PegaWorld 2025: Improving Healthcare Experiences: Driving Impact Through Increased Scale and Autonomous Optimization

PegaWorld 2025: Improving Healthcare Experiences: Driving Impact Through Increased Scale and Autonomous Optimization

All right, guys. Welcome, welcome. Last few people. Take your seats. Thank you. Sam. Hello. Uh, I've been asked to do a little bit of housekeeping just to kick us off here. So if you're here for the CVS fireside chat, you're in the right place, so. Well done. If you're looking to fill a script, you need to go back to the hotel across the street, and you'll go to the CVS pharmacy out there. Um, so I'm Christian Davies I'm the vice president of decisioning at Blend.

Um, I should probably tell you a little bit about Blend first before we jump into our great chat here with our with our with our panel. Um, we're a specialized 1 to 1 partner of of Pega. Um, we help clients like CVS Health just get the most out of their out of their marketing.

Um, we leverage data.

We leverage AI technology.

People also, you know, the key. The key, the key ingredient in the recipe, um, really, just to get measurable business outcomes for our clients. So we're in booth nine this year in the innovation hub.

Some of you might have been down there. If you haven't please check it out. We've got a we've got a pretty cool demo. Um, it's been a bit of a sleeper hit. I would say it's, uh, using AI to find out what kind of coffee you are.

So it's a bit of fun. I'm a flat white. I'm not sure what it says about me in particular, but people seem to like it. Um, and it's a bit of fun. So go down and see the guys, they'll be happy to talk to you. And, you know, maybe you can find out what kind of coffee AI thinks you are.

So on to what you've all come here to to talk about. So I want to introduce my panelists here. We've got, uh, we've got Rahul Kak. He's the vice president of customer health and experience, and we've got here Alpesh Alpesh Taylor.

He is the executive director of data engineering.

Rahul's been with CVS for the last seven years, but in this space for 15 years. Um, and Alpesh has been helping CVS with kind of all things data for the last eight years or so.

Um, and, and what are we here to talk about? Right. So we're here to talk about CVS journey, CVS Health's journey with Customer Decision Hub kind of where they've come from. They've come from a customer, a custom outbound solution which has worked great for them in the past.

It's unlocked an absolute ton of value for these guys. Um, and it's home. It's home built, right. So it's kind of built specifically for those guys by those guys. Um, and it's very sophisticated. It's, um, you know, it allows you to do very highly targeted campaigns. But with that flexibility, um, comes from cost, right? And the cost is highly sophisticated resources to run it. It makes some of these campaigns quite, quite expensive to run. So what's what's our goal? What's the goal of the project? Well, the goal is to reduce execution costs.

And it's not just to to save cost in and of itself for the sake of just saving cost. CVS has got an absolute boatload of innovative ideas on how they can tap into a well of customer value.

But historically, they've kind of not had that return on investment, right, with the with the cost of executing these campaigns. So if we can use CDH to bring those costs down, we've just got a whole wealth of opportunity and customer value to tap into.

So kind of above and beyond that kind of immediate and big and exciting goal, right? We've got other capabilities that Pega offers, right? True customer level decisioning optimization using adaptive models.

So you know what we really want to do Kind of going forward is kind of use dial back the business rules and really let I kind of get to work to get the right message in front of the customer.

So with all that in mind, let's dive in to the interesting, exciting stuff. So, um, give us a kind of understanding of the history of CVS health, the marketing outreach that you guys do, and kind of how did we get here? Yeah.

Hey, guys. My name is Rahul. Um, thanks for coming. During the peak of food coma, our, um. So we have a wheel there. Before I talk about that visual, I'll take. I'll take a minute to talk about the history of what we've been doing around what is not marketing.

Although I actually sit in the organization, the marketing organization of my company.

Um, and what I describe more as population health. So, so the fun stuff we're doing with customer decisioning and marketing technologies, it's not related to your traditional acquisition or retention.

