How Cedar & Twilio Are Rebuilding the Patient Experience with AI

0:00:08.2 Kailey Raymond: Welcome to Builders Wanted, the podcast for those boldly building what's next in customer engagement. I'm your host, Kailey Raymond. Today, we're talking about healthcare, but not the clinical side. We're digging into the part that's often overlooked, but deeply felt, the patient financial experience. Because no one should need a PhD or a customer service hotline just to understand their bill. Joining me is Dugan Winkie, Head of Commercial Strategy at Cedar, a company that's taking a builder's approach to modernizing how patients interact with healthcare billing. From automation to conversational AI, Cedar is rethinking the entire experience with empathy, intelligence, and a lot less friction. We'll talk about what they're building with Twilio, how it's working in the real world, and what it takes to innovate in an industry that doesn't exactly love change. Let's get into it.

0:01:03.5 Producer : This podcast is brought to you by Twilio, the customer engagement platform that helps businesses turn real-time data into seamless, personalized experiences. Engage customers on their terms, across SMS, voice, email, WhatsApp, and more. Power every interaction with AI, so conversations feel natural, not robotic. Adapt in real time, delivering the right message on the right channel, exactly when it matters. That's the power of Twilio. More than 320,000 businesses, from startups to Fortune 500s, trust Twilio to transform customer signals into conversations, connections, and real revenue. Reimagine the way you engage with your customers. Learn more at Twilio.com.

0:01:52.1 Kailey Raymond: Dugan, you're the head of commercial strategy for Cedar, and you're focused on new product innovation and strategic planning. And I know you have over 15 years in healthcare under your belt, a ton of great experience, very excited to learn from you today. Welcome to the show.

0:02:08.8 Dugan Winkie: Yeah, thank you for having me. Excited to be here.

0:02:10.8 Kailey Raymond: I want to jump right into it. What would you say is the boldest thing you and your team are building right now?

0:02:19.1 Dugan Winkie: Yeah, so I think we're doing a lot of interesting things, but if I had to pick one, I would talk about our new Agentic AI strategy. Our new AI agent, Kora, is specifically focused at how do we enable health systems to operate more efficiently and effectively in their call center, specifically focused on the financial experience or billing-related questions. And I think what makes this a bit bold for us is Cedar has always been laser-focused on the patient financial experience. But in the early days, it was much more around how do we reduce friction in the digital experience with all things related to your healthcare billing. So how do we consolidate statements? How do we minimize friction in order for people to digitally engage with us and digitally pay their bill? But I do think part of any good leader thinking about new strategy is always taking a close eye to what the feedback the market is saying. And for a lot of healthcare providers right now, they are under significant operational cost constraints, regulatory uncertainty, and budgets that are being cut in some cases. So we felt like it was our responsibility to bolster our existing value proposition and provide a new tool to providers in a way that both continues to delight healthcare consumers, but also is a real actionable asset that healthcare providers can use to operate their call centers more efficiently and effectively.

0:03:52.0 Kailey Raymond: Beautiful. So I think you're talking about your product Kora and this kind of like agentic experience. I'm super excited to jump a little bit deeper into that with you later in the show. It sounds like it's a win-win for you, for the industry to make sure that you're achieving efficiency across the board. It also feels like it falls right in line with Cedar's mission, which I know is all about humanizing the patient financial experience. How do you bring that to life in a system that isn't always known for empathy?

0:04:24.3 Dugan Winkie: Yeah, most people don't typically equate their healthcare billing experience with a delightful experience. I think it's gotten a little bit better, but I remember when I first joined Cedar there, we did a bit of an assessment and found that people consistently ranked their DMV experience kind of higher on the satisfaction level compared to navigating the healthcare financial experience. So there's room to improve, we'll put it that way. But what does that mean? Like, what does delightful mean in our minds? At the end of the day, we are, in most cases, communicating a bill to a patient. So there's kind of an upper bound threshold of like how excited are people going to be to receive a bill. But I think as you peel back that onion a bit, what you'll find is there is a substantial portion of the patient population who has both means to pay as well as a willingness to pay, but there's just structural gaps in the information that's creating dissatisfaction. So I'll give you like one very real example. You get a bill from a provider and then you get an EOB or an explanation of benefits from your payer that says it's not a bill, but it sure looks like a bill.

