Payer Panel: How Agentic AI Closes the Value‑Based Care Execution

Hello, everybody. This is Erica Spicer. Mason with Becker's Healthcare. Welcome. And thank you for joining us for our webinar today. How Agentic AI closes the value based care execution gap. So before we get into our discussion, I'll walk through just a few very quick housekeeping notes. First, if you have any questions for our speakers throughout the webinar today, feel free to type those questions into the Q&A box that you see on your screen, and we'll be sure to follow up to get you an answer. Today's session is being recorded and it will be available after the event.

You can use the same link that you use to log into this webinar to access that recording. And finally, if at any time you have issues with the audio or visuals, just try refreshing your browser. And you can also submit any technical questions into that same Q&A box. We are here to help. So with that, I'm delighted to welcome today's speakers. We have a great group of leaders with us today. And to get us kicked off, I wanted to ask each of them to share just a little bit about themselves and the work that they're doing in healthcare. Um, and Amy, you are coming up first on my screen, so I'll have you kick us off with intros, if you don't mind. All right.

Great. Delighted to be here. Thank you for for inviting me to join today. So I am an internist by background, and I spent a couple decades at Kaiser Permanente. So really grew up as a physician, as an executive in an entirely value based care system. Then I was recruited over to Providence Health System, um, up and down the West Coast, 52 hospitals, 10,000 employed physicians, um, to help really lead the journey to go from what had always been a predominantly fee for service based infrastructure to being more value based care. Um, so by the way, Brian, your background looks very familiar. Uh, the Seattle area is where I still have a home. Um, but, but then, uh, did a quick during Covid, a quick little touch and go with a health tech company focused on the experience of care because as we know, it had had some issues.

Um, and, uh, then a year ago joined CVS health, where I lead our medical affairs organization and our clinical strategy to make to simplify healthcare. Amy, thank you so much for the background. It's great to meet you and great to learn more about you. Uh, Rhonda, would you like to go next? Sure. Um, so I started out my career in the health care space, um, working at WellCare as well as Centene. And today I, um, working for a major health Caritas, um, which is based out of Newtown Square, Pennsylvania. And, um, I oversee marketing, communications, government affairs, compliance risk and the audit teams. And so I think that gives me sort of a unique vantage point to how value based care actually works in practice from a clinical operations and reputation perspective.

So I'm looking forward to the discussion today. Oh, absolutely. Rhonda, it'll be great to have your perspective. Thank you so much. And Brian, you're up next. Yeah. Thank you. Hello and hello, Amy. Yeah.

Our paths probably crossed because I, too, worked at Kaiser and back in the day. But currently, I am the regional vice president for provider contracting at Humana's Western Market. My team is responsible for procuring the contracts that allow us to have a robust provider network for our members. And I'm. Which is great because Humana's recognized as a leader in value based care and a significant part of my role is helping evolve and expanding that to make it available to more of our members and more of our community. Personally, my passion is about providing high quality, affordable care to as many people as possible, and this role allows me to do that because the relationship between the patient's provider, that contractual relationship is where behaviors flow from, right? And so that the bond between those two and the effectiveness of those partnerships allow us to serve our communities more effectively. Mhm. Another important perspective you're bringing to the table.

So, Brian, thank you so much for being here. All right. Monique would love to learn about you next. Sure. Good afternoon. Uh, Monique Stoner, I am with Oscar Health. Um, I am the national vice president over network contracting or we call it network management services. So I oversee everything that is related to regional as well as our specialty contracting, uh, network adequacy and also working with our out-of-network providers. So, uh, at Oscar, we're in, we're a really large player in the exchange marketplace.

Uh, we have, we're in 21 states and we have about 16,000 providers that we manage on a day to day basis. And I am a clinical psychologist by licensure and background. I started my career seeing patients every day, and I'm on a mission to really bridge the gap between payers and providers. I've worked on the provider side, I've worked in consulting, and then I've also worked for three other payer organizations. So value based care, I think is critical to being successful in network contracting. So I'm excited for today's discussion. Thank you, Monique, and it's great to hear that, that that you have that passion about bridging the gap between payers and providers. So looking forward to hearing from you in this discussion. And last but not least, Robert, you're up next.

Yeah. Well, thanks. I'm a pleasure to be on such a panel here. Um, Robert Conley, I'm the global market leader of healthcare for Pega and I've got a long past. It seems like I've probably crossed paths with a lot of you here, starting back in the late 90s, where I was a VP of web technologies for McKesson, and I invented the first physician portal, I guess, or one of the first. It became pretty much a dominant hospital platform for about a decade. Um, developed a number of products, patented a few things, and then left to form one of the largest health information exchange companies called Medicity. Um, interesting point there. We were acquired by Aetna just before the CBS merger.

So, um, I was a VP of innovation for Aetna for a few years. I'm creating other products, bringing things to market, you know, certified in EHR on some of the technology I had patented, um, retired from Aetna the first time. Apparently I have a retirement problem. My wife tells me, um, so, um, basically I wound up at Walmart for a short period of time helping launch their Walmart health initiative. Didn't want to leave the game. And somebody introduced me to Peggy. I fell in love with the technology. So I live the dream every day. I get to work with some of the largest organizations in the world and how we're using this new tech AI technology, agentic and otherwise.