We're trying to change health behaviors.

years, and we started this program about eight years ago when we were looking at sort of the. I think wealth of data that we had, and also a very large investment we've made in data science, data engineering resources to activate a lot of that data.

Um, using a lot of machine learning technologies to help do very precise targeting and outreach. And we wanted to build a nudge program effectively. Um, and at that time, eight years ago, and we built everything from scratch or like some enterprise platforms that are sort of pulled together. So we've kind of like evolved the maturity over the years. Um, but the recognition of that vision very at that point in time was like, we're doing something very difficult. We're not trying to get people to buy a widget or click a click a link. This is trying to change their real life health behavior, health behavior that we can measure in claims data, um, prescription data, things like that.

Keep in mind, um, you know, the scope of our company is not just CVS pharmacy, it's also medical benefits. And though I would say the wealth of our data is actually not from RX, it's actually from the medical claims, which is a much larger percentage of, of health cost.

Um, so so what we decided to do when we built that, we said we're going to need to build a program that has a lot of, um, takes borrowers a lot of principles from agile product development.

It's going to have a lot of iterative test and learn, and we need to get really smart about what's working. So then I go to this this figure, this wheel here. And so the first thing that we wanted to think about was, um, that we always think about. And we're building our portfolio of behaviors that we want to change. So I'll give you some concrete examples of, of behaviors.

So one would be around um treatments. So educating people on vaccines screenings that they want. We want them to get um site of care selection. So thinking about what's the difference between going to an urgent care versus a retail clinic versus an ER for my acute need, there would be something around condition management. Um, I think it's something like two thirds of adults in the United States, or a percentage much higher than that, even have at least one chronic condition. And it would be an example, like how do we educate people on how to manage their musculoskeletal condition or sinusitis or diabetes? And what are the actions that they'd want to take? Um, some other examples would be hospital readmissions. One of the classic quality measures that not just insurance companies, but also hospital systems have are what is the rate of readmission into a hospital within 30 to 90 days after a discharge? And what are the what are the actions or the nudges that we can provide to our members to keep them from getting readmitted for for something unnecessary? Um, other examples could be on behavioral health, which since the pandemic has become an increasingly hot topic.

So the things we could help on would be early detection, prevention. What are the resources that we can provide? Um, there's a whole category around women's health. There are things such as, um, endometriosis as an example of a very commonly underdiagnosed condition. Uh, and then when you have a lot of the data that we have, it can maybe even inform, um, there may be there may be a higher risk for someone to have this condition, like the one example I gave.

We might encourage them to see a specialist, because sometimes your pcp's won't necessarily catch certain things and so see the specialist. And just to, just to play it safe. Um, and like any healthcare company, we have a whole bevy of different care solutions and partners that we can get, that we can provide care management solutions.

Um, but people have to enroll in them. People have to actively engage in them. So then you get into maybe some more traditional marketing type stuff where you want to activate people into that.

Um, moving around the wheel, uh, the second category in the top right is around behaviors and facilitators.

So you want to think about when we're building a health population health program, what are the what are the topics that people are not aware about, that maybe they're not aware that this is something I should be worried about. Or maybe they're opposed to the idea of vaccine hesitancy is a common thing. Um, maybe they're incentivized or they think, I don't know what's in it for me to do this health thing. Move around the wheel. Another step. The bottom right. There's actions. So then we have a lot of practitioners and they think about all the time what are the tactics that we're going to be able to use to get them to change their behavior.

So there's of course your reminders. There's incentives like gift card type programs for trying this. One thing. There's behavioral economics.

There are um, there's support that we can provide in scheduling, decision aids or treatment aids. That could be helpful when you go to a clinician. Um, interactive digital experiences, all sorts of tactics that we're going to be testing and learning to try to change the health behavior.

