Oscar Health, Inc. (OSCR) Earnings Call Transcript & Summary

Welcome to Oscar Health's 2026 Investor Day. Please welcome Chris Potochar, Treasurer and Head of Investor Relations, to the stage.

Good morning. Welcome to Oscar Health's 2026 Investor Day. It is great to see so many of you. I know it's conference season. So thank you for joining us this morning. We appreciate you taking the time. I'd also like to welcome everyone that's joining from the webcast this morning as well. Thank you for joining. For your reference, all the materials that we will present today will be available a little bit later this morning, and you can find that on our Investor Relations website at ir.oscarhealth.com. As a reminder, any remarks Oscar makes about the future constitute forward-looking statements within the meaning of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995. Actual results may differ materially from those indicated by those forward-looking statements as a result of various important factors, including those discussed in our most recent annual report on Form 10-K and other filings with the SEC. These statements are based on our expectations as of today, and we specifically disclaim any obligation to update them. Okay. So before we begin, I just want to hit on and briefly walk through today's agenda. First, Mark Bertolini, Oscar's Chief Executive Officer, will share our vision and strategy, including how we are building the premier consumer health care company. Following Mark, Scott Blackley, our Chief Financial Officer, will discuss our financial outlook, including our updated 2026 guidance and longer-term financial targets. We'll then have a quick break. And then when we come back into the room, Janet Liang, President of Oscar Insurance, will discuss how we are strengthening our leadership in the individual market. Then after Janet, Mario Schlosser, Oscar's co-founder and adviser to the CEO, will discuss how Oscar's scalable technology and applied AI enables durable earnings power. So that will wrap up the presentations for the day. We'll ask all the presenters to come back up to the stage, and we'll do a question-and-answer session and then to close the meeting out, Mark will provide some final remarks. And then for those in the room, we will have lunch available. We'd ask that you would stay for lunch if you can do that. So good agenda for the morning. We're excited to get started. So with that, it's my pleasure to introduce Oscar's Chief Executive Officer, Mark Bertolini.

Please welcome Mark Bertolini, Chief Executive Officer, to the stage.

Thank you, Chris. Good morning. Thank you for coming. It's always so nice to see some of the smiling faces at an investor conference, you try smiling. The information will come back, they come across a little better. We're going to discuss our vision and our strategy today, but I want to make a few comments about what strategy really means. Strategy is not buying an asset. Strategy is not coming up with a good idea. Strategy really is finding fundamental capabilities that differentiate the organization and applying those to unmet needs in the marketplace. So the journey we've been on for the last 3 years has been very much about creating those differentiating capabilities, those insights. We bought an asset along the way, but it was all part of having fundamental capabilities that meet upcoming changes in a market where unmet needs have not even met. So we're going to talk about that a little bit as we go through the discussion here. 50% of consumers feel that health care fails society's needs. We have $200 billion in active consumer medical debt across the United States today. Unacceptable, I would think, in the 21st century as a norm. And our mission is to empower consumers to build health care around them. They will drive innovations and better value and quality. The whole stream of consciousness here is to think about getting to a place where we're going to have consumers, through curated pricing, be able to make value-based decisions that matter to them like every other piece of their household budget. Health care is #1 is the only place where they can. They don't get a bill -- they don't know what they're paying until they get a bill later on. And so if we can create that ability in consumers, they will do the fundamental work longer term of reducing the costs and driving out high rents in the health care system as they've done in every other market that they've had that opportunity. So our 2026 earnings, we raised them this morning to $600 million to $800 million. Our SG&A expense ratio is the same. We improved our MLR for the year based on what we're seeing in our underlying medical costs -- sorry. Wrong Slide. First -- sorry, good stage direction. We have a deep bench -- this is a gallery of rogues, but we have a deep bench, and we have the A team. And the 1 thing I want to say about this group is that this group gets up every day thinking about this business. They're not worried about other parts of the business. They're not worried about capital allocation between business, they're focused on the ACA. And that results in every employee in the organization being focused on the ACA. And one of the amazing things I found when I joined the company in 2023 is every employee in this company is engaged in a mission to make health care