0:05:36.3 Dugan Winkie: You call the provider to ask them a question about your deductible status, they can't answer it, and they redirect you to the payer. So that's just one example of the type of friction that I think is fairly unique to healthcare because there's so many of these like disparate parties that are involved in a patient's care and a patient's reimbursement for said care. But there's a lot of information silos simultaneously where we're expecting patients to know how to navigate a very complex system. And as you alluded to, I've spent my entire career in healthcare. It's still confusing to me, right? So I always try to wake up every day thinking about things from the average healthcare consumer that may not know what co-insurance is or doesn't quite fully understand the hoops that they need to jump through to get a claim covered. And really what Cedar wants to do is be a bit of that digital concierge Sherpa that is holding their hands and connecting the dots. And we think conversational AI and agentic AI is a great manifestation where we can leverage a lot of our core competency and network connections to help answer those patients' questions in a way that is delightful.

0:06:52.5 Dugan Winkie: At the end of the day, there's also, I would say, like technology enablers that we haven't had historically. So if you have a question about your bill at midnight on a Saturday, the call center is not typically gonna be open. Or as we think further out, the opportunities to detect human emotion and escalate accordingly or even translate our agentic AI voice into a myriad of different languages. We're very early days, this industry is very early days, but I just think of the opportunities into the future and it's so exciting.

0:07:28.1 Kailey Raymond: I am so glad to be speaking with you today and for you to validate my feelings and how frustrating or confusing this can be as a patient. And very often I find myself thinking about how having upfront access to some of this information might just make the bill feel a whole lot better once you actually receive it. So having some of that context and kind of knowing where it's coming from and being able to connect the dots between these different systems, as you're saying, is not something that's incredibly intuitive and something that, if somebody could help me, it's like in high school, it's like, what do they teach you? They don't teach you about finances, they don't teach you about healthcare, like all these life skills, you know?

0:08:10.9 Kailey Raymond: And so having an AI to be able to help do some of that work for you and not pile on additional stress and what could be an already stressful moment in your life sounds just simple, a little bit unconventional from where we are right now in healthcare and just extremely innovative. So very exciting to see where we're going. I'm wondering if you think back on your career path, if there's a moment where you might've taken an unconventional approach that ended up being exactly what the business needed at that moment.

0:08:43.1 Dugan Winkie: Yeah, absolutely. I would say even in a more recent context, Cedar in our early days was really focused at how do we help health systems increase collections and do it in a way that is a delight to the healthcare consumer as opposed much more of kind of a carrot versus stick based approach. So even us saying, hey, we have heard from the market that cost pressures are increasingly high priority and you need a novel solution that will help you manage this process holistically, that is pretty unconventional. And we had a good amount of internal dialogues around how we're placing our bets as a company, how we're allocating resources. And at the end of the day, is call center automation, the type of transformative innovation that we wanna put our name to? And I think the answer to that is, we think it is a great starting point that then creates a bedrock for future innovation. And what I mean by that is agentic, conversational AI is a new and ever evolving space right now. It seems like every week we're disrupting ourselves and health system providers I think are excited about it, but they're also a little bit leery of the rapidly evolving situation.

0:10:07.9 Kailey Raymond: Totally.

0:10:08.6 Dugan Winkie: So what we have heard from a number of our partners is, look, we could take a shallow but very broad-based approach to adopting AI, or we could take a deep and a narrow-based approach. And most are leaning towards the latter of the two there because they can set up better guardrails, we can work closely with their AI governance committee, and there's just such a monumental opportunity for disruption. So as one little proof point or nugget, we did an AI automation assessment with one of our partners where we ingested call recordings and then used our LLM to say, what was the intent at time of calling, and ultimately what was the degree of automation potential, and found that about 97% of inbound calls were not simply to just make a payment, but it was they had a question, they were confused, they needed to be educated. So for us, if you look at other industries, the number of questions people have about billing-related items is probably around like 80-ish percent, but 97% of patients, right? That to us just tells us that there is an urgent need that the current setup is not meeting the needs and expectations of consumers.

0:11:31.4 Dugan Winkie: So that one is, call it a bit more unconventional, but current state, look into the future. I would say looking a bit more historically, I had the privilege of being at Epic back in 2008, right around the time that high-tech and meaningful use legislation passed. So what it meant was there was additional funding that was earmarked for EHR system adoption, and it really was an exciting time to be part of that because I got to see what scaling looks like, and in particular, scaling within providers, which is not necessarily an industry that typically adopts change at a very rapid clip. But I think in a period of, call it 10-ish years, they went from single-digit EHR adoptions to well north of like 80%. So I learned a lot to think about how providers think about disruption, think about change, and think about risk, because it is a much higher burden that they have to hold themselves to, given the highly regulatory nature of this combined with patients' lives are at stake. So people always talk about, well, why can't we disrupt health system quicker, right?