Um, so just a pleasure to be amongst you all and thank you. Oh, Robert, great to have you with us. And thanks so much for rounding out introductions here. What a great group we've got together today. Lots of different diverse perspectives that we're bringing to the table. And I wanted to get us started by talking on something that I think most of you have already hinted at at some point in your introductions already. And that's that value based care initiatives are expanding. And alongside that expansion of value based care, there is also a lot of operational complexity right now that health that health plans and healthcare providers alike are facing. So from each of your seats, what do you see changing most in the day to day work of managing value based performance?

And where are you feeling the greatest pressure at this moment? We'd like to get us started. I can start. Thanks so much, Rhonda. Sure. So, um, I think most organizations have matured beyond just the strategy phase of value based care. Um, the real challenge now is ensuring that execution is, is scalable. So what's changing is the level of operational discipline required, including like coordinating clinical teams, community resources, data insights, um, provider partnerships in real time. And so these are continued pressures and, and they're absolutely data fragmentation, um, pressures as well.

So the data exists, but it isn't always actionable in workflows. So for member engagement, it's, um, it's hard to reach members who need the care the most. Um, especially, you know, in my space, in the Medicaid space, which is a big core of our business, um, for providers, um, we must align and operate under like different incentive structures. And lastly, I think we have, um, intense regulatory complexity, which, you know, is also played out in the Medicaid markets. The pressure today isn't knowing what to do. Um, it's building the systems and partnerships that make it happen consistently, which is why Amerihealth Caritas continuously, um, invests in our communities. Um, our members and our providers. And I forgot to mention earlier in my intro that I also have responsibility for our philanthropic efforts, um, at, um, at Amerihealth Caritas. So I'll stop there and let others join.

Great insights to kick us off. Rhonda and Brian, were you about to weigh in? Yeah. Thank you. I want to start with what hasn't changed and then we'll get into what has. So what hasn't is the core mission, right? It's serving our communities by focusing on prevention, access, whole system, health. What else? What also, what hasn't changed is the pressures we're under for this.

So we you know, inflationary costs, rising acuity, aging populations, regulatory expectations, maturing technologies, All of these pressures have been there. What I think is particularly different now is the acuity of those pressures. The inflationary costs are rising faster. The tolerance of our regulatory agencies and our members of having the same level of care for the cost of care is not acceptable. And so all these compounding pressures are really forcing us to go to that next level. And what I see is the biggest challenge with this is an operational one. I see many providers who are living in both worlds. They're still in that old world. Fee for service, volume based higher acuity care delivery while trying to get some extra money through a value based care.

And the two are competing, right? Value based care is all about prevention, you know. Right care, right place. Right time. You know the health of the population. That's what you get. Incentive for. That's where the money is. Is caring for that population.

And when you have an organization that has those competing priorities, you're, you know, you can't be successful. You're, you're sub optimized. You're, you have conflicting incentives. You can't invest the way you need to in doing the right thing. And so I think that's our biggest challenge is getting more to fully, truly embrace value and to do the hard work to be successful in it. Yeah, a lot of through lines. Can I pile on that one? Erica. Oh, please go ahead.

Amy. Yeah. Um, I, because I have to completely agree with both both Rhonda and Brian said that the complexity in the system has actually eroded trust in the system because patients get conflicting information. There's lack of shared data, there's lack of coherent systems that when you go from plan to plan or provider to provider. You get different messages from all of the above. Um, and, uh, the structures of the way we're set up today in the fee for service architecture that the most of the care delivery system is still architected around is counter to the architecture that supports effective value based care. So that complexity and the architecture of the ecosystem, um, is the biggest hurdle that we have to deal with right now to rebuild the trust and enable, uh, broad based, effective performance on all things value based care. Yeah. If I can add just from an Oscar perspective, you know, Oscar operates exclusively in the ACA exchange marketplace, where the average tenure of a plan member is approximately two and a half years.

So that level of transiency and the high turnover with our membership. It really disrupts that longitudinal care management. And it also makes it hard to align financial incentives with member outcomes. So as a QP, we face a lot of risks associated with downside financial arrangements with our providers. So in our line of business, it forces us to execute value based arrangements or as I like to call them, alternative payment models. Um, that's my preferred term, but we have to align that with something that's, uh, that meets the reality of our membership base. And another point that I'll make about QA is that, you know, moving too much of our book of business away from fee for service, it creates revenue unpredictability. And so that requires some significant investments for us in infrastructure to support value based care. So we we are being very selective.

And we have been at oscarj selective in which provider partners we choose to have value based arrangements with and the types of contract terms and incentives that I will agree to administer. And just to sort of add on to everything, um, I have a perspective that aligns with what everybody has just said, but also from the technology side. Value based care may be one of the most complex business relationships on the planet in any industry. So I think we, I've often heard healthcare get a bit the short shrift. It's always seven years behind the rest of the world. Well, I would argue it's seven times more complicated than anything else you're doing. So it's not so much behind. It's just the technology is now catching up. What I'm in right now, and we're all talking about this agentic AI, and that's just a piece of it.

