Good morning and welcome to the R1 RCM Technology Teach-in. At this time, all attendees are in a listen-only mode. A question-and-answer session will follow the formal presentations. If you'd like to submit a question, you may do so by using the Q&A text box at the bottom of the webcast player or by emailing your questions to questions@lifestyleadvisors.com. As a reminder, this call is being recorded and a replay will be made available on the R1 RCM website following the conclusion of the event. I'd now like to turn the call over to Evan Smith, Senior Vice President of Finance and Investor Relations at R1 RCM. Please go ahead, Evan. Thank you. Thank you, everyone, for joining us today. Over the past year, investors have asked several questions about the strength of our technology platform and data ecosystem, as well as our ability to leverage the platform across R1's comprehensive services and modular suite to further lower cost, improve quality, and drive new revenue opportunities with our customers. Today, Steve and Brian will provide insights into the strength of our platform, new AI advances to date, and our technology roadmap that we believe will keep R1 at the forefront of the revenue cycle industry and a partner of choice for providers. Now, let me move to the forward-looking statement. Given we will make forward-looking statements during technology and innovation organizations, which we are confident will continue to support R1's ability to drive value as we move forward. Now, let me pass the presentation over to Steve. Thanks, Evan. I'm excited to be here. When we leave here today, you'll understand our platform and data and how they're differentiators, our investments in technology and how they'll enable us to drive growth and scale, and how we'll leverage technology to develop new solutions that will drive growth for us and better results for our clients. Brian and I will walk you through our product and technology vision, our strategy on how to get there, and our key focus areas. At the end of the day, we'll open it up to Q&A. Now, let's turn to the vision. The vision for R1 is to be the provider of choice in revenue cycle. Our technology vision is to have the best tech-enabled platform in the industry of revenue cycle. It starts with data. We have a breadth and depth of data that's unmatched that spans the entire healthcare ecosystem, giving us insight into patterns and trends that others can't. We invest in technology that allows us to take an automation-first and increasingly AI-enabled approach to running the revenue cycle that delivers better customer results, better patient satisfaction, and supports our own margin performance. What's special about R1 and our business model is that we have three critical building blocks that come together to allow us to innovate at scale. Can we go to the next slide? First, our revenue cycle expertise. Second, our unmatched data. And third, our position of being deeply embedded in our customer workflows. Our revenue cycle expertise stems from the fact that we run all 14 steps of the revenue cycle. We have best practices around how to best operate every nuanced aspect of the revenue cycle, and that proprietary knowledge is embedded in our standard operating procedures, our standard methods, our rules, and our algorithms. We have an unmatched breadth and depth of data. We see over 500 million patient encounters every year, including 60% of the acute care market. We see them across all geographies, all care settings, virtually all payers, and all payment forms. That allows us to see patterns and trends and changes that inform how we optimize our workflows. We're deeply embedded in our customers' workflows, and that gives us access to the best point to take action. That's what enables us to deploy an automation of a task right when we need to, which we do hundreds of millions of times per year. It's also what enables us to provide the right guidance and recommendation to a revenue cycle expert when they're about to make a decision. Over time, we've built on these critical building blocks. We started with a global captive model. This allows us to deliver very high-quality results for clients at a lower cost. We started in India, and we extended over time to the Philippines. In 2018, we launched our intelligent automation initiative. This is an area that's been highly successful for us, even though many of our clients and others in the market have struggled here. We took a different approach to automation, where we have a dedicated team that builds automations in the form of reusable components that are tested and monitored rigorously. That approach has allowed us to automate the work equivalent of 5,000 people. Over time, as technology has become an increasingly important part of our value proposition to clients, we've expanded our technology solution set. One example is in the front end of the revenue cycle in the patient access space, where we've added patient self-scheduling technology, physician order processing technology, patient self-registration technology, and patient payment technology. This enables us to not only deliver a more satisfying experience for patients, but allows us to capture better information upfront in the revenue cycle, which allows us to prevent problems that we're observing happening downstream in the back end of the revenue cycle. Another example is our acquisition of Cloudmed, which gave us access to a huge breadth and depth of data and also to advanced machine learning and predictive analytics capabilities. Going forward, we'll build on this foundation of global technology and services, RPA, machine learning, and predictive analytics, plus our breadth of data to continue to develop new solutions to benefit clients and new modular solutions to drive more growth. With that, I'd like to turn it over to Brian, who can now tell you about our key focus areas. Thanks, Steve. Here at R1, we're really intentional in our use of technology. We know that building proprietary software is expensive, so we only do it when it is the best way to unlock value from the revenue cycle. What that usually means is that we've got a nice combination of access to data, some great best practices, and meaningful proprietary rules and algorithms. To illustrate our approach, today we're going to be looking at our use of tech in three ways. We'll talk about enhancing our operator platform. This is key. It's the primary way that we improve our internal employee experience, drive efficiency, and produce results for our customers. Second, we're going to talk about the fact that we've been a leader in automation for a long time. Using RPA, we have taken millions of manual tasks off the plates of our employees. Today, we're going to talk about how we're injecting next-generation technologies like large language models into our intelligent automation initiatives. These capabilities are opening up new areas of automation, whole new bodies of work in