0:12:52.5 Dugan Winkie: And some of the players, the Googles, etcetera, that have jumped in and maybe haven't been as successful as initially thought. I think the first, in my mind, the first thing is to really learn about what is unique about healthcare and tailor your approach to them, given it is a fairly unique business.

0:13:12.9 Kailey Raymond: Totally. I mean, in these extremely regulated busineses, it's a different ballgame. I think what's interesting about Cedar is that it is taking this technology-forward approach to a highly regulated industry, right? So I do think that you're really blending a lot of the principles. You mentioned the Amazons, you mentioned the Googles, which are really taking in those pressures of consumer demand, of personalization. I need it now, like the pace that you need to be able to deliver an experience today. People don't have the tolerance that they used  to wait on a phone line or to, frankly, wait for something to be shipped in the mail. You know, they expect that they want it and they need it now. And so I think that's just kind of a really interesting meeting point of regulatory industry meets like technology need and Cedars kind of coming into the middle of this. And  you've kind of mentioned this, but the Amazonification of the consumer demand kind of bleeding into healthcare with patient's expectations, I think also probably making sure that you are personalized, digital first in all of your interactions is something that I'm sure is on your mind every day. So what do you think is a recent patient behavior or demand that's pushed Cedar to think differently about the billing experience?

0:14:32.9 Dugan Winkie: Yeah, I think part of it is acknowledging the realities of a highly regulated industry, but also not letting it kind of dissuade you entirely. You need to adjust your expectations. And I think for Cedar, like the one thing I am incredibly proud of is I think we've tried to approach this from a blended perspective. I'll give you a couple examples. So  we have one of our co-founders  came from ad tech, right? So they've been looking at how do you dynamically engage consumers for decades and decades. And  in healthcare, I do think that there is a large portion of innovation that is simply just adopting and tailoring existing best practices and other industries for the healthcare industry. I'll give you like one example, the abandoned car checkout workflow. Pretty much any e-commerce site that you go to, if you get to the checkout screen and you leave, like you're going to get some sort of nudge in about 24 hours. But nobody had done that in healthcare, right? So as we kind of progress with having a more mobile first approach, ensuring that the expectations that consumers have, both in terms of like best in class user design, user experience, dynamic engagement, all of that needs to be in place.

0:16:01.6 Dugan Winkie: But simultaneously, like I think we need to build the solution based on  healthcare expertise, because it's not enough to say, hey, instead of having this button be blue, let's do dark blue, right? Like the issues ingrained in healthcare billing are just much more structural in nature, and it gets back to having all of these disparate stakeholder groups. So what we're trying to do is really position ourselves as a connector, like a hub, like our platform is here to help you navigate a very complicated process. And what could that mean? One, if you simply want to talk to an AI agent that has all the answers and not want to wait on hold, excuse me, like we can do that. Alternatively, like if you can't afford your bill, that is a challenging question. And there's actually a lot of pockets of dollars available in various  nooks and crannies across the healthcare industry, but we're reliant on healthcare consumers to understand that and to advocate for themselves. And I think that's a premise that is very challenging. So what can our role be in it? It's whether or not you are eligible for Medicaid enrollment.

0:17:18.6 Dugan Winkie: Is there an Affordable Care Act plan that we can match you with? Do those drugs that you're being charged for actually qualify for a copay assistance program? All of these are examples of areas of opportunity where if we can shoulder some of the complexity of this and just present a financial pathway for patients using a number of different variables that ultimately increases their ability to pay and decreases their likelihood of being sent to collections, that's a win.

0:17:51.8 Kailey Raymond: That's amazing. It's one of those things where I feel like patients might not even know the right questions to ask themselves. And so by gathering all of this information at AI's fingertips and being able to unleash it into the patient experience, you are giving people access to information and making them feel heard. And like they can ask good questions and be their own advocates. Because I often feel as a patient that when I'm walking into a room with a doctor, I might not even know what questions to ask. And then on the backside and the billing experience, it's the same thing of how did this show up here? Where can these funds apply? So I think that's a really great application of where you think AI can come in here. So I know that you've implemented conversational AI into your platform. What made now the right time to invest in that? And what impact are you really hoping for with this?