But that whole movement is really accelerating these solutions. We've always known how to solve it. We just not have not been able to scale it. And so what we're seeing now is the emergence of newer technologies that are helping amplify what we've been doing, um, to achieve that scale, that repeatability. Ultimately, you know, repeating what Kaiser Permanente did with their leverage from primary care initiatives and things like that in the past. It's just we know it works. We just don't know how to scale it. And I think this is where we start to see the emergence of technology helping. It's not changing anything overnight.

It's not a magic bullet, but it is starting to pick up in areas that help alleviate some of the pain and starting to drive some of that value of that holistic care experience. And value based care is the really the first one that's been told you're paying for it. Up till now, we've not had that position in health care. Nobody was responsible for this piece of it. Now we're seeing it for the first time, and I think the technologies are now emerging to address it head on. And that's what we're all looking for. Mhm. Robert, appreciate your high level perspective. And it sounds like some of the through lines I picked up throughout responses here is that operational complexity has increased perhaps faster and more significantly than in the past.

And so having a lot of discipline around what operations look like. And also, um, it sounds like being aware of the structures of, you know, fee for service and how it's set up versus value based care and how it's set up and trying to kind of, um, reconcile the two, reconcile the differences. It sounds quite challenging, but Robert, to your point, it's interesting to see how technology is coming into play here. And I want to go further there in just a bit, but I do want to talk a little bit more about, uh, follow through. So value based care, whether we're, we're talking from the payer perspective or the provider perspective, we know that there's a lot of focus on identifying gaps, whether that's gaps in care, completing follow ups, managing care transitions, and also preventing avoidable utilization. So I'd like to know where you're seeing the biggest gaps currently and how you're kind of managing, connecting those dots and bridging those gaps. I think. So Amy and I have crossed paths, um, in the past as well. Um, Yeah, I was on a management team and she was on the board.

So she might recall that even years ago. Um, we, we were getting better at identifying members who need support. But the, the real work is, is helping our members navigate the barriers that prevent access to, to health care, right? So some of the most common barriers we see today is still unreliable transportation, like housing instability, um, health equity or health literacy, um, scheduling, I guess, access and care coordination, but, you know, closing the, the, the care gaps often have less to do with analytics and more to do with addressing the structural barriers and the social determinants. So at Amerihealth Caritas, our, our community and our care management teams work directly with members, um, to, to address issues like housing, transportation and health literacy. Um, but our, our members want to access health services. So we try to meet them. And you hear this phrase all the time. We try to meet the member where they are.

Um, but we also try to provide that critical link to care. So yeah, remember crossing paths? Well, Rhonda and I'm not surprised you ended up somewhere so fabulous like Amerihealth Caritas. So congratulations on that. Thanks. You know, I think you you really highlighted something about closing gaps of care. And that's we only know what we know and we don't know what we don't know. Um, and that's where I think and Robert, something you've been talking about is how do we get the data integration much more effective than we have today? Something Monique said is, you know, the average tenure with Oscar is two and a half years.

The gift we had at Kaiser is our tenure was very long. And so we had we had significantly less care gaps because we knew when people not only were missing, you know, a hemoglobin A1-c. But we knew when they didn't have transportation, we could tell that from the data. When people change on a regular basis, we don't have that information. So more effective healthcare. Um, interoperability and health data exchange we believe is fundamental to being successful in doing what. Value based care is designed to do, which is to improve health outcomes. And it's one of the reasons last week we announced our health 100 work is to actually get an ecumenical like everybody can participate in, in one version of data platform so that we can exchange information amongst insurers, amongst pharmacies, amongst care delivery systems, so that we can have more effective identification and closing of those gaps. Couldn't agree more.

I mean, that's been sort of the vision that we've had for a long time in the health information exchange world. Can we aggregate and pull things together, what we're seeing now within the payers and, you know, as they build their BBC, you know, programs and whatnot, is that they're using these tools, whether they call it a consumer data platform, often see these in the hyperscalers now where they use that and they're starting to aggregate data through these interactions, these workflows. So using fire and some of the technologies and interoperability to connect to an EHR, it's less about just acquiring data, more about acquiring it within a workflow. So you have that, you know, I'm doing this only for treatment and payment. So you don't have those barriers normally just aggregation and we're seeing them use that effectively. So time's starting to come into the picture is what do I know now? What can I get now to help make my decision? So it's not just the aggregate of data, but it's the curation of data, adding in real time elements which we don't have enough of. We don't have those insights into somebody's daily life, like somebody said early on.

So completely agree completely. Yeah, I think I think you all covered it very well about the care gaps and the technology Enabling better care gaps. I like when I think about your question about the gaps. I'm thinking about the fundamentals that support the infrastructure to close the gaps. And when I think about value based care, some of the most significant care gaps I see are things like training and culture, proper incentives, timely embedded data insights, and front end capacity of our caregivers to close those gaps. And so, you know, investment in those really understanding how to build those critical elements are necessary. So you have the infrastructure, the capacity, the skills, the culture to do the right thing to, to focus on that preventative medicine and those care closures. Yeah, I completely agree with you on that, Brian. And from my seat, it's because I oversee provider network.