And then you move forward to the bottom left corner. And that's the last part of the wheel. And that's on automated outreach. So you can think of traditional outreach that we can use to nudge people. And that would be things like direct mail, phone calls, both inbound outbound IVR, interactive voice response type stuff. There's digital channels you can think of email, SMS, web app, and then there's also relationships that we have. And the nice part of the sort of space that we play as a vertically integrated healthcare player in the ecosystem and said, we'll have knowledge and the resources to, um, uh, leverage clinicians, doctors, pharmacists, care managers, care managers that will be on staff.

Obviously pharmacists, that'll be on staff to be part of that behavior change process.

So those are all of the that's kind of like our little you know, we'll of like all of the things we need to think about and the complexity around it. Um, and then uh, and then there's all the stuff about how do we make that happen, which is like the, you know, the technologies that we need for that. So I'll probably in a minute, I'll probably spend more time talking about like, what are the capabilities that we need in sort of the challenges we need to solve and to make making all of that a reality. But I wanted to set the stage of like, what is it we're trying to do? What is the vision? Which in a couple of words is just like changing health behaviors.

Very, very difficult. I'll, I'll feel free to add on to that. Yes. So hi everyone. I'm Paige Taylor and I lead the data engineering and CVS Health.

Um, so let's start talking about data.

Actually, um, we as May 2025, I think we have close to 9071 71 stores within the United States, and we cover almost 85% of Americans, which lives within ten miles of CVS Health. So overall, through this MBA program, our CVS ambitious goal is to be part of every American household health care journey. And how do we deliver this? Personalized messages to each and every American is where we are doing this MBA program. Um, MBA is next best action. Yeah, that's one I didn't define the term, but one of the terms we'll use for changing health behaviors.

What's the next best health behavior we want to change? Yeah.

Sorry. Go on. Um, and very simple example I think um, when we reach out to our members about, um, their preventative care.

Um, I think they diagnosed early in the stage and they help like the system, basically help them lead their diagnosis and they take care on their personalized healthcare journey.

Um, with this, uh, marketing, um, MBAs, we want to scale this personalized, uh, interventions across different line of business as well.

Um, and we want to make sure that we deliver a message at the right time to the right people, uh, to make their healthcare decision at a time.

Um, we, as I think Rahul mentioned as well, we embarked on this journey in 2017. Um, and we have scaled our MBAs, um, almost 400, 500 MBAs that we run, um, per year, uh, through this behavioral change, um, campaigns.

Um, and we are um, I think, um, we have use at that point, a lot of, um, machine learning capabilities. And as you can imagine, we have a vast amount of data at CVS Health. How do we chunk through this millions and billions of transactions to get to that? Right. Decisions. Right. Personalization message to our members.

Um, we build this customized solution to that fits our our supports, our MBAs. And we experiment a lot of a b testing through our this 400, 500 campaigns. Um, our data science team and data engineering team are very closely integrated to support these campaigns.

Now, as you can imagine that we are at the point where the cost of launching this campaign has not yet decreased. And that's the reason we are here to talk about how we're going to scale this program of MBAs and how we communicate to each everyone in their personalized healthcare journey.

Um, right now, we want to make sure that we, our team, the data science and data engineering team, can focus on more innovation and new capabilities and opportunities for our members, rather than spending time to deploy and maintain all these campaigns that we have been launching.

So thank you. Um, so you've you've outlined some pretty lofty ambitions and some goals. What makes Pega the kind of right tool to kind of lead this transformation for you guys? Yeah.

So I'll start on a few things. I'll build on our previous point about like and what you had mentioned about sort of decreasing the cost or what I'll sometimes call the activation energy of building a campaign. So let's say like right now, like I could, I could estimate the different types of practitioners who are involved in building a campaign which could be member engagement strategists.

There's marketing operations people or internal operations people.

There's creatives. This could be like digital asset creators, designers, copywriters, um, there's marketing technologists like architects, developers, coders.

Then you have your data engineers, data scientists that alpesh mentioned.