0:18:57.8 Dugan Winkie: Well, I always like to start by saying, like, when people talk about AI, I think that the majority of Americans think of AI corresponding with a ChatGPT revelation, right?

0:19:09.8 Kailey Raymond: It used to be Terminator, and now it's ChatGPT. You know, it's like a little bit blander.

0:19:14.7 Dugan Winkie: Exactly. But  in our minds  we have been an AI native platform for the past  seven, eight years since we started the company. It's just back then  AI was predominantly machine learning. And really what we were trying to do is create dynamic payment resolution pathways and engagement pathways. So if you're a college student who is 29 and has a $10,000 health care bill compared to somebody who is late career, 50s, and has $100 co-pay, those are two very different consumers with very different questions, with very different abilities to pay. So for us, it was all around how do we optimize how we're communicating, when we're communicating, and what we're communicating to those unique patient archetypes with the net goal of increasing digital engagement and  successfully completing the flow. What I think makes generative AI really interesting, though, is it's almost like the next evolution of how we communicate and receive answers. In the very early days  it was calling a phone and talking to a human. And then we had this, like, advent of  a number of SaaS companies and obviously the digital communication that redirected people to a portal or an app, right? I almost look at conversational AI as like the next evolution of we no longer need necessarily to go through  a digital portal or an app to learn all the menus and accelerate

0:20:58.9 Dugan Winkie: Self-service, but you still need to know where to find the information. It's just much easier to talk to an agentic AI agent, and they can provide all that information faster in a highly conversational context and without necessarily have as much expertise on, like, the digital, like, portal, which is going to increase usability at the end of the day. I do view this as something that will continue  to expand where it may be a voice agent, it may be a chat-based agent. We, in the early days, like, we're actually experimenting with both mediums and found that the voice medium was just better suited to achieve, like, very scalable impact because you need to meet the patient where they are. And  it's pretty much a 10 to 1 differential in terms of if patients have questions, they're picking up the phone to call their provider as opposed to logging into whatever their provider's chat portal is. So those are  some high-level areas that we're looking at. But I would say I often get, like, is this hype, is this reality? I firmly believe that there is a lot of hype out there right now.

0:22:10.5 Dugan Winkie: There is a lot of hype that's going to be around surrounding any technological advancement, but that doesn't minimize the fact that this is structurally, like, different inflection point as we think about our ability to engage and delight  healthcare consumers.

0:22:27.3 Kailey Raymond: I fully agree with you. And I think what you're telling me is that I no longer have to dial a phone number and incessantly hit one until I actually speak to a human. Instead, there's this really elegant kind of middle layer of human voice intelligence, which is an AI itself. So let's dig into Kora a little bit. So Cedars Conversational AI agent Kora obviously designed to do more than just automate responses. It's built to handle complexity. It's built to sound like a real human. So you call in and Kora understands and just that information and  routes you to the next best place. So I want to hear a little bit more about that. What decisions did you make when it comes to things like voice design or a patient interrupting the AI agent? What was critical to making Kora feel like it was genuinely helpful and empathetic to patients?

0:23:29.4 Dugan Winkie: Yeah, it's so important to build that trust with patients. So I think at the core, what we viewed as our long-term sustainable differentiation is the data that we have at our disposal to optimize our LLM models. Because my hot take is that the nuts and bolts standard LLM models are going to be fairly commoditized. It's already trending that way at a pretty rapid clip. So as I think about what Kora, our AI agent focused on patient billing experience, how they can be more effective than another point solution down the road, it's really its ability to successfully answer those questions. If you peel that back another layer, what is required to answer those questions? First off, you can't just have one side of the data. So we integrate very closely with providers. We integrate with payers. We integrate with HSA banks. We integrate with a number of affordability mechanisms such as automated Medicaid enrollment. So all of these sources of data is going to help increase the efficacy of that AI agent to answer questions. And what we found through user testing that is a fascinating insight as well is that you need the LLM to really approach this like a problem solver.