One of the unique approaches that we have at Oscarj is the partnership between network management and our medical management and care navigation teams. So while they focus on member care gaps and really understanding and having deep insights and knowledge about what the members need to have better health outcomes, my team then focuses on our network design and optimization, and that's how we move the needle with value based care. So when it comes to network optimization, this is really the process of how we recruit and refine the right mix of providers that we have in our network. And that's everything from primary care physicians to our specialists and some of our ancillary support, you know, even our virtual care services that we offer members. Um, it also means building the right coverage across all of our geographical areas. And right now we operate in 21 states. So we look at market intelligence, we look at, we do some competitive benchmarking. Um, and then we look at how, How? When we're a recruiting providers, what capabilities and qualifications do they have?

Do they have a strong population health management process in place that they would actually become a high performing provider in our network and be able to help us with closing access to care gaps and making sure that there are minimal wait times for members when they want to get or need to get an appointment with the physician. So I think, you know, we have a very unique strategy around network design. Mhm. Monique, I really appreciate what you said because I was starting to wonder, you know, I'm hearing about all of these, these wonderful ideas behind supporting members at a foundational level to work upstream and perhaps prevent those care gaps in the first place. It sounds like meeting members where they are considering the social determinants of health. Many of these strategies, um, they make complete sense. And when I think about scaling those strategies, I'm sure that's where some of the challenge comes in. But we've. So network optimization and that partnership with medical and case management.

Super interesting to hear of. Um, any other strategies maybe that we, we glossed over too quickly or anything that anyone would add where you've really seen, whether it's a piece of technology, an approach, a partnership that's really moved the needle in, um, achieving follow through like we're speaking to here. I think that there are, um, it's like a three pronged strategy, right? So it's, it's community based partnerships, um, which, you know, the local organizations, faith groups, you know, service, social service partners, those kinds of organizations. And then there's the integrated care teams. Um, and when I say integrated, I mean truly integrated community health workers, behavioral health folks, nurses, etc. and, um, sort of that pro active outreach models that make the most sense. So I think you have, it's from our perspective, it's like that three prong approach to strategy. It is not just one one leg.

It's all. It's a leg and a couple of arms, so to speak. So love that. All right, well let's go. Well, I'll tell you because. Because I also have three. But they tend to they're a little different than the three that you had on there, Rhonda, which are all are all great. But I was thinking ours are more like aligned incentives. How do we construct the contracts so that we actually have win wins?

Um, that for, for everyone. And we found that the best contracts are really two sided, two sided risk. Um, they turn out with the best outcomes for patients, um, structured collaboration that we come in to say, how do we work together? Not how do we have us versus them and then actually put the infrastructure in place so that we can have regular ongoing meetings and sharing with the, the healthcare delivery systems that we partner with. And then the third is sharing data. The. And something that Robert spoke to is the more effectively we can get to real time data and information sharing, the better off the partnership goes because we're all working off the same set of assumptions. Thanks. Well, thank you both so much for for those hook ons.

Really insightful. And I think it leads us nicely into where I wanted to take our discussion next, which is to go a little deeper on AI, as of course, that is really the heart of our conversation today. Um, we know there's of course growing interest in using AI, but not just to analyze information and generate insights, but we're also seeing growing interest around using AI to support decisions and drive actions across workflows. So I'd love to know from the group where you're seeing the most realistic near-term role for AI in value based care, and where should leaders be cautious about applications as well? I can start, but I can talk for an hour on this subject. Um, but we are seeing it being used fairly effectively, um, in, in spot areas. So this notion of a large language model and a super agent taking over health care and doing things that are large scale are pretty much not where the industry seems to be going, but in very targeted uses of AI that might help in the structure of things. So a good example might be to use an AI agent on an intake of a claim that comes in, you know, from a provider to be able to do things with the claim before it's adjudicated, you know, to improve it. And what that's going to do is improve throughput, improve the kickbacks, the pins, things like that.

So we're seeing use that way. We're seeing it used to help service agents that are accepting calls from, you know, members and providers being used as copilots just basically doing Google on steroids. This is not the high risk stuff that you would get with some of that. And so our model there, and I think you've got it later, which is the governance of AI. How do you manage AI? Um, because I sort of lead teams of technologists. I have to sort of abstract everything. And so I look at things like generative AI, like nuclear fuel. It's plutonium.

If you hold it in your hand, it'll kill you. The only way you can really use it effectively is put it in some sort of containment object. That's how we do it in nuclear submarines. We can contain it and just use a bit of that energy to support it. So what we're seeing is more of this encapsulation of AI in workflow. Um, that's the basic way to describe it because in the workflow case, you have all the, the auditability. What's the data that I'm working from? What is all the specifics that don't allow the LMS to just go off in a space somewhere and create something? This is more of what we call a knowledge graph to be able to give it the guidance to create some phenomenal output, but only when it's well controlled.