better. We have employees that have been with us the whole way and they're continuing to be involved. So this leadership team is a pleasure to work with. They lead great people who every day work really hard. Now guidance. So guidance is up $600 million to $800 million from $500 million to $700 million. Our SG&A ratio, same range, MLRs improved $18.7 billion to $19 billion in revenue, not much change there as well. But what it's done is it allowed us to make investments in other parts of the business and the new stuff that's coming along. We'll show you some numbers today. Scott will go through them in more detail. Those numbers for 2029 don't rely on a whole lot of what we do in the new businesses. But the investment in those new businesses is part of those projections. Those investments are the hard work that we need because you never really show your work until you actually can prove it works. And so all of the investments that we've been making over the last 3 years, what we'll make over the next 3 years, we'll prove to change the American health care marketplace and allow individuals to purchase their own health care the way they want. Three years ago, 2.5 years ago, we gave you these commitments. At the time in '24, we had $9.2 billion of revenue and a 0.6% operating margin. Since then, we've had a 43% CAGR, which exceeded our 20% CAGR, and we've had 260 to 360 basis point improvement with a commitment which we are reaffirming of at least 5% in 2027. Here are our long-term key strategic objectives, to become the #1 consumer-preferred AI-native carrier through our strategic ACA advantage. And the point about this -- the issue around this point is we are not spending billions of dollars like the rest of the industry, health care industry, rationalizing platforms and cleaning up our data. From the very beginning, this company built a platform that was single-threaded, cloud native, with one version of the truth, which has allowed us to implement AI at scale with significant bottom line results to the company. And Mario will come up a little bit later and talk a bit about that. Second is to unlock the full potential of the individual market to support health care needs for all Americans. We do not believe that the ACA is the last place for this company to be. The TAM is not big enough. The market opportunities aren't great enough, the competition would be too fierce. So our view is, is that we need to continue to expand the market opportunity for individuals to have the same opportunity that people in the ACA have in having to pick their network and also to pick their product. And then finally, build the leading healthcare place to serve people's health and financial goals in a larger individual market. It means more than insurance. It means GLP-1s. It means all the supplements that people buy, it means all the lifestyle things that people buy in order to impact their health and live a healthier life and enjoy life. That's what the Lucie marketplace is all about. And I'll show you some of the statistics around that a little bit later. So let's talk about the future of the consumer health care marketplace. The era of employer health insurance is ending. 94% of employers are looking at -- looking for a new solution for their health care costs. I talked with the CEO of large multi-state company with 40,000 full-time employees, and he says, "I can't buy health care, economically any longer. My rates are going up on a self-funded basis, double digit. There has to be a better solution." And on top of that, we want to be able to build the leading payer-agnostic marketplace for health benefits, products and services through Lucie. So the secret behind what we built is not that we built it for us. We built it for everyone. We're making a market. Lucie with all of our competitors engaged, all having access to the same technologies including some of Oscar's technologies, think about the new +Oscar. They will be able to use those tools to compete in the marketplace, and they will pay rents on that capability. So the Lucie marketplace becomes a place where all competitors can build a broader market. It takes me back to the early '80s when [indiscernible] friends when and I started an HMO in Detroit, and we got together with all the other HMOs and said, how do we take all that business away from the big carriers, who have all the money, all the distribution. And by the time, 10 years later, all those big carriers, other than that, we're out of the health insurance business for all intents and purposes. So this is our opportunity, create the market for the better good of the whole industry and all Americans and then compete within it. So currently today, the ACA market has a total TAM of 49 million individuals. There are 19 million in the ACA, and there's still 30 million on [indiscernible]. That's our target. You can see the rest of it, and we'll talk about it as we rest through the presentation, but the opportunity goes to all 350 million Americans. The far right column is 100 million Americans and public programs. And today, we're already having conversations with state-based exchanges and governors about doing demonstration products in Medicaid that get them out of math map and allow those Medicaid members to buy through their local state-based exchange. Governors are tired of waiting for the federal