So you have a lot of different skill sets that are involved. And if I were to chunk out the approximate amount of time all of those skill sets, people are involved, the amount of time it builds, the amount of time that they're spending to build a single campaign from ideation through development, through a lot of internal iteration of getting it right and getting it approved internally, um, to execution and then ongoing measurement.

I'm not going to share the actual number, but like we've sized up, let's say it costs, maybe it costs like $200,000 to launch a campaign. Just illustratively if, if if you know what that activation energy is, then you need to. If you want to have like A3X ROI, you need to have at least $600,000 of return on the program that you're launching. However you're measuring that return. Maybe maybe it's medical cost savings, or maybe it's like a quality bonuses from the government or all sorts of like clinical outcomes. But you have to find a way to measure it, and then you end up creating and again illustrative numbers here you're creating a very high bar. So what we want to do with technologies like Pega and partner, we're working with partners like Blend. We want to reduce that activation energy and unlock value for this long tail of health campaigns and behaviors that traditionally, we wouldn't be able to touch this long tail because we're focused on maybe our big blockbuster things of the 20 largest psych spaces of like health topics, let's say. So, um, so that helps us activate the rest of the value there. Um, other key things that this enables? Uh, I'd say autonomy. So. Oops. I didn't mean to touch that part of my chest. Uh, as touching my chest as I, as a marketer, as someone who thinks about member experience and engagement.

Sometimes I want to change. Like the creative on a campaign. Uh, and maybe it's just something like, we're refreshing our brand and every company does it, even if it's a fluffy thing to do. These things happen. Um, and, you know, stuff that maybe our data engineers or data scientists are like, why are we thinking about that? Um, there's all sorts of things there could be, like legal considerations where we need to change the language, um, or the copy or even just disclaimers, and we say we need to make this change, and then we have to have a whole negotiation of prioritizing it in our big like, portfolio of campaigns.

And like, that means, uh, working with, like, our technical partners in order to work through their prioritization. And that's very difficult. And even though I would say we're a highly cross-functional team and we operate in these cross-functional agile pods, um, you still have to have that negotiation.

So if you have the technology that can have non-technical business users, um, launch a campaign more readily, set up an audience profile and put something out. Um, that's great because that makes us more independent to handle some of these use cases, some of which have business reasons, some of which have like, um, kind of more on the fringe reasons where we say we need we want to make an update here. Um, there's some of your more traditional things. If you kind of get the long tail of campaigns, we could build out more personalization. Um, solving for like, different, uh, do more tests of, like, different types of behavioral economics principles or different types of cohorts that we're seeing in the data, uh, based on like health activities or decisions that they're making.

Um, generally, there's a bias towards flipping our focus on here's a long portfolio of Alpesh mentioned something like 400 plus different campaign ideas or like health behavior topics.

And how can we flip it more to like a member centric area, like for member A, what is this? What is this the best, the best action to take for this person? Um, which is always the intent of a program like this.

But it gets difficult when you're building from like sort of a campaign view versus a member or a consumer or patient centric view.

Um, Alpesh mentioned it, but I'll give you this idea of like kind of rightsizing roles. How do we enable our technical people to be focused on their highest value skill set? If I take the example of data scientists, um, you know, right now, a lot of them are very much focused on like kind of the day to day, um, mechanics and logistics of keeping a campaign running.

And we would rather have them focus on strategy, focusing on experiment design, um, you know, building randomized controlled trials for and everything we do has randomized controlled trials. But, you know, focusing more on that. And then being able to focus on, um, uh, on on sizing new opportunities, things like that. Um, predictive models, of course. Uh, and then also autonomous optimization.