And we're seeing that being used more and more and more. So the bigger organizations are looking less at transforming using AI, but applying it in places that are transformative but somewhat limited exposure for now. Um, we do use it in predictable analytics with those big databases that we talked about earlier that Amy was mentioning. This is definitely there, but the actions that are triggered by these other elements are what we're seeing going in now and starting to be used more and more. So for the the most vulnerable population, like the members that are that we see on the Medicaid side, the, you know, AI is a, is a tool. And, um, we see it that way, but the human interface for them provides peace of mind and connection. So we can't lose sight of that, right? So in my mind, AI has enormous potential to really improve, um, value based care, particularly when it comes to identifying patterns, predicting risks or, you know, prioritizing interventions, but it's most powerful, um, when it provides data to support the decision making. So for example, um, we're using AI to identify at risk members, um, to predict, um, hospitalizations and prioritize, you know, the outreach to those members.

Um, we do not have an over reliance, so to speak, on the model. And, um, we're continuing to provide transparency in our, our decision making, which is a part of what I think he was referring to in terms of the governance. Um, but overall, I think that our position is that AI should, it should help our care teams, um, focus their, their attention on data, but the relationship with our members still remains personal and human to human. Yeah. And in that case, it almost acts like the copilot for the social worker or the community health worker that's there and saying, I can help you with data while you interact with this person. It's almost like what we see in the ambient AI that providers are starting to embrace so heavily right now, is that ability to take off that workload of documenting the encounter. You know, that that that thing is happening is we see that spread more and more. It's actually making humans do more of the human job instead of the computing job of recording all this stuff in an assessment and trying to understand all that. So we've been using this in behavioral science for a while just to record the voice itself.

And what can we learn from the tone from all those things like that? This is, I think, the combination of ambient AI with some of the workflow orchestrations we're seeing with the other AI tools can be really interesting how it's going to affect us over the next decade, I imagine. Yeah. At Humana, we're taking a very similar approach that to what Rhonda and Robert have outlined. We see AI as an enabling technology, not a replacement one, right where it works hand in hand with our teams. And an example of it aligns really closely with what Robert just articulated was we implemented it in our call centers where it's monitoring the call. The AI is monitoring the call real time and providing information so that we can respond and provide accurate information just in time. Instead of saying, I'm going to connect you with someone else or I'm going to I'll call you back when I pull that information. It's it's it's complimenting the care experience versus replacing it.

Right. And I think that's a critical element in care delivery. I'll also say, you know, tied to what Rhonda was saying earlier is AI is really good at servicing gaps, right? Identifying patient risks, right? Making sure that we do proper follow up and integrating those kind of prompts into the EMR. The electronic medical record makes it easier for our care providers to respond. Similar to that example on the with the call center, I just gave the next piece. I see it really being valuable is in reducing administrative friction, right? We can leverage it to take care of some of the things that we don't really want to do, but have to do to make the system work effectively, like assisting with routing of referrals, right?

Insisting with care transitions. Right. Um, suggesting alternative pharmaceuticals. Uh, for the providers to, to consider. Right. You know, different things like that are really helpful. And then the last one is, I see it as an expanding of capacity, right? We, we have a limited care providers, right? When we have an almost unlimited amount of care needs, right?

And so this technology can be really effective at helping us with outreach reminders, uh, follow up, you know, providing the right documentation. So it's really exciting times. Um, with a cautionary note. Yeah. I have to say that there's a big through line coming in here that your hope, it seems like all of your organizations are, are viewing AI tools as able to be applied to a diverse range of use cases, really. But with members at the center at the forefront, it is a means to unburden the clinician, unburden. You know, the social worker, whoever is serving the member so that they can focus on those face to face or more personalized interactions with members and care teams. Um. And Oscar, we are, um, even.

Though we're a health plan, we also have a very robust tech, uh, side of our business. Uh, you know, we build a lot of tech solutions. Oscar actually started as a health tech company, so we are finding some very creative and innovative ways to use AI, um, both on the business side as well as on the clinical side. I can give briefly give you two examples, um, of how we're using it for decision making and value based care. So one way is through proactive care management and risk reduction. I think others have already spoken about this, but we're exploring ways that the AI agent can monitor member data for chronic care gaps to identify high risk individuals and also to automatically alert clinicians to initiate interventions. Um, we're also looking at ways to leverage AI for what we're calling intelligent financial and claims management. So with AI tools like super agent, we believe that we'll have the capability to audit medical codes, you know, handle prior authorizations and even identify documentation gaps that accelerate reimbursement. And these are the types of capabilities that can reduce the manual effort that most of us know is involved in managing value based incentives.

I think you guys are all using just fantastic examples of how AI is going to change care. You know, taking a step back, um, I, I, it's, it's good having some, uh, long history behind you. I became a physician in 1990. And when I, when I, um, got my MD, you know, there was a person who needed help and a person who could help, like right next to each other. And then over time, we just put, you know, electronic medical records, prior authorization, narrow networks, formulary decisions, and like just kept putting all this stuff in between the person who could help and the person who needed help. And that complexity has gotten so frustrating. And I think all of you have spoken to the fact that, um, and you can tell, I think in frameworks like AI can do two basic things. One is to do things better and the other is to do better things. And the doing things better part is, is starting to, to take away that administrative complexity that you all have spoken to and put it in the background through automating and a lot of what our head of it calls is, um, does, does the, the stupid stuff, like we put all the stupid stuff behind the scenes so we can do that kind of matching, but then we also need to think about what are the better things we can do because we have AI.

