How Gen AI and Automation Are Reshaping Healthcare

Prankur: Hello and welcome to an all new episode of the Zinnov podcast, where we dive deep into the latest trends and innovations, getting unique perspectives across the Intelligent Automation space, bringing in the thoughts and perspectives from industry leaders across the world. I’m your host, Prankur Sharma, Principal Automation and AI at Zinnov. Artificial Intelligence and more specifically Generative AI is poised to play a pivotal role in automation across the healthcare industry. It has the potential to transform the way we diagnose, treat, and manage diseases. As we delve deeper into the era of Generative AI, it has become increasingly clear that it holds promise for not just improving efficiency and accuracy of healthcare discovery, but also driving breakthroughs in the prevention and treatment of diseases.
Ultimately, the goal is to enhance the quality of life for patients worldwide. Today we have with us two guests to discuss how a healthcare organization can make the most of AI and automation. Seth Kutty, who heads the Predictive Analytics and Data Science group at Kaiser Permanente, and Arun Hiremath, the co-founder and Chief Business Officer of EvoluteIQ.
Welcome Seth and Arun, to today’s session.

Seth: Thank you, Prankur.

Prankur: Let’s get started. So Seth, let me start with you. Can you talk a bit about your role and how are you leveraging Automation, Predictive Analytics, and Generative AI in the healthcare domain today?

Seth: Yeah. So first of all, Prankur, thank you for having me. So I lead Data Science, Engineering, and AI at Kaiser. But within Kaiser’s health plan group, Kaiser is actually organized into three main entities. One is what we call the ‘health plan’. That’s the organization that sells the insurance component and we’re what we call a vertically integrated health care provider. The other group is the Kaiser Foundation hospitals that focuses on the care delivery part of our business, which is managing the hospitals, the medical office buildings.
And then the third group is what we call the physicians medical group, which is really the largest group of physicians in one organization in America.
I’m on the health plan side of the house and my role really is, me and my team, we develop solutions for our Sales and Marketing organization for go-to-market strategies and we use AI specifically to look at accounts that are risk. We look at retention risk. We look at opportunities for growth and we use AI in a lot of different areas within the administrative part of the organization to really drive some of that go-to-market opportunities.
And in terms of Generative AI, it’s fairly new for us. We started looking at it right after ChatGPT came in, and we’ve started to think about what are some of the opportunities and use cases within the health plan site. I know there’s tons of opportunities, obviously, within the care delivery and healthcare, in general.
My colleagues on the other side of the house about basically organizing that, but I can talk a little bit more about the Generative AI use cases as we go on.

Prankur: Great, great to know that Seth. Yeah, all very exciting areas to work upon.
Arun, considering your role and work, according to you, what’s been the paradigm shift that AI and Automation has caused? And I think the term we’re using today is Intelligent Business Automation. So how does that fare for enterprises in your experience?

Arun: Yeah. Thanks, Prankur. Actually, it is truly a paradigm shift and I see this as an inflection point for the enterprise software as such. I know the enterprise experience, so as I see it from all the discussions and observation of so many enterprises, one of the most important thing, especially with Generative AI is the democratization of innovation and problem solving. So it is not in the realm of some small group of people. Now it is spread to almost anyone in the enterprise to be able to create content, solutions just by simply describing what they want, right? This actually takes innovation to a different level altogether.
Second piece is mainly automating this complex workflows. Before there was a lot of human intervention to be able to come up with the flows and then trying to automate that with whatever the platforms, but generally, it makes it a lot more feasible where the AI can mimic your judgments, human behavior, human skills to be able to help with your workflow automation.
The third point was very important where hyper-personalization at scale, right? So all the enterprises can now look at all the data, behavior, all the content they have to be able to personalize all the content and the applications to an individual, whether it’s external or internal customer, right?

Then of course, the faster iteration experimentation, that’s possible with the newer tools where you can actually try it out, try out your idea and in an order of days to be able to see whether your idea actually can be realized or not. And of course, finally, one of the key things is empowering employees.