So being able to. Continue to build this portfolio, it gets to a point where it's so unwieldy where you just. Need so many people to manually handle all of these different campaigns that you have. So if we can. Kind of keep these things kind of running in an optimal way. And also like alerting if there like. Issues. It's a highly sensitive space. I mean you think of like HIPAA concerns and we're dealing with. Um, private health information. Um, it's a, it's a high stakes place to work in. Um, so you got to get it right. So those are, those are some of the things that come to mind for me. As I think you might be hearing about the campaigns, as I mentioned as well, our current operating model is always at the campaign level, and we always think about which member is going to be part of this campaign.

We never think a member from a member's perspective and Pega, which operates at the member level, it's going to be a key distinct.

I would say, from how we operate in today's world. As we all know, Pega is a golden standard for a real time decision making, which is very advantageous when we want. The ideal customer experience for our member is to learn from their recently available information and drive the conversation while they are on a call with our customer representatives.

You know what MBAs they are eligible for? What should I discuss next with this person while I am at. We want to think about more, I would say proactive approach and how we can communicate efficiently with our members.

And that's where I think the Pega comes in. A picture to make help us drive the real time decisioning.

Um, I think as we know that we have existing campaigns that we do, um, a lot of a B testing. But this I would say the learnings from each of the campaigns are right now shared manually across different campaigns. What to do next for um, I would say POCs, etc. but what we want to do is we want to remove the humans out of this process. How can we use Pega's out of the box capability to drive the right content to the right person? Uh, autonomously? And that's where I think we are partnering with Blend Blend team, where they are helping us drive these decisions on how we want to configure these campaigns and start thinking about the member level aspect of our campaigns, rather than still continue to focus about like, this is the campaign, let's figure out how which members are going to be eligible.

And that's where I think we have a key partnership with the Blend team as well.

Thank you. Alpesh. There's no secret that transformation is quite challenging, right? It's you know, I would imagine probably everybody in this room has been through one transformation or the other, and it's never plain sailing.

So, you know, I'd love to hear from you guys kind of what's your experiences and the kind of things that you've learned over the journey so far? Yeah.

So a few few categories. There's one whole category on change management, which is which is always difficult. I used to work in a consulting firm, and we had an entire practice around change management. And it's it's one of those things I feel like any, any technologist or business leader who's working with technology really needs to get like at least a one on one on one on change management. So that's why we wanted to start with. When I think of change management, there's one whole area around, um, you're always thinking this framework people, processes, tools. On the people side, organization is a pretty critical one. So we have these, um, traditional roles that we've developed over the past decade in our teams to run this program.

And now that we have a different technology that may have different skill sets that can change. So, um, you know, we had to think about, for example, when we brought in Blend what, uh, you know, and we wanted to bring in Blend because we needed like some of the skills and the expertise. We also needed some of the capacity, um, in order to maintain a current program without putting everything on pause to do a migration. Um, but then where does Blend fit into that, like organizationally? And what are the roles and responsibilities you could imagine creating a RACI for this type of a thing. So organization as a whole thing that we had to we had to be very thoughtful about. There's upskilling, of course, like getting our people to like build their skills. And when you do that, you know, there's there's even a whole question of like, what are people's titles on that? Like what do they take away from this? And even just how do you, um, encourage people who are like, and I'm talking about like individual contributors who are on the working teams, how do you get them excited about this? And, you know, I mean, at a basic level, you know, we should probably change your title because your role is changing in some cases.

You may have traditionally been an architect in this one platform, but now you're working with like multiple platforms, like there's something fundamental that's changing about your role in terms of managing leaders.

You have to sell the long term vision. Of course, you have to build your business case and sell the ROI, but you also have to manage expectations and kind of work people through the sort of the people being like our stakeholders, our leaders, um, you know, our like internal investors, what have you getting them, sort of bringing them along the journey because you kind of go through this valley of despair period, as part of change management, which means you start from a less efficient place than where your status quo is, and it'll stay that way until things get more efficient.