which we can reduce errors and eliminate manual activities. Finally, we're going to walk through some of our new areas of innovation. These solutions combine our massive data scale with our expertise in every stage of the revenue cycle. They can be deployed for our current customers, both in our enterprise end-to-end business and also in our modular RPS business, and of course, in net new logos of the future. Now, let's turn the page and take a closer look at our platform. On the right-hand side of the screen, you can see the 14 steps of the revenue cycle. These steps cover the complete process, from the moment of patient scheduling and appointment through the encounter itself, all the way through to when the health system is finally reimbursed, and every step along the way. To keep it somewhat simple, there are three key aspects to our platform we'll discuss: our data, use of leading-edge technologies, and our own proprietary revenue cycle optimization engines. We'll talk about those quite a bit. The key point to keep in mind is that we leverage the platform to embed ourselves in the customer workflows on the right-hand side. We can inject ourselves anywhere. Now, let's dive into data. For me, as a technologist working at R1, one of the most exciting things is our access to large amounts of data. R1 has already built thousands of data connections into the healthcare ecosystem. That means payers, clearing houses, banks. We work with every type of client EMR and a huge array of other clinical administrative systems. All of these are feeding us data. These data are continuously updated, and we can track what happens to an encounter from the very beginning of the process all the way to the end. When we take an action, for example, the appeal of the denial of a hospital's claim, we record the action and the result we receive. The sum of millions of these actions and responses creates a really powerful data set. We consolidate all of those data into a single ecosystem, which is available to our applications, our algorithms, and our automations. We see over half a billion patient encounters. We work with over 500 hospital system customers and connect with more than 95% of payers across the country. As a result, we can identify trends and actionable insights across the full U.S. health system, the entire country. The real magic happens when we apply our proprietary intelligence to our data ecosystem. Let's take a look at that. Recall a moment ago when I talked about how we record an action and the result of the action. That's request and response. Those data and the application of our intelligence to the data create a powerful feedback loop. What we do is we use that feedback loop to optimize our processes. We incrementally automate tasks. We improve procedures on workflows that we haven't automated yet. They're still manual. We course-correct when we observe issues occurring. Each of these changes flows back into the data, and they provide us with feedback we can use to further improve. This is the most important thing of all. We use the flow of feedback to generate new intelligence and insight. We refine what we do. We watch how other players in the ecosystem respond to us. That is lather, rinse, repeat: continuous improvement of our processes and our technologies. To facilitate this, we use highly scalable modern technologies, which are based in the Azure Cloud. Some great examples of this include our use of Databricks and Snowflake. We also have selective but close collaborations with Microsoft on artificial intelligence and Automation Anywhere for robotic process automation. We drive our learning into those revenue cycle optimization engines I mentioned before. Those engines are where we turn our rev cycle expertise into proprietary software, which we leverage across the full breadth of the cycle. Now, I want to turn our attention to a foundational aspect of our technology. Let's talk a little bit about our journey to the cloud. Leveraging these leading-edge technologies we're discussing today and adopting born-in-the-cloud design patterns, managing these huge pools of data is hard if your core infrastructure is lagging behind. Fortunately, two and a half years ago, we began a journey to accelerate our engineering teams by centralizing and standardizing on best-in-class cloud platforms. This month, we completed the journey as planned. Right now, our most important application and data are live and in production in the Microsoft Azure Cloud. I'm proud to report that we completed this multi-year project on time, under budget, and with the original scope. We're already seeing the benefits. Azure makes it easier for us to control costs by scaling capacity up and down dynamically. It makes it much simpler for our developers to get access to the latest and greatest technology and to build new solutions leveraging that technology. Finally, it's simplifying the work of integrating with our clients, many of whom are on their own cloud journeys. We talk a little less about the cybersecurity benefits of this work, but infrastructure consolidation decreases the attack surface area, and the use of our public cloud allows us to take advantage of the massive cybersecurity investments that our partners are making. Leveraging cloud technology isn't the only way we can gain scale and efficiency through the use of tech. We also build and consolidate technologies that we can leverage across our service lines and the revenue cycle. So let me turn now to one of our most important revenue cycle optimization engines. I'm talking about our shared pricing service. This is not only an important effort for consolidation and enhancement, but it's also a handy example of one of those key engines. Let me talk a little bit about what we mean when we say pricing. In the context of revenue cycle, pricing means the calculation of expected reimbursement. That is, what a hospital system is owed for providing care to a patient. This is a really complicated process. It's hard to do, and it's critical to get it right. This is particularly true in our RPS business, where our bread and butter is correcting errors or finding missed opportunities for reimbursement on behalf of our clients. Today, we have four revenue cycle pricing engines in use across major service lines. Now, while these are a core capability that allows us to plan and optimize our work and achieve good outcomes for providers by knowing what reimbursement should be, keeping all four of them up to date and working correctly is a deeply non-trivial task. To do this, we have to review all of the relevant contracts, the letters of agreement, rate schedules, and the other documents that exist between our customers and payers. We turn those into structured data, and then we convert the structured data into models. The models are what enable us to accurately price accounts. Doing this work across four pricing engines is a coordination and researching challenge, to put it