Prankur: A lot of food for thought in terms of how companies can think about their future roadmaps of using these Automation and AI technologies.
Seth, just bringing you in here, we see that healthcare companies lately, especially over the last couple of years have taken a lead in leveraging Intelligent Automation technologies and both of you mentioned, that you are using it on the sales and marketing side, so more operational use cases, but even on the patient experience side there are areas where there is disruption being caused.
What are the key outcomes that companies can think about when they look to implement AI and Automation technologies?

Seth: Yeah, that’s a great question. I think, if we think about the healthcare industry, we’re not, and you’ll see this is across the board, across everywhere in the world. It’s always a lagging industry. It’s not always the cutting edge. And the reason for that is it’s highly regulated. Change management is always a very complex process, small changes affect really people’s lives. So there’s a lot of construct around what really we can change and how quickly we can change and that’s been the traditional thinking. And what we saw when COVID happened was this ability this cross-collaboration across the world where people got together for drug discovery and an unparalleled, unprecedented level of cooperation started happening and it was a wake-up call. It was a wake-up call for organizations that have been lagging behind to say, ‘Hey, you know what? There is a new approach and a new paradigm on how we can collaborate, innovate, and change our business processes faster. And when I think about where we are now as an organization versus where we were a few years ago, there’s a lot more discussion that’s going around on, you know, agile transformations, digital transformations, and I think, it was part of the vocabulary, but it wasn’t part of the psyche and now it’s part of the psyche where we’re beginning to start thinking about, okay, number one, you know, how do we use AI Automation in every part of our organization?
On my side of the house, when we think about go-to-market and what the sales account management teams are looking at, there really is a very strong driver towards summarizing content, right?
We’re not generating content on one side. It’s more about the fact that Kaiser as an organization generates a huge amount of highly curated market ready conversational topics for our sales account management teams, things that we want to be able to share with our accounts and our customers around what we are doing, how are we progressing? And this, and when we cut across the entire nation and we’re looking at different regions, that’s a huge opportunity for us to be able to say, how does a salesperson in 10 minutes understand what is going on within the organization and being able to articulate that effectively to our customers?

Because our customers are dealing with a huge deluge of information as well, right? So what we want to do is we want to be able to make it crisp and Generative AI gives us that summarization capability. There’s a lot of governance that is also being applied towards Generative AI. It’s not just, you know, let’s go build something and make it work because we’re healthcare we have to be extremely careful about what is being communicated.

Prankur: Yeah, I think the possibilities are truly endless. I think whichever functional area you pick up, I think there are use cases that you can go after. Arun, from your experience, can you help us understand… when I think a lot of companies go into these specific use cases, and they start with a certain goal and an outcome in mind. How do you ensure that that benefit that they were envisioning at the start of the journey is something that they can realize at the end of the deployment?

Arun: Okay. Yeah, no, that’s a that’s a very good question because it’s very overwhelming. The technology’s pace is very fast and then enterprise are still trying to catch up. And some of the best practices that again from all the experience that we’ve been talking to various industry leaders is more what I’ve been seeing is something the new technology is also allows you to get away from this long release cycles where you have an idea and then you see the results in years. So get away from them and then get into test and learn something more agile. A completely agile approach is one of the key things that you want to try out and see what the technology holds for you, whether it is meeting your expectations.
And also then second is more about focus on augmenting people. So don’t look at technologies as replacing them, but empowering them. Bring people into the automation cycle. So that actually helps both adoption as well as an ROI from the automation investments.
Other key thing is prioritize integration. So when you bring in new technology and then you have tons of legacy systems and all the investments that you already have, trying to see how you can actually connect both of them instead of just trying new technology in a silo, so that you have a unified data and workflows which is very important from an enterprise standpoint to be able to see the full benefit of that new technology, as well as whatever their previous investment.
And as I mentioned earlier bringing human in the loop is extremely important. It’s not just automation as in, it just takes care of it completely independently, bring the human in the loop, and especially for an oversight or even for the key decision making, so that also helps with the overall effectiveness of the process.
In terms of what all other things that you can do is in make sure that cyber security is also at the high priority. Because new technologies bring new vulnerability. So that’s one of the key concerns many people have about any new technology is that make sure that you take care of the security part of it. Then make sure that the enterprise is able to achieve the results that they want. Another key thing is, it’s interesting that everyone wants to try something new and trying to reinvent the wheel. But instead, try to cultivate partnerships, dig into the ecosystem. So now make sure one is empowering your own employees, but also look at the ecosystem, whether technology providers, ISVs, all the new ideas that are coming in. Try to see, make use of it and leverage the technology that’s available to you.

