as successful. So that's kind of like the fundamental principle, like, platformize your data in every which way, operationally, from a product perspective, how you think about your business, and so forth. And you'll be successful.
[00:26:57] Todd Crosslin: So, on a personal note, I kind of joked at the beginning, you know, it's kind of like, to be into healthcare, you're either crazy or you care.
We like to, we like to think it's a little bit of both. While your, you know, your journey is, has, has only been into healthcare recently, what kind of drives you day to day? And what kind of, you know, is there a story in there or stories in there that kind of make it so real for you?
[00:27:20] Rajesh Viswanathan: There's uh, there's certainly stories and scars as well.
I, I think, you know, a couple of months in I, I, I came in with frankly, the hubris of someone who's kind of worked with large systems coming in and saying, Oh, this is, this is a technology problem.
[00:27:34] Todd Crosslin: Right.
[00:27:34] Rajesh Viswanathan: And like, you know, all I need to do is to sort of like get these, these systems. to sort of interoperate better.
What I've learned is it's not a technology problem at its core, right? Like it certainly is that, but it's really a fundamental mindset and a cultural shift, uh, about how we think about this, this space and, and, and to gain an appreciation of the real world that kind of sits around this is sort of huge.
It's been a lesson for me. I've had a lot of sort of angst around sort of like problems that I approach from a pure technology lens and then kind of realized it's actually not that, right? Like here's, here's all of the reasons things are the way they are. And the core of it lies this, this sort of human, the loop point that I mentioned earlier about it.
You don't need humans to kind of provide that like validation of some of the decision making that's happening in healthcare and so forth. So it's really kind of been liberating for me to kind of accept that, as that is one of the bigger challenges. And that's why, despite being, what is it, 18 percent of our GDP, healthcare is not a technology industry, right?
Healthcare is a, is a patient customer, uh, end user space that we have to kind of keep in, In our mind, and the technology is an amazing enabler for it, but like, we have to remember that. Right? And that's, that's been a huge learning for me. And like, I think it's super important to keep that in mind as we kind of build this thing.
[00:28:53] Todd Crosslin: Absolutely. We can't thank you enough for coming on the program. Any parting thoughts for our guests?
[00:28:59] Rajesh Viswanathan: Oh, and first of all, thank you very much, Todd. I really enjoyed the opportunity and I want to thank you and everyone at Snowflake for all of the sort of support we've gotten, the genesis of some of the things that, that have happened at Inovalon, thanks, thanks to Snowflake.
So I want to sort of. Record my appreciation of that. And there's, there's a whole lot we have, you know, happening and more coming up. Well, thank you again for joining us. Thank you very much, John. Enjoyed it.
[00:29:24] Producer: Calling all developers, business leaders, IT execs, and data scientists. Snowflake World Tour is your chance to learn and network.
Discover how Snowflake's AI Data Cloud can transform your career and company. Experience the future. Join us on tour. Learn more at snowflake. com slash world dash tour.