0:24:58.3 Dugan Winkie: Because a lot of times patients don't even necessarily know what the issue is. The conversation may start saying, I'm calling in because I'm confused about my bill. And as you start walking them through that, there may be some definition like, I don't understand. I paid $200 of coinsurance at the front desk. I thought that was it. What is this $1,000 bill that I'm getting afterwards? So those are fairly easy to answer. But then as the conversation progresses, it may come down to, I can't afford this bill. Now I understand why I owe this, but I'm telling you it's not possible. So how do we put them on an extended duration payment plan? How do we connect them and enroll them in some of these affordability mechanisms? Those are all inputs that are predicated on us having data hooks into all of these different ecosystem players that is going to move the needle for us. So I would say that that's backend infrastructure. On our front end, we really wanted to take a partner-based approach on the best-in-class text-to-voice AI out there. So we've been partnering with Twilio, leveraging their conversation relay service, which really enables developers to create robust natural voice AI agents and integrate our data and our LLMs directly into it.

0:26:23.5 Dugan Winkie: So we're not approaching this all as a 100% build. I think it's a combination of a build and a partner, looking for those partners who have really asserted their superiority in certain aspects of the value chain.

0:26:38.7 Kailey Raymond: One of the things that I'm really glad that you spent some time on touching on is the quality and strength of the data that you're feeding into this, because at the end of the day, what that does is it makes sure that people trust you and that if patients feel like you actually know them and that you are paying attention to them. They're much more likely to engage with your services. So building that trust is essential. And it goes without saying that within all of this, safety is essential to bake into the AI systems. Of course, within highly regulated industries, that is necessary to kind of be a standard. I'm imagining along the way, this isn't easy stuff to build, but you've probably encountered some roadblocks in trying to modernize the patient experience. How did you navigate that when you're kind of running into these roadblocks along the way?

0:27:36.5 Dugan Winkie: I'll give you a real-life roadblock that I think pretty much every provider is grappling with right now and obviously has a lot of healthcare consumer implications on top of it. So we are seeing two parallel trends, one of which is healthcare is getting more and more expensive, right? Like the average bill size is going up, not down. Those reasons are multifactorial, but when you take into account inflationary pressures, as well as just the rising costs of care, we see it in our data. Like the amount that providers are billing for services are going up. And then the second component is the amount of liability or the amount that patients are responsible for on the hook for is also increasing. So  if you go back like 20, 30 years, the majority of people had a job with a great health insurance plan and the amount that they were responsible for coming out of their checking account was minimal. Like the average health system that was like single digits. Of all the outstanding payments owed to providers, the amount from uninsured patients were like single digits.

0:28:56.3 Dugan Winkie: Just this last  year, across our customer base, that number is now about 35%. So what does that mean? Like out of all the outstanding payments owed to them by patients, about 35% is from patients without insurance. So this number has increased a lot. The amount that we are asking patients to pay has increased substantially. So I really don't think we can just approach this using the same tools that we had at our disposal  years ago. And there is absolutely a technology role here. I alluded to the fact that we had HSA and FSA integrations. This is just like one kind of discrete proof point here is all we did in those cases were link, like for the patient, when they say they have a bill due by them, when we go in there and actually share the HSA balance and say, Jenny, you actually have $300 in your HSA available, would you like to use that to put towards your payment? We see satisfaction go through the roof. We see payment rates go up by close to 10%. It's really because a lot of consumers maybe had a funded HSA, but they lost their car. Maybe it's hold away in some junk drawer.

0:30:19.3 Dugan Winkie: They've never logged into their HSA banking partner. They have no idea that they have some of these funds available. So that's just one example of like, ideally the types of things that we're trying to connect where there are ways to make your bill more affordable, but we need to do that ultimately in a way that doesn't create additional patient friction and just pulls it all together and put a bow around it.

0:30:45.3 Kailey Raymond: I feel like I needed this conversation right now. I was recently hit with a bill that was larger than I expected. And that frustration came at the payment versus upfront when if they could have given me a little bit more context, it would have been much better to actually stomach it then. So I'm very excited about what you all are building and how you're bringing a lot of information together. We've touched on this a little bit, but there's this obvious balance that you need to strike when you're innovating, especially with newer technologies like AI. We talked about this regulated industry. You have to do it with a little bit of caution, perhaps. So how do you balance those two things? Caution, innovation, especially with Gen AI