Prankur: Understood. And yeah, I certainly agree with a lot of those approaches and from your perspective, what the specific challenges that you came across when you got started with the overall digital transformation, automation program, and how did you overcome some of those?

Seth: I don’t think we have overcome it completely. It’s an evolving landscape, but I have to say that I think Generative AI is a game changer for us. If I think about a year ago when we started putting a roadmap together for what AI can possibly do, I mean, everything has been probably accelerated or rationalized differently because of the way we can now do things with this new capabilities, right?
My team’s really responsible for developing predictive models to help sales marketing teams understand which segments of the market to go after, which accounts are at risk. But there’s a lot of textual data that’s captured in every interaction that they have with a customer, which has been fairly difficult to process and understand, because it’s a huge volume of information. And where I think we are probably going to go is to really to tap into and mind that information using the Generative AI capabilities and develop more insights about our customers so that we can have more meaningful conversations. We can develop very customized products for them. It’s going to change our entire interaction with our customer base as well.

The other thing, and I want to just tag on to what Arun was talking about a little bit earlier, was we’re starting off fairly small, right? So there’s change management within our internal teams, the challenges of them getting up to speed with a new technology, And again, we have different perspectives on what we can use it for, where we think we will not introduce bias and cause conflict even in understanding or in real life, I think the way we’re taking it in small steps and we’re thinking about first, let’s figure out a way to hyper automate areas of our workflow. And I think that has a lot of capabilities. We have a ton of use cases around that already, and then moving a little bit further up, which is settling around hyper personalization. And Arun talked a little bit about that.

Now we weren’t able to do that primarily because if we think about our business processes, if they don’t change very often and moving a complex, first of all, health insurance is extremely complex product, right? It’s not like selling a car like a Tesla. It’s extremely complex product to sell and to design. And so the business processes to support them, we can’t just change them on the fly, but we can test experiment, design new business processes to help with that level of personalization, you know, at a specific level, like we might be able to do it for a specific account much faster than we did before. And I think that capability is now available and we’re really excited about that piece. That we can literally design and test a new business process powered by AI in a way that’s never been done before.

When I used to design business processes in the past, it was always about, okay, we’re going to do this. Now let’s worry about change management. Here we can mitigate a lot of those risks on change management because we help with bringing the stakeholders into the design process and getting them to understand where AI can help and where a human gets into the loop to be able to drive that change process.

So it’s a very different strategy for us. Again, we’re very positive that if we’ve tested it out, we know it works. We’re now trying to do this at scale.

Prankur: Yeah, thanks, Seth, for really giving us a deep view into how you’re thinking about navigating your journey. Arun, from your perspective, is the platform approach really helpful, right? I think we have various ways in which we can design and deploy solutions. You can have something which is right off the shelf which is very point specific, whereas you have these other more platform centric approaches, which gives you a more horizontal capability to play around and customize and build a solution that really caters to your needs. What has been your experience of designing these solutions and then deploying these solutions at scale at large enterprises?

Arun: The platform approach certainly makes a lot of sense in general compared to a custom solution or a custom code on two ends of the spectrum, your platform gives the best of both worlds, right? So first thing foremost is it decouples your business cadence and technology cadence completely. So that way you don’t need to worry about one or the other up to being ready, right? And then, of course, the platform approach simplifies your complexity. So technical complexities are taken care of by the platform with an easy to use interfaces, automation, and all this technologies offered so that you just focus on the business.
So simplifying that whole overall complexity. Then most importantly also this whole approach comes with the scale automatically. You don’t need to worry about your approach, how much scale that you’re ready to do when you’re even trying out small things because the platform is there and most of the platforms can actually scale to the enterprises.
And of course, flexibility and Seth was talking about change management, right? So change management is an inherent feature of a platform, in general. So when you’re talking about, do I have to determine everything I want to the last T and I, you can try out, start, and then continue building on top on a platform versus a ready-made solution, you have to decide everything up front.
And in terms of also the platform allows a collaboration, right? So it is not like, some team is assigned to get this done. Then people can add various pieces as and when they are ready and as and when it is applicable to their particular domain. So that that way you can actually get a faster digital transformation.