So we can do things better is through simplification, automating, um, making, making those phone calls easier because the call center reps have the information at hand that gets surfaced at the right time, right? That's doing things better, but doing better things is rethinking how we, how we, um, help people access, navigate and pay for care. And accessing care is like when they go on our website, can we actually predict what they're calling for? Are they calling for heart failure? Are they calling for for their knee pain that they need to get a joint replacement? And we can like surface the right information and maybe not calling, maybe they're calling online. Can we surface the right information and the network of people who can help them, um, with their particular need? AI can do that with the access world navigation, like ensuring that we know who's in network and, and bundling together things like prior authorizations that that they might need because we know what the typical pathway would look like. And so why make it hard for them?

And then payment, like it's so frustrating for patients to not understand what things cost and what their cost share is. And by the way, what's the difference between deductible and copay? Oh, like, um, so, so like really simplifying what that user interface is for healthcare is I think what AI fundamentally can do. And it's, it actually, it's rehumanizing healthcare. It's getting, getting back the relationship based care that I always loved as a primary care clinician. Um, and, and, uh, taking what is frustrating for people, you know, but here we are coming from the plan perspective, but taking the plan out from between that and instead being a support to that relationship. Part of my work works in the world of AI orchestration, not the AI itself. We've got it everywhere. But orchestration looks at that experience over time that every member has.

It's where all these actions fall through the cracks or could be predictable, but are not things like that. We don't have that data that really happens in the ether out there somewhere. But what we're seeing AI do is starting to use tools, marketing type tools, maybe developed in other industries that are increasing those touch points. So using our marketing channels, our outreach, our digital interactions, if we could bring those together. So it was like one brain orchestrating a lot of that stuff, not doing the agent part's not doing what Oscar does or doing what everybody else does in their own agentic worlds. Those become part of the things that we do, but the experience is greater than any one agent. And so we're focused on that as sort of that next level. So as we do look at those, we can predict better, we can tie them together better and really get what we're doing now is getting beyond the departmental level solutions. We're really starting to look now at enterprise or even enter enterprise solutions in healthcare.

Interoperability brings that to the fore. So some pretty exciting times there. But it is starting to look at the problems in a higher level way. And this is where value based care lives. It doesn't live in a business. It lives across businesses. And that's where I think AI will help as well. Yeah. Robert, thank you for those additional comments.

As you were speaking, I was starting to think, um, how Agentic really is. I would say newer in terms of what organizations are implementing. You know, I know it's not a new technology, but we're, I think it's earlier days compared to other technologies that plans are leveraging. So how and Robert, I'd love your perspective on this as well as well as everybody else, but how do you know that these tools that you're investing in are actually making a difference in a positive way? You know, how do we know that work isn't just getting shifted from one area to another? What are some of those markers or signals that that tell you that it's an effective intervention? So the approach I've seen the most successful is what are you doing with the AI, where most people focus their attention on measurable KPIs that they want to improve? So you start there. What am I trying to improve?

Not what AI am I trying to test or experiment with. And that will lead you to a few good spots right there. So if I just simply improve this intake so that I could basically just summarize the complexity of the case, like a claim that I could get it in the queue at the right level so that the adjudicator has enough time to complete it before the SLA is triggered. Little things like that are where are immediately measurable. So we saw a large payer. I won't mention their name, but they did this in their one claim segment and saved like $48 million in one year. Um, simply because they prioritize the work queue of their highest paid experts and said that actually had a great payoff. So that kind of thinking is what we're seeing more and more of is using AI. Set up a successful example of it first.

You know, look for something that has value and then you build up from there. We're seeing more of that thinking when it comes to it, instead of let me just solve the business problem and then look for those repeatable things that they can do. So using AI in that case, and then just repeat it in another business unit that uses the exact same thing for compounding value. So I'd like to build off of that because I totally agree with Robert of, you know, look at your key performance indicators, build around that monitor, adjust all that sort of thing. And organizations often go to the obvious, right? It's like, hey, I'm trying to reduce turnaround time, right? And that's, that's hugely important. I want to highlight an area that probably doesn't get as much attention as it should, particularly with AI. And that is ensuring that things like hallucinations, silent failures, um, uh, bias, you know, Rhonda has done a really good job of emphasizing equity in this conversation, making sure that the models that we're building this don't reinforce inequity, right?

And so those are harder things to build measures into the system, but they're no less important. And so I just, I would like to challenge everybody to be thinking along those lines when they're implementing these types of solutions. Such great points we have. We have two themes on that human in the middle. You know, those decisions ultimately have to be arbitrated by human and not AI. And then just, you know, limit its its use how you're doing it that way. But we find that to be one way to sort of reduce those things, especially the hallucinations that seem to be less and less in that that world. Super helpful. Thank you, Robert and Rhonda, did you want to add on to anything that Brian mentioned?

Yeah. So, I mean, I think everybody. Um, I think we're all on the same page and I think we are reinforcing each other's thoughts and the work that we're doing. And for us, really, it's like, how are we measuring that? We're using AI to, to a point where it adds value, right? Just not using it because it's a trend or just not using it because. Everybody says we need to use it. Like what are the outcomes? And I think it's, you know, things like, um, better provider engagement, improved member outreach, um, things like making sure that their caregiver gap closures, um, rates are improving, that there's avoidable admissions, all those, you know, follow up compliance and member engagement rates are all kind of reflective of the fact that we are investing in AI for the, for the outcome outcome, not just investing just because people say we should invest.