government to come up with a complete solution. They're starting to take their own action. And we've actually made regulatory changes in those markets for ICHRA to make that happen. The average American changes jobs 13x. So the other thing that's going to happen and people are going to be moving to the right on this spectrum, AI is going to displace the workforce or cause people to be more gig workers or more part-time workers and there isn't a simple solution for those people today. And we need to find a way to do that. And ultimately, the 1,000-plus employer market will come along. The 2 complaints they always have is how do I know my employees will pick the right plan, and secondly, how do they get the right network. And what I've told them is because all of our competitors are on the Lucie marketplace, when an employee goes to defined contribution, they can buy any one of the networks from any one of our competitors. And those networks are generally narrow networks. So we have the best PPO in the country at narrow network rates. That's a fundamental difference than where we were before as just the ACA. But we'll talk a bit more about that in a minute. So the ACA has proven to -- it has improved access in a meaningful way. 73% of individuals in the ACA rate their experience as positive. I can't say that for the rest of commercial insurance. The foundation of a more modern consumer-driven health care system is the ACA marketplace. Not everything about the ACA works, but I believe the network model and the way we underwrite the network to price versus underwriting the members because we can't underwrite them and the risk adjustment mechanism that we use to level out the risk across all carriers at the end of the year creates a very solid and stable market over the long run. And when you get to 200 million, 250 million, 350 million Americans, morbidity changes don't impact the model. It becomes a very stable marketplace. The ACA today, here are the people that are in it today. 27% of U.S. [indiscernible] are ranchers rely on the ACA. 17% of gig workers rely on the ACA. 48% of enrollees are from small business owners, entrepreneurs or their employees, where the small business has just given up on insurance and their employees have gone into the market to get products through the ACA. It's the backbone of the U.S. workforce. And by the way, these populations represent more than 50% of the GDP of the United States. And they have the weakest support system in the United States. So it's essential to the U.S. economy. We've taken almost to half the uninsured in America since 2014. We're instrumental reducing economic burden of $245 billion worth of uncompensated care as [indiscernible] from [ '14 to '25 ]. And we continue to address national issues of 28 million Americans are still uninsured, still haven't found their way. And being in the hospitality business that I'm in now, I can see it because the employees that we serve as that work as waiters in our restaurant, they don't know how to get the ACA. They didn't know they could get it, the part-time employees at [indiscernible] supermarkets, 34,000 of them, more than 80% could get subsidies, but they didn't know it. So the whole idea of helping people understand how to access the system with products that matter to them is going to get at that 28 million uninsured. It's the fastest-growing market of any other segment in America. And we expect that to continue. It has more choice than any other market in America. If you work for a large employer with [indiscernible] lot of you do 2, 3 plans. And it has greater competition. More and more people have entered the market, while some of the big carriers have left the market in large part because they can't manage their networks effectively to get to a narrow network model. It has created great price stability in employer premium in 2020 through 2025. So here's our first projection, 2029, 23 million lives in the ACA. 2 million will come out of through Program Integrity. We're in the middle of that process now as a lot of you have heard with CMS in spite of what's going on in the legislative side, through regulation, so to get at some of the concerns they have. But we still believe that 36 million Americans will grow into the marketplace, 3.4 million new members under the ACA in 2026. Choice adoption is increasing to 22.5 million in 2029, 200,000 new independent freelance workers market -- enter the market every year, 900,000 new individuals in the ACA due to unemployment each year. So there's a basic underlying growth that occurs in this marketplace. And so we're calling for 6 million growth over the next 3 years for a total of 23 million lives in the ACA in 2029. But how do we lead the ACA market? We have a seasoned leadership team. We have a culture of product and network innovation. I'll give you an example. In 2025, we bought a product out of Connecticut company that had 3 assets. It had an EDE, it had a agency, a brokerage agency, and it had the second largest lead generator in health insurance, Healthinsurance.org. We bought that company for a relatively small amount of money. It was even within my authority as the CEO of the company. And so the result was when we bought that, people said, "Well, what are you buying that for?? Some of our board members had the same question, "well, why do you need