mildly. It makes it more difficult to leverage the intelligence we get from our data ecosystem. Fortunately, this is a problem that we are fixing. Through 2025, we are rolling out a single best-in-class capability for us to calculate expected reimbursement. Our new pricing service enables scale, and it enables us to have the ability to deepen pricing coverage, improve our accuracy, and reduce the total volume of manual work. Today, we are live for multiple service lines, and we will continue to roll out through the rest of this year. Now, let me turn it back to Steve, who's going to talk about some other ways which we are enhancing productivity. Here are two examples of how we leverage our platform to deliver significant savings and satisfaction for our users. We're constantly looking for ways to make work easier and faster for our thousands of revenue cycle experts who are doing work for our clients. The example on the left is for a part of our business called underpayments and a set of users who are doing what we call validation. On the right is an example from our DRG validation and charge capture businesses for users who are focused on invoicing. As I mentioned earlier, we run every nuanced aspect of the revenue cycle. And when we do that, our team goes deep to try to help our users make their job easier. You can see we do things like change the screen layout, add in additional data they need, deploy mini automations behind the scenes, and provide recommendations and guidance to make it easier for them to make decisions where work can't be automated. In each of these examples, you can see there's about $1 million of margin impact. What's important to you is that we're constantly looking for opportunities like this across our thousands of users. That sounds great, but I always tell my team, "If you want to know if you have good software, ask your users." You can see from the quotes here that our users are really delighted with these changes because it makes their jobs easier and faster. That's critical to us because it helps us to retain and prevent attrition from employees. Now, let's turn our attention to the second focus area, which is AI and automation. We believe there's a lot of runway for us to go in AI and automation with the combination of RPA, machine learning, and generative AI. Here, we have eight opportunities outlined on the chart. On the Y axis, it indicates the value potential in each of the areas. On the X axis, it indicates the complexity, which you can think of as level of effort and time required. The colors indicate to what extent each opportunity area is a priority for customers. You can see that customers tend to gravitate toward the areas that are most directly drive cash yield. Opportunities fall across the spectrum here. And I'll walk you through three specific tangible examples in a couple of slides to give you a flavor for what we mean by the opportunities are in each of these areas. Before I do, let me highlight over to the right that there are future opportunities that we're going to keep our eye on. An example is patient registration. We have thousands of registrars on site across hospitals in the U.S. helping to register patients as they come in for service. We can imagine a world in the future where some of that work is automated by AI. But the technology has significant maturing to do before we feel like we're ready to put it on this chart. That still feels a few years away. So we'll continually evolve this chart and our evaluation and assessment of these opportunities over time as the technology matures and we listen to clients and learn more as we implement our solutions in each of these areas. So let me tell you a little bit about how we approach AI and automation opportunities. We take an approach that allows us to deliver fast incremental impact that can grow over time while we learn about how each of these technologies work and how effective they are at driving results. We take a four-phase approach. The first three phases include a human in the loop. The last phase is full autonomy. The first phase is what we call summarization. This is where we use an AI to provide a summary of unstructured text to a user. Or we use an AI to intelligently search a document and provide the information that we think the user needs right when they need it. That saves the user some time and makes them more effective. The second step is what we call quality assurance. This is where we have an AI predict an actual output. It might predict a code or it might create a document that we think is that we are hoping is complete. And we compare that output to what a person generates. We use this to assess the quality of the human of the work that the people did. And it allows us a good way to assess the accuracy of our AI. The third phase is a staff or AI assistant. This is where someone doing work gets a recommendation for an actual output. That could be a recommended code or a draft document that's created by AI. That's helping that person and perhaps a reason, a rationale for why that is the prediction. This helps our team become more effective and more efficient. The last step is full autonomy, which we'd only get when the AI is sufficiently accurate and it makes sense given the rules and regulations around that process. So let me turn to three examples that are real-life examples of live in-production use cases that we have that are driving results so that you get a sense of what we're doing in each of these areas. For the first example, automated clinical appeals, we have a video to show you. But first, let me walk you through the middle and third column. The middle column is physician coding automated quality assurance. This is where we have an AI take a medical record from a client system and automatically predict a physician E/M code. We then compare that predicted code to what a manually coded chart's code. And where there are differences, we funnel that volume to a quality assurance group. Previously, we had the quality assurance group random sampling 5% of all the charts that are manually coded. Now, we're effectively QAing 100% of the volume. And with that subset that goes to the QA group, where there's a difference between the predicted code and the manual code, our QA group is finding that the AI is correct a little more than half of the time. I can tell you, when you talk to clients about physician coding, they say it's a hot-button issue because inaccurate physician coding directly drives compensation for physicians. And if physicians think there's a coding quality problem that's affecting their compensation, they are very vocal about it. So clients are extremely excited about the solution and the idea that we're leveraging AI to dramatically improve quality checking. The next steps in this application area, we've already started to extend this to other types of coding segments, and we'll be moving to provide an AI assistant with a recommended code to the manual