So. Absolutely. Thank you, Rhonda. And I know that we've also touched on governance already throughout this conversation. And Brian and Robert both just spoke to a lot of some of the cautions that or I guess, advice that they've given to other organizations around. You know, considering bias, hallucinations, other ways to really govern AI and use it responsibly. But we'd love to hear from the group. Any other, uh, maybe more granular pieces of insight, um, how you're approaching transparency, auditability, accountability, and any other watch outs that you would share with peers as, as you've developed your own AI governance framework or. I know a lot of organizations are also currently working on this too.

So any other lessons learned here that would be helpful for our audience? Well, this is part of our business, so we wrestle with it every day. Um, and how do you manage that sort of aspect of introducing this nuclear technology into your business in some sort of meaningful way? Um, and it's a, it's a juggle, quite frankly, because you mentioned bias and that's probably one of the most, you know, interesting terms because, you know, bias could be good or bad. Um, it's just, you know, how are you applying it? When and where you're applying it. And I, I know that Ron was talking earlier about. You know, especially in the, you know, the Medicaid communities where you, you really have to. Worry about somebody saying, well, I will evaluate your projection of outcomes.

And I may use that to actually influence what I spend on doing stuff for you. It's a cold thing, but it's a fear out there. And it could be the kind of thing that it could be, you know, just naively done by AI. It's not some intentional bad thing. It's just that that concern. So our answer there is not an answer to the problem. It's to say you've got to audit everything. Because the way this actually turns out is in some sort of situation where you better know what happened. Where did AI make the decision?

Was it done? The worst thing you could do is let it happen in a black box. So what you'll see us talking about is that AI can't work that way. You have to put in a glass box that says it's incorporated in something that's completely visible. I can see what was asked of it. I can see what data it was fed to make that decision. I can see its series of decisions and information points that drew upon to make that decision. Um, by doing that, at least you have a defensible thing if this is challenged. So it's less about preventing it from happening, although you can put all kinds of.

You know, lead boxes around the thing. Um, there's still going to be stuff that gets through. So that's our approach in the end. Don't let AI make the big decisions. Put a human there and say, are you responsible for this? I'll do all the work to make it as easy for you to understand as I can. But in the end, um, we can't prevent that from happening. So that's sort of the theory that we have with it, which is more operational. So we try to help the teams as we work internationally, it becomes even more critical to have this sort of perspective, because they're a little bit further ahead of us in some ways, especially when it comes to this sort of level of security.

Robert, I'm going to be quoting you forevermore on we can't use the black box. We have to use the glass box because it's exactly the right phrase. Um, you know, something else that we're doing is because AI is so ubiquitous. I mean, I don't know about you guys, but on my personal phone, I now have ChatGPT, open evidence, Gemini and Claude all downloaded because I use them every single day. Right? Everybody we know is using these tools. And so how do we make sure not just through like organizational principles that we have oversight and responsibility, but we're actually rolling out to the, the organization AI training on how to use AI responsibly because we know it's, it's out there in the wild. And so we can't, we, we have to ensure that we're using it effectively and responsibly with some guardrails around because like the, the train has gone way down the track and we just need to make sure that as we use it, it is used in a way that is effective and safe and compliant. Yeah.

To that point, um, Amy and Oscar, we spent a lot of time last year focusing on our governance model around how the AI tools are used, um, to validate BBC results, outcomes findings for our providers, because one of the things we did not want to do is create any provider abrasion. Um, we, we did have some problems or saw that providers were challenging some of the purely algorithmic AI tools. So when you're looking at, you know, EDI analyzing EDI visits and, um, looking at, you know, inpatient stays and using these tools to calculate, you know, what a shared savings, uh, percentages and, you know, even ibnr it, it can create problems if you're not transparent. So we also wanted to make sure that we could share the information as much as possible, be very transparent with providers, give them an opportunity to provide feedback to us. Um, and we wanted to assure them that just because we're using AI tools, um, you know, these. It's not taking away from medical decision making. Um, we're not questioning in any way the judgment of our providers because we do respect them. Uh, but it has led to a really strong partnership with our providers because we have been so transparent and because we have documented our governance model around how we use AI tools, uh, to make decisions for BBC deals. Um.

And I'll just say one final thing, because, you know, this is where I put my left and my right, um, brain together because I, you know, I have that the creative side and I have the, the lawyer governance side, um, and the compliance side. And so it's significant to me and my team because we need clear like oversight structures as it relates to this stuff. So cross-functional review committees, um, you know, transparency, accountability, making sure that humans are maintaining responsibility for final decisions and ethical safeguards. And this is where Brian was alluding to, um, my reference to you. No bias, no bias monitoring and regulatory compliance. All those things are really, really, really important because, um, I say, you know, human, human in human out, meaning we, we input the data going in and we make decisions based on the modeling and there's no lack of human touch in the process. And so having these review committees and setting up the governance structure is, you know, what's absolutely needed in order for this to be successful. Um. So well said, Rhonda Monique, I really appreciate all of these insights on governance, especially because I've been sensing kind of this element, especially Monique, when you were speaking to, um, it sounds like there's this transparency that you're, that Oscar has with providers, especially those in value based arrangements, transparency around how AI is being used.