So until you get past that valley, things are tough and you kind of have to work through that and be comfortable with the problem solving. Um, there's a whole second category. If I think for us it's been starting small, um, an MVP approach I mentioned before, we didn't want to turn off our existing now very large, mature, high business value and clinical value program. We didn't want to turn it off to focus on a migration. So we focused on sort of like a continuous art. And we are still doing this, focusing on a continuous migration approach, which is to say, um, you know, at a point of a natural refresh and a natural iterative cycle of one of our campaigns or next best actions, we'll we'll start to bring that into the new technology and the new process working with Pega. Um, and so, so kind of starting that way rather than kind of like doing a large kind of step change migration all at once. Um. Let's see. I would say there's a whole category of thinking we have to do around capturing and cascading the learnings. So being very strategic about which is the first cross- functional pod that's going to work on this and what, you know, for what are the topics, is it lower risk, that kind of stuff, but then also capturing the learnings in, uh, in artifacts, process documents, building the process or building building the plane while you're, while you're flying it, that kind of a thing. Um, but really making sure that you're cascading those learnings so you don't have different teams that have to pick up that skill and kind of start from scratch each time.

Um, because if are like heavily matrixed model, um, we had to be really thoughtful of this. So, so when I say matrix like we are matrix in terms of like our, our technologists, that would be like our, our data science and analytics partners and data engineers.

Matrixed in with marketers or population health people. That would be like my team.

We are still different team organizations and we have all these different skills, although we like put ourselves together in cross-functional teams.

On top of that, these cross-functional pods, we also have different campaign working teams. There's kind of another layer where, like even those individual teams will have a whole portfolio of different, let's say, conditions are working on we have this musculoskeletal campaign or sinusitis campaign.

So it kind of list goes on. And it's like kind of a lot of cuts of different groups of people and different use cases. And you know, we just need to make sure that information and best practices and learnings are not getting siloed in any of these little pockets throughout. Um, so so this gives you a sense of some of the challenges that we've had to work through and kind of kind of plan for. Yeah. I think change is necessary. Right. As you can see, even the healthcare overall healthcare industry is moving from fee based services to the value based services, where the long term vision for the NBA is to move away from our reactive, siloed, um, interaction from to a proactive predictive communication with our members.

Um, and the change is not just, I would say, simply a technology implementation. It's a strategic transformation on how we want to serve our members and this community.

Um, when we talk about the challenges, um, as you can imagine, NBA represents more than just a technical capability.

They are actually a Blueprint, um, not the Blueprint of Pega. They're a Blueprint of our healthcare system.

Uh, that turns, um, I would say, all the healthcare messages into the smarter and more responsive system, where NBA success depends on more integrated data and within Healthcare system.

As you know, the data comes from a vast variety of systems and they are not integrated.

So how do we take that challenge? In today's world, when we activate these campaigns, our team has access to this vast amount of data sets that they work upon and build this hundreds of pipelines to activate these campaigns.

Um, how do we move from this hundreds of pipeline into one strategic data model? And that's the key question that we need to answer when we try to fit into the Pega, because the Pega usually, as I think we talked about, always work at the member level, and driving that data model is going to be a key strategic decision on how we can go from this flexibility of using all the data sets.

Vast amount of data that you have available in your hand to one data, which is going to be a little bit tighter, but you still need to grow your number of campaigns.

How do we do that? And I think that's the challenging question we are trying to address. Gotcha. So we are we're rubbing up against time here. So in one sentence each right. What's next for CVS. We've we've hit our medium term goal. We've unlocked that well of value that we're aiming for. You know what's next for you guys. Well one sentence. What a constraint. I'm going to repeat a sentence from before. But I would just say unlock more value by capturing the long tail of health topics that are too difficult to reach right now, the current model or our status quo model, I should say. And I would say right now, as we said, 400 MBAs, we are losing a lot of opportunity to interact with our customers.