coder. The final column is another example. It's automated call quality insights. In our scheduling contact center, we take every call transcript and use an AI to evaluate that transcript and assess the call reason code, the caller sentiment, and the call quality. Again, this is another example where previously we had a quality assurance group random sampling a small percentage of our calls to do quality checks on them. Here, we now are effectively quality checking 100% of the volume, which is leading to better training and education and coaching for our agents, which will drive better patient satisfaction. Also, it gives our management team the ability to look at those AI-generated call reason codes and its assessment of the call to come up with ways to drive more call volume to patient self-service and to come up with ways to reduce average handle time. This is important to clients as it drives patient satisfaction, and it's important to us for our own efficiency. Now, with that, let's give you an overview of what automated clinical appeals is with a video. Today, R1 handles nearly 433,000 clinical denials per year. That's about 36,000 per month for our clients. Managing and appealing denials is a tough process. The old way was to spend up to an hour or more for every appeal. Investigating why it was denied alone involved reviewing a patient's medical record, EMR notes, and payer documentation. After that, clinicians still had to devise an appeal strategy and start the appeal process. R1's AI-assisted appeals engine drastically streamlines appeals, reducing our clinician processing time to about 15 minutes for some high-volume clinical denials. The engine uses generative AI to automatically ingest, parse, and summarize a patient's medical record. It doesn't just pre-populate templates. It creates a concise appeal document, including two paragraphs that accurately reflect the patient's medical history and treatment and persuasively articulates the case for medical necessity. The model summary must differentiate between essential and unnecessary facts. Accuracy is paramount, as is brevity. The model also needs to be fast. Even if the medical record is hundreds of pages long, the summary must be created in less than a minute so the clinician can move to next steps quickly. This technology evolves the role of our clinicians from appeal authors to appeal editors, fine-tuning the document with additional arguments that may be required based on a specific denial rationale. After editing, the system then automatically sends the appeal to the payer. By spending less time searching for information and devising compelling arguments, our clinicians can create more appeals in a day, which accelerates cash collection for our clients. Over time, as more clinicians use and edit the auto-generated appeals, the R1 platform records and continually learns from that data, improving the effectiveness of auto-generated appeals in the future. AI efficiencies can increase productivity, free up time for tasks like identifying gaps or areas of documentation improvement, and accelerate cash flow to clients. Great. So in summary, now you have three examples of where we are deploying AI with live in-production use cases that our revenue cycle experts who are using them are thrilled with today and makes them optimistic about all the opportunity we have going forward to use these and other tools. With that, I'll turn it over to Brian, who can talk about how we're developing these types of solutions. All right. So clinical appeals are an example of an area where our approach to building these scalable and modular capabilities can really pay dividends. We've got a simple and repeatable pattern that we use for this kind of development. We start with a concrete use case. In this instance, with appeals, that means drafting appeals for use in our modular denials business. We're taking advantage of generative AI technology in this instance, but we can use this pattern generally. Because our tools are directly embedded in the provider workflow and the users of those tools are our actual employees, it's really easy for us to deploy this, learn, and improve rapidly as we see results. We can do this a lot more rapidly than if we were simply selling the technology. Remember, our tools are deeply embedded in the workflow, and it's our own employees running the tools. Key to keep in mind is because we operate across the full revenue cycle, we can reuse the same services in multiple lines of business. That's an important point. Our technology investments can be reused across multiple solutions and lines of business, regardless of where we started the development effort. We take that approach with all of our AI investments. Of course, we don't confine ourselves simply to building brand new capabilities and services. We have a pretty significant body of already existing technology, and we're constantly looking to iterate and improve, find new opportunities to innovate, and deliver new values in those areas as well. Let me turn your attention to what I consider probably one of the top three gnarliest challenges in all of healthcare. This is a place where we have existing technology, a lot of it, and that's prior authorizations. Let's ground ourselves in what an authorization is. You've probably encountered this in your life as a patient, but authorizations are pre-approvals that insurers require before services provided to a patient. And if the authorization doesn't happen, payment or even treatment can be denied. But these are really tricky. Payers don't require them for every single service, and payer policies change really frequently. We have to keep up to date with all of those changes basically in real time. It's doubly challenging because many payers outsource the approval process to third parties. The third party can vary by the service being provided even within a given health insurance company. We've got to stay on top of exactly which payer portal we need to use to request authorization. A third complication in all of this is that requesting an authorization often requires answering what comes down to a variable series of questions. For example, you might initially be asked to answer three questions about a patient's condition. Depending on exactly which answers you provide, you're going to get asked additional questions. For example, consider an order for an MRI. The imaging might be for diagnosis, or maybe it's for planning treatment. Depending on exactly the reason the clinician ordered the scan, the insurance might want more clinical details. They might want to know if alternative methods were considered. There's a whole range of things that might be stimulated simply by the initial set of answers. The intent of this is to gather sufficient details to know that the clinical decision is justified. As the MRI example shows, auth requests can require providing a lot of evidence, and payers are really picky about exactly