0:31:35.1 Dugan Winkie: In the mix. This comes up a lot. And I think what gives me a lot of optimism is providers and healthcare systems at large. They know how to assess new technology in a way that meets the highly regulatory nature of their business. In the same way that new treatments, new medical devices, new clinical practices are coming out on a daily basis, these health systems, I think, recognize that they cannot afford to wait, which is for us a really optimistic sign. But simultaneously, they know that they cannot just sit around and wait for some national governing body to come out with a AI governance structure and assessment criteria. So what are they doing in the interim? Pretty much every health system that we're working with does have that. They've stood up that AI safety and governance body to really look at efficacy, guardrails, transparency. And we think that's great because we want to build this with complete and full transparency. At the end of the day also, I think what's interesting about conversational AI, if you talk to any of the vendors out there, it is rapidly maturing, it is getting better. I would argue it has much more raw clay or putty to create really delightful interactions than a typical call center representative that maybe was hired two weeks ago and has no healthcare background.

0:33:14.3 Dugan Winkie: By the way, they have to go through and hunt and peck in the EHR to try to figure out what bill you're even talking about. The status quo is not great. And I think we need to ensure that we are providing the transparency to improve upon that. Is it going to be perfect out of the gate? No. But what are the things that are most important to health systems so that we can tailor it? And so we've had explicit conversation of what percentage of hallucinations are appropriate. Is it zero? And all of these things are controllable, but we need to have an open dialogue. And for me, I think as the technology is rapidly evolving, it's great to actually have providers as partners because they bring up questions that maybe we hadn't thought about and vice versa. So we try to approach this very lockstep, but ultimately trust-based. 

0:34:07.5 Kailey Raymond: That's great. I love that full radical transparency approach. And I know that one of Cedars' kind of frameworks is also around scaling empathy. And I think that running the course throughout this entire conversation is you're really taking an empathetic approach. You're putting yourself into the patient's shoes to make sure that you feel it every day, like everybody is a patient in some sort of health system. So I think that also kind of applying that principle to your AI methodology is incredibly important. So trust, empathy, those are things that go across every different type of company. It's not just healthcare. It comes back to any customer experience, a good lesson to fall back on.

0:34:50.4 Producer : Great customer experiences aren't magic. They're built. And Twilio is the platform that helps you build them. Every customer action, browsing your site, opening an email, reaching out to support, triggers instant AI-powered personalized engagement across SMS, voice, email, chat, and more. No delays, no guesswork, just the right message at the right moment. From automated messaging to seamless authentication to hyper-personalized customer journeys, Twilio's customer engagement platform powers millions of interactions daily, helping businesses drive loyalty, optimize marketing spend, and create experiences that people remember. Twilio is the ultimate toolbox for customer engagement. Ready to build experiences that matter? Visit Twilio.com.

0:35:42.1 Kailey Raymond: I want to bring It a little bit back to Kora and some of these kind of real-world examples. So I want to hear how you think AI has already improved the patient journey. You can give me an example of personalization or automation or even smart escalation. Any examples of where Kora is kind of jumping in and helping?

0:36:03.6 Dugan Winkie: I'll start by maybe going back to a comment you made a few questions ago around this idea of trust and consumer expectations. Because one of the things that we discovered very early is many consumers actually prefer a digital or more anonymized experience as it relates to financial experience. Because what are they talking about? They're saying, I can't afford this bill. And they would actually rather talk to an AI agent than a human individual at the front desk when there's somebody standing right behind them. Because it's not a comfortable topic, right? It's highly sensitive. So I do think what we have found is that there is a willingness to engage with AI agents. But the careful caveat is as long as they are helping me and answering my question. As soon as you stop being a value add to the consumer, the hammer on the zero until you get escalated to a human agent absolutely kind of remains. And we expect, honestly, about  call it 20-ish percent of patients will want to engage with an AI agent. But if we go back to that premise of 97% of patients are calling out with questions about their bills, like there's still a lot of opportunity where we can engage with people who are willing and open to engaging with an AI agent, provided it is answering their questions and making them more efficient.

0:37:38.9 Dugan Winkie: So what are some of the things that we're doing? As I alluded to earlier, we had that AI automation assessment where we funneled in hundreds and hundreds of hours of live kind of patient call center interactions and use that to triage our use cases. And one of them that was very non-intuitive was just authenticating a patient. So what's your name? What's your date of birth? Call center representatives would spend roughly about 20 % of the call just doing that.

0:38:08.4 Kailey Raymond: Wow.