We want to be part of that healthcare journey, and that's the goal we have launched. Fabulous. Thank you guys. Really appreciate it. So we do have roughly ten minutes for a bit of a Q&A here. There's microphones just I think in every aisle. So if you've got a question run up and ask away. We've got a couple of minutes for you guys here. And don't all rush at once. Thank you. Hello, this is Dora from Glidewell. I have a question around the practices. So let's assume you have always on actions. And I remember you mentioned 400 different MBAs.

And usually the greatest challenge is to really maintain those campaigns. Right. So ideally, if you can just have one campaign, you know, through all these eligibility rules and the arbitration process, you can just limit all your actions sent through the funnel. And one single campaign could be triggered. It's easier to maintain in terms of perspective. Right. So I want to hear more insight about your existing challenges and what sort of solutions you have provided for the CVS customers as the Blend company as well.

And how was this work together happen? So gotcha. So I can have an initial crack. So really what we're talking about here is there's 400 campaigns. We want to basically run one schedule, maybe once a day, and put the right thing in front of the right customer at the right time.

Right. So that's easier than having 400 different schedules that we're going to run and manage and operate. So that kind of lowers our total cost of ownership starts to drive down those operational costs. But how do we do it. Right. So one of them is obviously operations manager. But the the more challenging piece, certainly from my perspective and certainly from from your guys perspective is how do we know the right one to put in front of the customer.

So we know we've got these things called eligibility rules, right. We we want to strip those away to a degree, but we can't strip them all away. Right. We want to make sure that the, the things that they get are actually things that they should get, especially in, you know, in the HIPAA climate that we've got here where, you know, sending the wrong kind of email can have some pretty serious consequences. So really, it's finding the balance between stripping back the right kinds of rules. Um, but actually making sure that the ATM model works within guardrails and constraints. Right. So we don't want to we don't want to let them run, run rampant. But we also, um, we really want to use that as the kind of driving mechanism, the forcing mechanism really to to kind of lower that total cost of ownership. That's kind of my perspective. I will love it if you guys can jump in as well. Yeah. Your question seems to be more about what are the challenges on it. I mean, there's there's some unique challenges in our space too, that we have to solve for. Um, and I would say the seasonality is a challenge that any business has to face but ours. I'll give you an example. Seasonality for us would be, um, the Medicare business has an annual enrollment period, which starts mid-October, goes through the end of the year. And that is if you're a Medicare eligible all beneficiary. You're going to be hammered with all sorts of marketing from all sorts of companies.

You are a very high target prospect in a very big, very big market right now for Medicare Advantage. And we have to be thoughtful of that because that, you know, if we just set up an always on program and it's running year round and oftentimes it's triggered by certain health actions they're taking or not taking or predictive model like these are things that could happen at any point in the year. But then we have to be really thoughtful because it's a very crowded and noisy time of that period of the year. It's just one example, but it leads to a lot of interesting discussions within the company about like, what is the what are the business rules, how do you prioritize? And then, um, you know, any company is probably going to have this, but I think a company of our complexity, um, especially as a vertically integrated healthcare company, you're going to have some sort of you're going to have some competing, um, competing business values that you have to organizationally have your leaders come together and come to a decision. Objectives would be like in that one example of like an annual enrollment period. There's of course like there's like renewal and there's like kind of the revenue aspect of bringing in a member.

But then there's also the um, uh, there's like medical cost savings that come from certain health actions that are taken that are like clinical value levers, like stars. Haedus there's some like, jargon stuff for you, but there's like different teams that are associated with different, like quality metrics that we have to kind of balance.

So so it's it's definitely not easy. Um, so it's something we have to, we have to work through for every single one and certain things programmatically and certain lines of business, we have to have, um, um, some pretty thoughtful conversations at a leadership level about how to fix that.

Yeah. I don't know if you have anything to add. Yeah. I think, um, from the technology perspective, I think, um, as I mentioned, I think we have a highly customized solution that supports our current existing campaigns. The key challenge here is, um, how do we we we need that large team to continue to maintain this custom solution.