that?" Well, that EDE has become the rail for ICHRA for the whole industry. We created an EDE called ICHRA X. All of our competitors are on it. The fees are cost-plus unlike the $4 they're paying to other EDEs in the marketplace. It does both ACA and ICHRA. But more importantly, because they're all on the exchange, all on our ICHRA X. When an employee converts from defined benefit to defined contribution, they can pick any one of those networks as their product. It doesn't have to be Oscar. And this EDE became the genesis for Lucie. So we believe -- and what we did back in 2025, is while everybody believes that somehow the federal government will come to a conclusion that the audience enhanced subsidies, we created a plan where we knew they weren't. We advocated but we didn't see the highway to get it done. And what we created for the brokers was a vault that allowed them to take their members. We gave them all of their members that were affected by the enhanced subsidy going away and product alternatives for them to move those members into and the ability to make that action before open enrollment because brokers work on one simple principle. I want the highest level of earnings associated with a joule of energy. And so we did the work for them. And when others didn't, they went to those other carriers and moved those members to us. That was the secret of our growth. Because while they had all their other members ready by November 1, push the button, they all loaded into the system. Our line of growth went straight up. They were out taking other carriers business and moving it over to us. That was that EDE. That's why it mattered. And that same EDE today is working on program integrity efforts and DMI efforts for those same brokers with tools to be able to do that. So again, building capabilities ahead of meeting expectations of unmet needs in the marketplace. So we believe we have a clear path to achieving 18% national market share by 2029 with Oscar and better positioned to lead a market than our peers with differentiated state of capabilities and a single focus on the ACA, just in the ACA, and Janet will talk a lot about that when she gets up. But the most important part is that we believe these tools set the stage for ICHRA and build a platform for ICHRA to be able to make that market a seamless move. And by the way, a few weeks ago, as [indiscernible] got on stage and talked about the new choice plan that the administration wanted to develop. We have been working with them for a long time on that including the idea of having a separate wallet that has KYC capabilities that ensures eligibility instead of CMS having to build their own program integrity efforts. So we're committed to building new experiences in the health care consumer. This is another part. Our NPS score is now 71 today in an industry that averages 0. We are launching products every day about multicultural needs, clinical needs and lifestyle needs. And you can see sort of the TAM in 2026 associated with a number of members who think of these kind of products. And by the way, in these products, we have as high as 89 NPS. And so our retention is higher than the industry average. And so we have a less of a hole to fill every year when we're managing our growth. Reshaping the traditional insurance experiences requires delivering better experiences with less administrative burden, faster payment and more issues resolved on first pass. So we're doing a lot of programs around eliminating provider and member friction. And it doesn't have to have be perfect when we launch it. We're talking about the progressive elimination of failure. If we can impact more and more populations as we go in implementing our new tools, we're advantaging our customers every day instead of waiting for a big bang. So we don't have an announcement deadlines. We're implementing and again, Mario will talk about this, we're implementing AI every day. It's not a different department inside of Oscar. It's part of teams inside of our technology and business groups where we have a business person, a technologist. We have a product management person and somebody who understands AI, all working on improving each of our systems, which are owned by the owner of that business or of that process. So we're looking at reducing lag time from service to payment on 100% of office visits and lab tests. And we don't do that today, but the industry does it today for pharmacy. Why can't we do the same and deliver superior experience with 24/7 AI member support for 90% plus of member issues on first pass. And where we do have tools implemented that we're testing, we're seeing as high as 82% first pass resolution from members. So there's a lot of hope in the kind of things we're developing for the platform. That has been driving down our administrative costs that has created the kind of SG&A ratio that was only an idea in '27, but we have exceeded it already. So I talked a bit about AI. We have very fast AI value adoption. We have unified real-time data sets, modular scalable architecture, we have an AI-driven automation process, the way we ideate, test, implement. It's very important to the organization and how we roll this out. We've seen a 33% improvement in operating leverage from 2024 to 2026. And we've already seen $3-plus billion in total medical cost savings in the same