what documents and information should be provided. The bottom line is that the variable nature of authorizations makes them difficult to automate, and they're labor-intensive and error-prone if you do them manually. Authorizations are a major point of friction in the system. We estimate that about 92% of delays in patient care are attributable to authorizations. The volume required for auth is growing rapidly. We've seen about a 40% increase in requests for auths over the last 18 months and about a 31% increase in the amount of effort it takes to fulfill the auth request. Automating prior authorizations could save providers over $21 billion a year, so it's a big opportunity. We've actually automated a lot of this process already, but we believe we can do a lot more. Thus far, we've spent most of our time and effort on automating the process of determination, that is, figuring out whether or not an auth is actually required for a given procedure, for a given payer, at a given health system. That saves our team a lot of time just knowing if they have to pursue the auth process at all. We've also made good progress automating authorization request status checks, whether or not an auth has actually been approved, but we have more room to run there. So what's really exciting to me is that now, thanks to Generative AI, we can actually envision automating a significant portion of the auth submission process itself. That's because we believe we can use large language models to effectively navigate the sequential questions payers ask and extract and format the right information from the patient's medical record into the appropriate responses for the payer. These tasks are hard for knowledge workers, and they're really training intensive. They've traditionally been really difficult to automate as well. But the capabilities of large language models are really extremely well-suited for attacking the problem. Now let me turn it back to Steve, who talked a little bit about new solutions we're building. Our third focus area is new solutions, where we leverage our platform, our AI and automation capability, our data, and our revenue cycle expertise to develop new modular solutions to generate growth for us and great results for our clients. The first example is Insurance Discovery. We launched a new solution earlier this year called Insurance Discovery. What it does is it finds patient insurance coverage, of which our health system client was previously unaware. This leverages our payer connections, our data, our deep expertise in the revenue cycle, and obviously our platform. When we deliver coverage back to a client, we only deliver it if the coverage is relevant to the service that was provided to the patient. For example, if a patient comes in with a broken arm and needs surgery, but the coverage we find is dental insurance, we're not going to provide that coverage back to the client. But doing that for every type of service, for every one of our clients, for every payer across all the nation is a very hard task to do at scale and can't be done without the type of expertise we have to design the business logic and our platform like ours that scales. We previously at R1 used a vendor to do this service and have been replacing that with our own self-developed solution. We found that our solution finds 10% more coverage than the previous solution we were using. That message has resonated as we brought this product out to market to new prospects, as has our plans to further differentiate the product with our automation capabilities, where we can automatically post the insurance we find back into the client's host system. We are offering to provide billing services to clients to prevent them from having to do any work at all as we deliver incremental revenue back to them that otherwise would have been lost. The second example is modular coding. Our clients struggle to find and retain high-quality coding talent. We have an offering that leverages our global technology and services and will increasingly benefit from our AI-assisted coding initiative. Clients can choose for us to perform their coding function for them, or they can choose for us to be a supplement to their coding function to take volume that they can't handle. Where we've deployed the solution, we found we have the ability to accelerate coding completion by 30%, which directly impacts patient cash collection rates and timing. The third example is a patient receivables collection service that we launched just toward late last year. Our clients struggle to maximize the collection of the patient financial obligations that are owed to them. We leverage our combination of data, our billions of historical payment transactions data, and our machine learning capability to accurately model a patient's propensity to pay and to automatically create patient payment plans that are tailored to that patient. We couple that with our global contact center technology that optimizes the way information is provided to the patient and the way payments are collected from the patient to make it really easy for patients. Where we've deployed the solution, we've been able to increase patient payment collection rates by 10%. And equally importantly, we do it with high patient satisfaction because we make it convenient and easy for them. And this solution is resonating with clients. We'll continue to innovate new solutions as we invest in our technology platform and capabilities, and as we listen to clients and find new opportunities to help them in ways that meet their needs. That's what we call meeting clients where they are. Now that you understand our product and technology vision, our strategy, and our focus areas, you can see how these come together to create four differentiators in the eyes of customers. First, our platform. It's at the center of the healthcare ecosystem, and it's at the center of everything we do. It contains our unmatched data that allows us to see patterns and trends that others can't given our scale. Our expertise is embodied by our 30,000 global revenue cycle experts. We run every aspect of the revenue cycle, and our proprietary knowledge is embedded in our revenue cycle optimization engines, our standard workflows, and our algorithms, all of which are applied to the data. Together, our expertise and our platform form the brain behind our R1 operations' ability to deliver great customer results. Automation will continue to be a specific, intense focus going forward. It's not just RPA. It's also machine learning and generative AI, which will allow us to achieve the next level of efficiency and automation. Finally, we'll continue to innovate new solutions that leverage our combination of our platform, our data, and our expertise as we listen to clients and find new ways to help them that supports our growth and delivers them great results. As we go to market, these differentiators will deliver better customer results, will help expand our ability