Plus, if somebody has a new idea to enable this, that new idea, you need a lot of data engineer, a lot of engineering work, as well as data scientist work in order to make that first POC activated.

Right. We spent a lot of time and for example, let's say that we are not able to customize that one campaign. The team is not going to wait for you to fit into that customized solution.

They're still going to outreach to the members.

So what happens is you have this system where all the all your intervention with the customers are not centralized.

So you don't know how many times we're reaching out to the member.

You're not able to do any learning about the member.

And that's where I think the Pega comes into picture. Absolutely. Thank you very much. Thank you. Good question. Thank you. We've got time for 1 or 2 more. There we go. We've got gentlemen here. Thank you for this presentation. I had a question about the channels that you guys have used to activate these nudges. One what channels have you found, um, prevalent for your campaigns? And second part is, how is Pega helping in activating those channels? Yeah. I'll, uh, I'll, I'll answer more on the first question. Uh, so first off, I mentioned it before it, uh, we have traditional channels, I'll say is like things like direct mail outreach, um, call based outreach. That's outbound inbound calls. Outbound calls can be both like live outbound and also like, you know, like robocalls, right? Voice, uh, interactive voice recording type stuff. There's the digital channels, there's email, SMS, EMS within that. Um, there's RCS business messaging is a newer, newer channel in the industry. Um, What am I missing, Webb? And we'll have, like, you know, interactive, 1 to 1 personalized microsites, for example, is, of course, app like notifications.

Um, and I mentioned there's another category around what I'll call relationships. So relationships with like a care manager pharmacist physician depending kind of where they are and what the topic is, we may have certain like connection points to different like clinical relationships. So those I would say is mostly the ecosystem of channels. I probably got most of them there, um, at least for like population health stuff we're talking about. Like I didn't mention, I didn't mention like social media, for example, because it's not necessarily appropriate or doesn't have like the targeting capabilities that you need for like some of the stuff that we're doing that's involving Phi. Okay. And then the question I always get from people that I think is like part of what you're asking is like what you said most prevalent. Um, and maybe by that you mean what's like the most effective channel? Um, by the way, I'd say most prevalent. Like, we, we love we love email and mobile messaging because digital is good. I would say in general we see people who have their digital permissions on in the healthcare ecosystem, um, have better health outcomes.

And I think there's a little bit of selection bias here that people who have their permissions on for a health system or like healthcare payer, like we are we or at least part of our businesses, they're more likely to be engaged with their healthcare, be more likely to maybe change their health behaviors.

But I mentioned there's a selection bias there. I, I will say I have never I can never say that like 1 or 1 mix of channels is more effective than anything else, because it actually, I've found to be very dependent on the use case, the specific topic that we're on. So that's that's another reason why I think this, like agile iterative test and learn approach is so important because there is no silver bullet when it comes to channels. I would say it actually does depend on the topic. Depends on the audience that you're reaching out to. Um, so that's kind of my, uh, my, my regular diatribe on channels. I don't know if you guys want to add anything on the second part about how Pega is maybe enabling. Channel. Activation. I think right now, as you can see, that we are using just member preferences to activate the channel and, um, reach out to with their personalized content.

But with Pega, I think it will be more useful to see how the member is behaving.

Um, if we send out email or SMS and it will help you drive the next set of communication to the member to what is the preferred channel that member is like.

Member can be reached out to multiple various channels, but to make that decision and learning from the past engagement is key.

And I think that's where the Pega comes into the picture. Yeah, I'll just add on to that for two seconds. Um, you know, one of the cool things that we can do, obviously, is we can have adaptive models that select the next best channel for the customer that's kind of out of the box as long as we configure it that way.

Um, but there's there's a step further, which is not just what's the best channel, you know, remember, this is a big chunk of what we're doing here is outbound.

So what's the next best time to actually make that communication? So we're running this thing on a on a daily batch. We're going through many, many, many millions of customers here.

But what's the actual right time to land the message for the customer.