time period. we are applying AI against medical costs. And we have a team that meets almost every day. We report on it every month on the scores that we put out for our team to be able to get the numbers below the trend that we put into our pricing. So expanding the individual market, unlocking the potential of the individual market, the tools that we built to make health care more important for all Americans. That's part-time workers, gig workers, 1,000 less employees, 1,000-plus employees in other government programs. I think Medicaid will happen before Medicare. But that's the next chapter of the ACA from our perspective. And it's part of what the administration offered back a few weeks ago called Choice. So I'll start using choice from here on out because ICRA is a terrible name. So here are the number of opportunities that we see in 2026, 0.5 million lives in Choice. We think it's 2.5 million lives by 2029. Small, midsize employers are leading that Choice growth market, but we see large employers also adopting it, sort of a barbell. So this is the opportunity. We think there is a tipping point here somewhere we don't know where it is. But when it happens, we're ready. It's the way we think about it. We can't tell you exactly when. We haven't put a lot in the numbers other than these and so from our point of view, when the tipping point happens, we're ready to handle the volume, we have the capital to do the work. Employers need innovation solutions for the changing workforce. 73% of American -- 73 million Americans identified as [indiscernible] independent workers. They changed their job on average 13 times that may represent some of you in the room, maybe not. And employers don't want to manage their own health care risk. 62% of large employers are exploring active or actively planning to shift to Choice. 91% of Choice adopters say it was the right move for the company. Now they're early adopters. Here's how we think about the network. And I talked a little bit earlier about the network being the largest PPO. 73% of active physicians participate in an ACA marketplace planned network. 57% of local PCPs are included in the average single employer network. So the difference is huge. It's a big opportunity. And there's greater purchasing power for the employee. And this is a little complicated math, so I'll try and describe it for you. Currently today, an employee pays 15% of their premium, and they pay all their out-of-pocket costs out of their own pocket. With the transition to defined contribution, the employee actually gets up to 26% savings of the employer amount they're given because they get to pick their plan because they get a narrow network, and narrow networks are cheaper than the networks employers offer. And that 26% can then be used to actually pay out-of-pocket costs instead of having to borrow for it. So it's a net gain for the employer in a number of ways. And these are some representative numbers. But it's a huge difference and it's a difficult thing to explain, but the wallet for employees get better when they have a defined contribution plan. So we have united our competition around a shared industry technology structure. Now we built that together. They're paying cost plus, which is less than $1 per member per month versus $4 for other platforms they could use. They get that as part of their joining our group, but they also then get the opportunity to lean in to Lucie where we have over 70 other carriers available for them to partner with. And I'll show you a list of those in a moment. So scaling Lucie's leading marketplace platform for carriers, brokers, employers and consumers unlocks this consumer power to be able to use -- to be able to use curated pricing to make better decisions about what they buy and have some people like hospitals stop selling things like certain meds or retail meds that could be bought cheaper or DME, durable medical equipment. So unbundling health care from funding is the big idea here. So I grew up in Detroit. If you want to do an auto dealership and said, "I've got $350 a month for a car. What can I buy?" You're going to get cheated. It's the way it works in Detroit. We love people like that. But if you go in and say, how much does this car cost, these are the issues, these are kinds of attributes I want? How much is it? And I'll tell you how I'm going to finance it later, that gets to a better purchase decision because you now have made a trade-off on value with your own money, giving consumers a health care wallet that does LSA, HSA, HRA and HSA all on 1 wallet, maybe even a credit card something to think about. Wouldn't it be great then I get to decide how I spend that money. And if my situation changes, the people funding my bucket of money changes, not my plan, not my network. So the opportunity to keep members for life by letting them keep their network and change their product as they age is the bigger idea. What is the value of the lifetime value of a member and their ability to keep their network and to change their plan as they grow older. Young and mortals, some of you in the room, young families, older families, empty nesters and then seniors like me, all of our plan difference -- all our plan requirements are different. Wouldn't it be great if not only my health insurance, but all the supplements I buy, all the things that I