0:38:09.1 Dugan Winkie: Another example is getting a copy of an itemized statement. So oftentimes, like health care insurance company will like ask for a copy of the itemized statement. It's a bit more detailed, but roughly 10% of calls for one prospect was just people asking for that. So those are really the areas that we're trying to start with, which is like, what are the lowest of low hanging fruit where it's inefficient for the human call center representative, it's inefficient for the patient, like the answer is pretty clear. Let's just do it in a way that is a win-win. As we look at like higher kind of complexity use cases that we're tackling, answering questions about their payer and their benefits is certainly one of the higher priority components, integration into the various like funding mechanisms, the affordability mechanisms that we talked about. But I like to use that user verification and itemized statement example just because it shows you like there are some very easy wins out there if we can do it in a way that  ultimately make this win-win for the patient and the provider.

0:39:18.5 Kailey Raymond: Super low hanging fruit, like very obvious application of where AI can help versus a human. And I think that's the interesting part is like identifying what a bot should be doing versus what a human should be doing, those more complex things that  a human brain is probably still better able to do. And I always find that conversation really interesting when it comes to these types of kind of decisions and how you route calls and why. I'm wondering if you have any insights into something that surprised you when you were kind of looking into some of this data about from either a provider, a patient, maybe your own team about how this technology is being used in practice.

0:39:59.4 Dugan Winkie: Well, one, I would say one thing that was really surprising is like when you added up all of those incremental kind of easy use cases, you start getting to a number that is very exciting. So like we're on track and targeting to automate about 30% of inbound calls by the end of the year.

0:40:18.6 Kailey Raymond: Wow.

0:40:19.1 Dugan Winkie: And why does that matter? Because  if you look at some of the large providers, they are answering thousands of these calls a day and are spending millions of dollars to staff their call center, which is increasingly challenging right now from a labor retention side. I think that is one thing that is absolutely exciting for us is this matters. But I would say another thing that was really surprising more from the patient side is I think we had an ingoing hypothesis that maybe patients, slightly younger demographic patients that are more tech savvy would interact and engage with the AI agent, maybe more so than those who are a little bit less like digitally literate. But that was not really the case. Like I think what's really exciting for me about conversational AI is there is not that high cost of admission as some of the previous technology of the past, where even if you look back, right, the early adopters of  iPods, right, like you needed to figure out how to load your music library. And like there was like a certain level of tech savviness that you needed to have to be an early adopter.

0:41:33.8 Dugan Winkie: This requires no level of savviness. Like you just need to interact and talk the same way that you would in a normal conversation. And as a result, we've actually seen very, very high satisfaction from patients that are substantially older and may not necessarily have like an active EHR portal or engage a lot via text. But  just one like fun anecdote, like we were listening to a call recording. And at the end of it, this nice lady was saying, like, I'm going to hang up and call my grandson because he is going to be thrilled that  I just had a conversation with a computer. And it's just like  like as long as you're answering the patient's questions, there is a little bit of novelty here. But I think it could really be an equalizer in the sense that you don't necessarily have to be as technically proficient to get the most out of this technological advancement.

0:42:33.9 Kailey Raymond: I love that. This often comes up on the show is like the demographic assumptions of applications for technology and how it diverts from what you might originally think. I love that insight. I appreciate you sharing that. I've learned a ton from you today. Thank you very much for your time. I have one more question for you, which is, if you had any advice to someone that's trying to build innovation into a system that wasn't necessarily built for it, what would it be?

0:43:04.5 Dugan Winkie: Yeah, I would say persistence and empathy go in expecting that there's going to be ups and downs. And I think persistence is the key. I would say how empathy kind of fits into this is I think oftentimes  we come into a process that may seem inefficient on the surface. And first knee jerk reaction is to say, this is the wrong process. Let me show you what the right process is. And for me, I'm a very strong advocate of like, let's actually look at like the inputs into what made that process challenging. Right. Like what were the regulatory hurdles? You know, what are kind of the people process components that are driving maybe some of that inefficiency? Like it doesn't minimize the fact that like you will likely be the change agent in that organization. But I think your odds of success go up drastically if you can convince people that you understand maybe handcuffs that have been placed on them and you want to work with them to remove those handcuffs. Developing an advocate as opposed to being an outsider threatening to disrupt their day to day job is a really key difference in my mind.

0:44:23.1 Kailey Raymond: I love that. Great insights. I've learned a ton again. Thank you so much for being here. And thank you to Cedar for building a more human, intelligent and personalized healthcare experience. Appreciate the time.

0:44:34.8 Dugan Winkie: Yeah, thank you.