And, um, it's interesting because I think that we hear so much about transparency in the patient to provider relationship, but not so much. What is that AI? Transparency look like amongst the organizations and stakeholders in patient care. Um, so really excited about where where this is headed. And speaking of, I wanted to also before we close get the group's outlook on the next few years ahead. You know, I feel like when I used to pose these closing questions just several years ago, I would say five years from now, what do you think? But I think that realistically anticipating what time will look like that it's narrowing. So we'll we'll look ahead a year or two from now. I'm curious what you all think.

Um, where you think value based care execution is really headed and what new expectations do you think that will create for how payers and providers are collaborating? I can start on this one. Um, you know, and it probably comes from starting out in practice at, at Kaiser Permanente, which is, you know, so I practice value based care for over 20 years and, um, see it as values based care. It is the way to pay for care that aligns what clinicians are doing and the payment infrastructure. Um, and I do think that AI can do what Monique was talking about, which is, um, rather than have the healthcare delivery and payers be at loggerheads with each other and be adversarial, um, when used appropriately, it can actually reduce the friction and allow us to deliver on a better experience for the members and patients and consumers that we all care about. So I, I see a blossoming of trust enabled by AI through better collaborative practice in the future, through using value based care as the tool. Well, I hate to go after Amy because she's always so articulate in her answers, but I'm going to risk it because it ties to what she said. Uh. I believe in my heart of hearts, the path forward is value based care and doing it more effectively.

So it serves more of our communities more effectively. Um, and to do it in collaboration. And part of what that means is we've got to acknowledge where it works and where it doesn't. And part of where it's designed primarily at the primary care level, because that's where the bulk of the care happens on the prevention side. But there's a lot of provider groups out there that have large hospital systems that are in rural communities that are heavily specialized. Right? And most of the incentives aren't designed around those groups. And so where I see a great opportunity and evolution is us making sure that we look at all the different types of configurations of our provider partners and customizing so that there's proper incentives to do the right place to provide the right care at the right place, at the right time. And, and I'm excited because we're actually starting that work.

And we've got a number of additional incentives that we're working on right now at Humana. And so I see some great progress in that regard. The other piece that I hear in this is it's been touched on is, and how does technology both interoperability and which we haven't talked about as much, but is equally as important as AI. Right? And how does that information shared to allow us to do that more effectively and, um, make these systems serve our communities better? So I'm excited about the future. I recognize the challenges, but I, I'm confident we're moving in the right direction. And I'll be brief because I know we're trying to wrap it up, but in my mind, value based care ultimately comes down to trust. really is between trust between the payers, the providers, the members, the people, and the systems that deliver the care.

When our members need it the most. So mostly I want to leave you with a thought that, you know, technology should accelerate accountability but not replace it. And our biggest challenge going forward lies in our ability to help people overcome the barriers that prevent their care from happening in the first place. So well said, Rhonda. Monique. Robert, any final thoughts before we hop? Sure. I'll just say briefly, um, you know, one of the things that we're focusing on is differentiation and also having really innovative programs, value based programs for our members. Uh, so we're, we're actually looking for where we have put out RFPs, we're rfis, we're looking for providers who actually have some very innovative programs where we can partner with them and have BBC or alternative payment models.

So, you know, we recently introduced a women's New Women's Health program, our hello program, where we have these value based agreements that have customized approaches. Um, and just different ways of serving our members where they are. So we will continue to do that moving forward. We see that as the future for in healthcare and also for Oscar. But I will say on the other side is that we're also focused on striking the right balance because we're in the a, C, a between fee for service and value based deals. Um, and we see it as, you know, being in a river where you have to have one foot in a fee for service canoe and the other in the value based care canoe. We have to strike a balance in order for us to stay afloat in the ACA marketplace. Tough balancing act. Monique.

Thank you. Robert, would you like to close this out? Yeah, absolutely. Um, love the conversation, though. That's from everybody. You know, this this world of AI and how value based care are sort of coming together. I think what we're going to see AI impacting is it's not going to change the complexity of these, these systems right now. It's not going to take the pain away. There's no magic bullet here, but what it is starting to do is making it easier to make things better.

Um, and that's where it starts. And we're going to see as you know, value based care evolves. I think the technology is going to enable it to get from more of the shared savings types agreements, maybe into the earlier ACOs, but now into the real, bigger aspects of population health. That's what I'm excited to see this start to, to move so we can start more proactively identifying and helping people. I mean, and we've got to do it soon. My age is the point that I really need this now guys. Let's get moving. Robert thank you. And thank you to the full panel.

Your enthusiasm for this work comes through so clearly in this conversation. So I want to appreciate all of the time that you spent with Beckers today, with our audience and all the insights that you've shared with them. Um, and of course, we'd like to thank Pega for sponsoring the session today. And audience members, we hope to see you at another event soon. Thank you so much again, everybody. Take care. Thank you. Thanks, everybody. Thank you.