buy to keep myself healthy, the gyms I belong to, the coaches I have, we're all -- they are all available for me to purchase through a marketplace where I can trade off value. An agnostic marketplace where all of our competitors are working to get the best cost for everybody. So we're working with state and national policymakers to [indiscernible], the right regulations to make this wallet happen. And we think that's an essential component of delivering on Choice. Talk a little bit about reshaping the marketplace. Today, Lucie, a payer-agnostic health care platform is the premier place for brokers consumers offering ACA and supplemental plans. We will expand the platform to offer broader health and wellness products to employers and most consumers, and we'll have guided experiences through AI that empower consumers to take control of those health care decisions. We already have a couple of those, one on pharmacy and one on imaging where people can shop to find the best solution for them from a convenience and a quality standpoint. So here's the marketplace. It's all of them, 350 million Americans. Even people in Medicare and Medicaid can take -- they have access to this marketplace. Today, on the platform today, we have 70-plus carriers. You can see some of the names there. We have a broad supplemental product ecosystem. These are just a few. There are others on there, Aflac, Allstate, Cigna, Pivot Health, and we have consumer shopping and decision support. Those all exist in the marketplace today. Where we're going is to frictionless setup for consumers and brokers because we believe that brokers are the important second step for employees after their employer moves to fan contribution, helping with using benefit selection tools to get employees into the right plan does 3 things. For the employee, it allows me to get extra cash to pay my out-of-pocket costs, or extra cash to buy other products or services. For the employer, it stabilizes the defined contribution over time when people are in the right risk profile. So we'll have [indiscernible] personal AI guidance for broker support and new health products with broader consumer choice. You can imagine them whatever comes along. And we're talking to all sorts of people like Eli LillyDirect and others who want to get on to the platform and have direct relationship with consumers. We're talking to supermarkets who want to have a section for a healthy food and healthy eating. So it's an interesting new way. And on this system, everybody pays rents. I'm going to see. So the #1 marketplace we have by '29. We expect 2.6 million ACA in specialty products sold in 2029 and 28,000 active brokers on the platform. It will include retail peptides and GLPs and -- which today is 48% cash pay, lab and diagnostic services, consumer wearables, lifestyle and wellness goods and services. Lucie will turn choice for chaos into choice with clarity and transparent guidance. And we have a video here to show you next. [Presentation]

One of the pieces in there you missed was the -- you import your medical history in will help with this analysis in guided buying. Now the next slide is a dimensioning slide for you. And it's a dimensioning side because we don't know how quickly these marketplaces take off. We already have 800,000 policies for open enrollment coming through the supplemental carriers already for 2027. So we're not sure how quickly it will happen. But in the ACA supplemental enrollment, there's $30 billion in revenue opportunity. That's the total available market. In choice enablement services, moving employers from defined benefit to defined contribution is $125 billion of total available market, and that's an important part of the pie that we believe we need to be engaged in and involved in. And then Consumer Health and products was $500 billion for a total available market of $650 billion. Over on the right, we dimensioned for each 100,000 members the total available purse, $60 million for ACA supplements and enrollments, $60 million for Choice enablement services, $30 million for consumer health products per 100,000 members and the margins of 35% on the first 2 and 80% in consumer health products. That's the economic opportunity. Question is, when does it tip? How fast does it move? But the reason we made these investments in this marketplace is we believe this will actually be the more dominant part of our capital structure and our market cap in the future. This is where we really make a difference because when we unleash Americans on health care, with curated pricing and information like you've just seen, which is available, we can then turn around and we can reshape the cost of the underlying health care system and get better costs under control. We can chase out excess rents. So we're building the new consumer health economy. We're scaling the #1 consumer preferred individual market. We are today. We're unlocking a larger individual marketplace by looking at Choice and Lucie. We're creating a leading health marketplace for consumers, brokers and employers for the industry. And we believe that by 2029, we will continue to deliver on our 20% plus revenue CAGR, a 5% to 7% operating margin depending on how we invest in capital and price our products and a $4-plus minimum EPS by 2029. So with that, the man with the numbers, I'll turn it over to Scott.