Prankur: Got it. And last couple of questions from my side. So Seth you spoke about a lot of use cases across various functions of the enterprise. What are the top three use cases you are personally excited about, the ones probably you are going to look at next in your Automation and AI journey?

Seth: So I think the most of my use cases right now are centered around retrieval augmented generation. So, at least with Gen AI, I think the goal for us is how do we take a lot of this content that’s already available within the Kaiser’s ecosystem and make that readily available for our customer facing partners, right?
I think those would be number one use cases. And again in many companies, which have a lot of off sign conversations and documentation, this would be a really good opportunity. I think the second thing we’re looking at is orchestrating our data automation processes. We do have a lot of different automation tools and, because we have both a cloud as well as a ground operations, we do on-prime as well as we do cloud, we needed a tool to be able to bridge those two right to have a hybrid architecture to move data across and automate and orchestrate processes across the various cloud and on-prem applications. So I think, as we think about automating many of these things, there’s going to be an AI component to drive that orchestration all the way through. So that would be, I would say, a lot of our use cases would be centered around automating some of these data components as well.
I think the third one of it is really going to be self-service analytics. And when we talk about self-service analytics, we have thousands of dashboards that we’ve built within the organization doing to inform, guide, prescribe what are the next best actions and recommendations for our users, and that’s changing, right? It’s changing because now people are coming in and usually when you build a dashboard and you make this available, the next people always have like, okay, what about this? What about this? Can we do we have analysis around this next thing? And there’s no end to that level of questions. And the reason for that is because as much as we build dashboards, they’re still fairly static, right? You know, there’s a lot that goes under the covers in terms of building it and exposing some of that. And there was not much of an insight as you can’t infer data directly out of it.

And I think Generative AI is beginning to change that, right? We’re beginning to get to a point where we can have self-service analytics with this new technology. So going beyond the dashboards, augmenting dashboards, augmenting workflows so that people can ask, the customers can ask questions directly and say, you know, if it is not just, you know, our penetration rate at this account is X percent, you know, why is it X percent? What are some of the opportunities for us to increase? And there’s a ton of information that’s already available. So having the self-service, learning, understanding, insight generation, I think it’s probably the next thing for us. As an analytics organization, we think that will be a game changer for them.

Prankur: Got it. Any last key takeaways from your experience, anything that you would want our audience to take away as they think about scaling and starting on their AI and Automation journey?

Arun: Right. I mean, I think we touched upon many items, right? So, especially in a Healthcare domain where we were talking about earlier, just to add on to Seth’s comments, we have seen the possibility in almost every aspect of the Healthcare industry, and it’s a huge industry, of course, if you look at from the patient, whether it is their experience themselves, or whether it is diagnostics from the doctor/provider standpoint, how they can provide better care for their patients or the overall operations, right? Technology is actually going much faster than the one enterprises can think of adapting.
There is a general concern about data, how the privacy is handled, HIPAA, and all that. I think those will be resolved. What we see is that take these technology pieces in bits and then start testing them out in small areas where you can feel comfortable and then with the platform approach, you can actually go further and further and integrate into your mainstream.
And that’s what we see. The technology is there and it’s up for the enterprises to leverage it and get ahead.

Prankur: Yeah, you summarized it well, Arun. There’s certainly immense potential across the various areas of healthcare enterprise today. Whether it is in some of the horizontal functions like Sales and Marketing, HR, or even in patient care, in hospital operations, in revenue cycle management. But the key is to actually prioritize some of these areas, go in a more structured manner and take an approach which can be cost effective and can give you the envisioned outcomes in the most optimal fashion. Thanks a lot, Arun and Seth for sharing your insights.

Seth: Thank you. Thanks for having us.

Arun: Yeah. Thanks for having us.