to win, and give us great confidence that we'll be able to achieve our margin goals over time. Thanks for the time. Now we'll turn to Q&A. Thank you, Steve. As a reminder, please feel free to continue to submit questions as we move through those that have already been submitted. Brian, I'm seeing several questions that have come through regarding cybersecurity. Could you address how R1 thinks about its current cybersecurity infrastructure if you plan to invest significantly more, and how much cybersecurity protection or compliance is a focus for our customer base? One thing that's true when you are in the healthcare business is that protecting patient data is very important. We think about it all the time. We talk about it all the time. We look at it as one of the most important trusts that we carry, and it's very important to us. If you've been paying any attention at all, you've noticed that the last, I'll call it, six months in healthcare technology from a cybersecurity perspective have been super eventful. R1 has been right in the center of that, helping and supporting our clients who've been impacted by some of these events. One of the things that we've had the opportunity to learn is that our cybersecurity posture, our infrastructure, the technology in which we've invested is really sound. That's been a great thing to see and to experience personally. We've certainly learned things from that, but we're very satisfied with where we are. A key takeaway in learning from that is that we got a lot of benefit out of the fact that we've been very intentionally keeping up with technology, continually modernizing, assessing, and improving the tools, technologies, and the team that we have. I think that's paid a lot of benefits for us, and we'll continue to focus there. I'm satisfied with our level of investment, but we're always thinking about it and doing whatever we need to do to make sure that we're in a good posture. In terms of customer retention, I think that everyone in healthcare has been aware of and has thought a lot about cybersecurity. I think one of the things that's interesting for me is to see the degree to which people are starting to think about this, not just as a reputational risk or a regulatory risk or a compliance issue, but as a real operational matter that is a major point that you have to be thinking about within the business. Two days ago, R1's CISO and I were meeting with the team from one of our large enterprise customers. We had their CIO and their CISO in the room, and the meeting was entirely focused on cybersecurity. That doesn't sound surprising. What might be a little more surprising is that also in the room for an hour was the CEO of that client and a lot of the other key members of his leadership team. The meeting was sponsored by my boss, our CEO, Lee Rivas. The only intent was to talk about cyber. My point of view is that there's a lot of attention on this right now. That's probably not going to die down, and that R1's investment and posture in this area is a strength of ours that I'm happy with our position today. Thank you, Brian. Steve, the next question is for you. Generative AI has obviously received a lot of hype over the last year. What would you say are realistic expectations for the technology over the near, medium, and long term? As we said in the presentation, we're very optimistic about the potential for AI and automation, including generative AI, going forward. It gives us great confidence in our ability to achieve our margin goals over time. We shared the eight opportunity areas, and we gave you very specific, tangible examples that are live in production use cases to make it real in terms of how we're using this today. We have a lot of optimism around that. We didn't share any specific estimates. As you can guess, in our position, this is still newer technology that is evolving very rapidly. We have a great relationship with Microsoft. We have a lot of visibility into improvements that are coming down the road, and they are fast and furious. It's hard to even imagine where the capabilities of the technology will be in a year or two years. That makes us bullish on the overall opportunity. Great. Thank you. Steve and Brian, I will open the next question up to both of you. Does RCM own its data for the purposes of training modules? And if not, what rights to claims, EHR, and other data does RCM have? I can start. We have the right to use the data from our clients and from our operations for the benefit of providing the services for which we're contracted with that customer, including AI and automation use cases. Great. Yeah. And for me, the powerful thing is that we've got access to this huge range of data covering the whole healthcare ecosystem and that the rights we have to use that data in service of our clients give us what we need to build a lot of exciting new technologies and services. So one thing I don't feel at R1 is that we haven't got the data, which is a fantastic way to be as a technologist. Great. Thank you. Steve, I think this next one is for you. R1's intelligent automation now unlocks the equivalent of more than 5,000 FTEs, which you mentioned. Looking out 5-10 years, where do you think R1 can get to in terms of automated FTEs,% of shared service work being automated, and how would you quantify that margin contribution so far? The important thing is going forward, we have a lot of optimism about our ability to use many technologies that come together to drive more efficiency. We've been very successful with our intelligent automation initiative, which is largely RPA. But I expect that going forward, it's going to be more of a combination of technologies: RPA, machine learning, generative AI, workflow solutions that come together to deliver not just automation, but also efficiency for our users. We're very bullish on this. We don't provide specific estimates, but you can tell that we feel very confident in driving efficiency. The second thing I would mention is sometimes there's too much focus on just that one benefit, which is efficiency. Hopefully, you saw from the examples I shared that in many of those examples, we can deliver better client results. If we have more capacity to appeal more denials, then we'll be able to appeal more denials, and more will get overturned. And that results in more cash generation and reimbursement for our clients, which is good for us and which is good for them and a competitive differentiator. There are many examples there that I shared that provide a better patient experience and increase patient satisfaction. And when I go talk to clients, that's top of mind for them because our clients are trying to make sure they maximize their share of the care that they're giving those patients. They're trying to retain them and make sure that that patient is loyal to them and thinks of that health system as their health system. Great. Thank you. From a technology perspective, what tends to cause the