Please welcome Scott Blackley, Chief Financial Officer, to the stage.

Good morning, everyone. I'm Scott Blackley. I'm Oscar CFO. It's a pleasure to get the opportunity to stand up here before you today and talk about some of the great results that this company has generated, most over the past and, more importantly, what we expect to be able to deliver going forward. So this is one of my favorite slides. I -- whenever I'm having a bad day, I pull this thing out and take a look at it because this is really the evidence of what this team has been able to deliver. Revenue increased 7x over this horizon since our IPO. During that period of significant growth, our medical loss ratio has just been dropping. We've done that through disciplined pricing and through affordability tactics. And in being able to do that, we've been able to offset trend and drive margin. Lots of people told us that you wouldn't be able to grow a business in health care and have dropping MLRs. I think these slides conclusively prove that we've been able to do both. And we've always believed that our technology would allow us to drive efficiency as we scale the business. AI is further powering that opportunity. We'll show you the clear evidence of that. But our expense -- our SG&A expense trends have been cut in half over this time horizon. I mean really a breathtaking amount of improvement in that ratio over the short time horizon. So we're incredibly proud of these results, and we think they position us well for what's ahead. So all the trends that I just talked about are culminating in a strong 2026 performance, where many of the KPIs that we are achieving are actually approaching or beating what we had expected to be doing in 2027. So today, as Mark talked about, we're improving our 2026 outlook by $100 million to a range of $600 million to $800 million of operating earnings, which is double our original guidance. And with 8 months of experience under our belt, strong underlying utilization trends. We're also improving our full year MLR guidance by 80 -- or excuse me, 50 basis points to 81% to 82%. And the remaining guidance that we have for the year remains unchanged. Let me give you some color on the trends that we've observed through August. First off, overall utilization through August is favorable to our plan. Secondly, on MLR seasonality, this year is tracking according to our plan. We anticipated a more pronounced step-up in MLR from Q1 to Q2. We saw that. That was driven by the mix of new members and lower SEP than what we experienced in prior years. Third, we've been closely tracking the progression of member cost shares given the change in our mix year-over-year. Importantly, what we are seeing on that metric is that members who are reaching their maximum out-of-pocket costs are in line with our expectations. And from here, we typically see a very consistent progression through the end of the year. Given where we're at, at this point, we feel like we've got good visibility that, that will continue to move in line with our expectations. And then finally, I want to give you an update on what's going on around market morbidity and specifically the CMS 1 million member program. So CMS has completed their industry-wide review of 1 million members in the ACA. We received our member termination list, and we have processed that from that program. All of what we saw through that termination process was consistent with our expectations. And we have processed all of those terminations that will be reflected in our second -- our third quarter results. We do expect that there's going to be a second smaller CMS program later this year. We think that will cover something in the range of approximately 500,000 ACA lives across the entire marketplace. The CMS reviews are really focused on eligibility, making sure that they are removing unauthorized enrollments based on things like missing social security numbers or other data matching issues. Just want to make it clear that we have fully reflected the impact of these programs in our -- the guidance that I just walked you through. And we feel like we've got good visibility into the remaining performance through the end of the year. So as we look to next year, we are positioned to increase our market share and drive top line growth. Our performance to date in 2026 sets a strong baseline for revenue growth and for margin growth. We're encouraged by the rational pricing environment that we're seeing in 2027. Our low teens rate increases are largely in line with national averages and should position us well to take and exceed on gaining market share. And turning to the market, what are we expecting for the market in 