most friction during an onboarding process, and what are the factors that tend to determine the complexity of an onboarding for a new customer? Sure. I can take that, Brian, and then you might want to touch on it. The most friction is likely caused by just getting the data from clients. That can be driven by a number of factors. The data we need from clients varies depending on the specific solutions that we're deploying for our client, the extent to which they've consolidated their own internal systems, the extent to which they have their team and have capacity to work with our team. And so that is a variable that can be a hurdle to getting a solution set up for a client. We have a large team working on this, and that's a specific focus area for our product and technology team to not only be able to make sure we have the best quality and most granular data from clients, but also to make it easier and faster for our clients to be able to provide us that data. Because when we get data faster from clients and are able to stand up our services earlier, we're directly able to provide more benefits to the client sooner, which is important to them. Great. Thank you, Steve. I think this one is for you as well. You've had a partnership with Microsoft for several months. Have you seen meaningful evolution of the Generative AI technology throughout the partnership? Has it assisted in accelerating timelines for projects? Have there been any unexpected shortfalls? And what does Microsoft contribute to the partnership? Sure. I'll comment on the relationship and the partnership from a product perspective, but I'd love, Brian, to comment maybe on the technology benefit that Microsoft offers. We have a great relationship with Microsoft. It's been a great collaboration. Our team is meeting with Microsoft team weekly. We have a quarterly steering committee with them. They have absolutely accelerated our ability to design these solutions effectively, scale them, and to get them deployed. They are working hand in hand with us in the weeds, not just on architecture, but we have had in-person working sessions with them where we actually collaboratively try to develop and work on developing a new solution. So it's been really, really helpful to us. We have a great relationship. We've talked about additional ways to work together, and I'm looking forward to pursuing those going forward. I agree with Steve on the great partnership that we have with Microsoft. We do a lot of work with them on AI, but really, we're collaborating with them across a whole range of concerns in our business, everything from cybersecurity to infrastructure concerns. We meet with Microsoft constantly. It's a great partnership. From a tech perspective, thinking strictly about AI, I think that one of the real powerful things about the partnership with Microsoft is their posture as a platform company. We think about our expertise as being very specific to healthcare, healthcare technology, the revenue cycle, the data in that area. And we think about Microsoft as bringing to the table access to a lot of technology that would be more difficult for us to acquire and learn totally independently. So Microsoft gets a ton of attention around the OpenAI relationship and their investments there, but possibly underappreciated is the way that the use of the Azure platform gives you quick access to a range of other models and other technologies and the degree to which Microsoft is making investments in other aspects of AI as part of its own roadmap. And that by partnering with Microsoft, we get access to those technologies. It allows us to focus on what we do, our unique advantages in terms of access to the workflow, access to the data, presence in the revenue cycle, and leverage the technology that Microsoft brings forward. It's a really powerful model and a powerful partnership. We're super satisfied with it. Yeah. Can I just give one example? So when we start using these new technologies for generative AI, you come across challenges. For example, to summarize or create an appeal, a draft appeal in an automated way from a 50-page medical record is a very different challenge than doing it from a 300- or 400-page medical record. You quickly get into a lot of complexity with what's called RAG or retrieval-augmented generation in terms of how you store the data and how you make sure you achieve accuracy while you're also achieving scale. It's similar. When we have problems like that, Microsoft is on the phone with us immediately because they've encountered this pattern of problem with other clients as well, or maybe even internally, and they give us great guidance rather than our team having to figure it out themselves. That's just a specific example of where at our scale and that kind of relationship is really required with new technologies like this to be able to deliver the type of solutions we aim to deliver at the pace that we're doing. Great. Thank you both. While AI will make your technology more powerful, could it potentially disintermediate R1's end-to-end platform as the value of offshore labor diminishes? Said otherwise, could AI shift health system preference to more modular software-like solutions as opposed to outsourcing? Yeah. I have two comments on this. First, as a reminder, we have two sides of our business. We have an enterprise business where we will run all or most of your revenue cycle. And we have a second side of our business, our revenue performance solutions, our modular side of our business where clients can just engage with us for modular solutions. And as Brian explained, the way we're developing our AI-enabled tools, they're relevant for both sides of those business. So no matter what clients want, we have solutions for them. The second thing I would say is that we see health systems increasingly wanting major providers that are going to reduce them from having to deal with all this complexity. They don't want the complexity of many, many point solutions. They don't want the complexity of lots of relationships. They understand the limitations of point solutions of being siloed. What's attractive to them and what's attractive to us about what we do is that we have many solutions, and we have all the data, and it's all in one place. Our AI and machine learning is all working on that one combined data set. We have visibility into patterns and trends that point solutions just can't. I think health systems going forward, the trend we see is that they want to minimize complexity and simplify their world so they can focus on the clinical side of the business and finding great patient care. I don't see that trend changing. I agree with that. I would also say that I'm a couple-time healthcare entrepreneur. I understand how healthcare technology works. Startups bring strengths to the table. I think that it may be underappreciated in R1's model that we solve what I think are the three primary challenges in sort of healthcare entrepreneurship, technology entrepreneurship in that space. You've got this question of, how do I get paid? How do I get access to data? How do I get people to use my technology? Often you blow past the first thing, try to get access to data, find that's hard, you end up with trying to get people to use your technology. When I think about R1, we have all those things already. I think that that gives us a huge advantage in the innovation cycle in that we're already in the payment flow. We know what our business is. We have an attractive TAM and SAM. We've already got access to the data, and we have the ability to deploy it in our own workforce. We don't have to convince anyone, and that enables us to iterate rapidly. So totally agree with Steve's remarks on the sort of platform desire. But I also think as you think about the modular business or just our own ability to innovate, given our position in the marketplace, I think it's really attractive. Thank you, Brian. How quickly do you see a return on investment dollars in terms of performance? When you look at any individual initiative on our roadmap, on average, our return on investment, our payback period is a little less than 24 months. The way we approach development and deployment is to deploy and then fast. We're biased to deploy fast and then iterate. So we're typically generating business results well ahead of that and earlier, but that's a general timeline you can think about. Great. And then would love both of your thoughts on this. Can you talk about the advantages in scale in terms of your ability to invest in technology, access to vast data, and what that is relative to point solution vendors? Brian, you want to take it? I mean, I wish I'd saved my answer for this question because I feel like it's really the same thing that I was addressing a moment ago. I think the advantage that we have is that we have scale, that we have access to data. I really think a lot about our ability to experiment and test and drive things into the workflow really quickly. So I do think those advantages give us some opportunities that they certainly exist for other players, but we've got a built-in, baked-in audience and capability that one would struggle to create if one was creating it from scratch. Okay. Thank you. And then one of the benefits of the Cloudmed acquisition was that it provided the ability to combine R1 data with Cloudmed's platform. Can you comment on the status of the combination and any tangible benefits that have come from the integration? Sure. I can comment. Yeah. The integration has gone really well. We've made a ton of progress on it. It's been a huge focus of Brian's and mine to bring all the data into one place. The platform that we speak of, the foundation of it, is the Cloudmed AI platform. So having the data together, it's generated new insights. We've been able to generate new solutions, including some of the ones we talked about today that are leveraging the power of the data from both of those solutions. Moreover, we have lots of examples where, for example, in Cloudmed, we have our rules and algorithms that we're running as a safety net provider where we identify mistakes that were made or opportunities that were missed to generate more revenue for our clients. That could be, for example, a coding mistake that was made. We've been able to deploy those rules into our end-to-end business to improve our coding upfront, upstream, which is better for clients, which is better for patients, which is better for reimbursement. So we have a lot of examples of where we've been able to strengthen the quality of what we deliver to clients and our performance for clients with a combination of the capabilities. Thank you, Steve. I have a question regarding full autonomy, kind of going back to the approach you discussed in the presentation. At what point do you feel AI applications can have full autonomy? And is there some sort of metric that makes you comfortable to make a step towards that final kind of full automation step? You really have to consider each individual use case. I'd love to say when we achieve 95% or 98% accuracy, that's good enough, but that's really insufficient. That's just one criteria. Another criteria is, well, what happens to the 2% or 3% or 5% or 10% that is not perfect? What if you're submitting an appeal to a payer? What happens then? What is their reaction? What is the additional thing that we have to respond to? So you really have to consider it within the context of the entire process to know what is good enough. We do that by applying specific criteria per use case. We do it in collaboration with our operations team, who are the revenue cycle experts, to make sure we consider all angles and all the implications. The second thing I would say is the real compelling thing to me about going through those four steps is that even if you land on step two or step three, for sure, we can drive a lot of efficiency and much better customer results just sitting in one of those steps. We don't necessarily have to get to full autonomy for 100% of the volume or 100% of use cases in order to drive a lot of value. And so we'll evaluate each opportunity as we learn more, and we'll see where it makes sense. Great. Thank you, Steve. I think our last question based on time will be, from a technology perspective, what is the bottleneck that could prevent you from moving faster to implement items you've discussed? You want to try, Brian? I mean, it's one of the interesting things about innovation is that you're never quite sure where the punch in the face is going to come from. I think it can be a little bit challenging, right? There are occasions where we have to spend a lot of time and energy getting the data into the right state or where there are unexpected challenges in connecting into the ecosystem. I think a bottleneck is always going to be just how much capacity we have to take on the opportunities. As I look at the list of things that are sort of super interesting opportunities, there's just a lot of them. And we can't do them all at once. There's a physics problem you run into in terms of available resources and the amount of intellectual capacity you have to do so many things at once. So we're a little bit spoiled for choice sometimes. And so I think that an aspect of the bottleneck I think about is less technical and more just the importance of keeping focus on the most important things and proceeding incrementally and iteratively to address them and deliver value. Great. Thank you, Brian. I'll turn it back over to Evan to close us out. Thank you, everyone, for joining us today. I hope you've learned more about the strength of the platform, our AI advances, and the strength of our roadmap to keep us at the forefront of the revenue cycle marketplace. As always, if there are additional questions, given time we had to stop at the hour, you can always reach out to myself or Megan. Again, thank you very much, and have a good day.