2027. We're projecting the overall ACA market in 2027 will be very stable. Specifically, we expect by the end of 2026 that the market will end up at around 17 million lives in the total ACA. And then we expect that the market will stay at around those levels through the end of 2027. So going back and talking about a little bit of the successes that we've had and last time I stood up here, we talked about targets for 2027. I'm here to give you an update of how we're doing against that -- those targets. I told you last time that our 2027 targets were ambitious but achievable, and I'm very pleased to say that our strong performance to date and our current view of next year suggests that we are on track to broadly exceed those targets. So specifically, we anticipate that our revenue CAGR will be between [ '24 and '27 ] will exceed 20%. We now expect that our 2027 operating margin will be greater than 5% next year. And we expect that our 2027 EPS will be greater than $2.25 a share. Just as you're doing your models in math, we expect that our 2027 effective tax rate will continue to be in the mid-single digits. So all of this is being driven by our differentiated strategy and our market focus. And that is cumulatively producing these strong results. I look forward to giving you specific guidance on 2027 at our investor call next February. All right. So let's talk about what '29 is going to look like. As we look out over the next 3 years, we believe our earnings will be materially higher than where they are today. First, we're targeting revenue CAGR of more than 20% through 2029, driven by a growing individual market, expanding our footprint and increasing our market share. We expect that revenue growth as well as an improved MLR and SG&A ratio will help us to achieve a 5% to 7% operating margin by 2029. And then finally, we see a clear path to achieving EPS of greater than $4 per share by 2029. This assumes that by 2029, we will have a tax rate that is at that time in the mid-20s and that our diluted share count will grow at around 2% to 3% per year from today's levels. Let me go through and outline a little bit of how we're going to achieve these targets in more detail, and I'm going to start with revenue. So with revenue, we're obviously keenly focused on growing the top line by more than 20%. We certainly [indiscernible] ship at the house of scale and know that the larger we are, the more we can leverage our AI and technology innovations to drive efficiency in our business. Our underlying revenue assumptions -- underlying our revenue assumptions is an assumption that the ACA market grows from 17 million lives at the end of 2026 to something around 23 million lives by the end of 2029. During that time, we'll be expanding our market share and increasing our footprint in our existing regions. And those are our largest growth opportunities. We estimate that those growth opportunities will drive between 17% and 19% growth per year. And on average, today, we have a 30% in market share in these regions. And so there's plenty of opportunity for us to continue to increase in our footprint. We also have a large expansion opportunity in front of us in terms of both entering into new counties as well as entering into new states. We typically enter into new markets, with a 5% to 7% market share, and then we grow from there, which translates to around 2% to 4% revenue growth through 2029. And then finally, we anticipate that new Choice and Marketplace products will create revenue growth vectors that can compound over time. And while we've not specifically called out this in this waterfall, I'll note that cash and investments now generate a significant investment income that's included in our overall revenue. We expect that NII will continue to increase through 2029 at a rate of around 10% to 15% per year. Let me turn to margin. So we've made really meaningful progress on our path to achieving our 5% margin target for 2027. From today's level, we have a clear path of getting to up to 7% operating margin by 2029. We will continue to maintain our disciplined approach to pricing and we're going to price to cover rising cost trends. We believe that the inflationary environment that we're currently in is going to remain over the next several years. And so that 5% to 7% trend is higher than what we've seen historically in the 3% to 5% range. We think that, that will continue over the forecast period. And while pricing will play a part in driving margin, medical cost management programs will be a key to driving and increasing our margins. A big part of what drives our margin improvement is the continuing leverage and scale from efficiencies in our tech platform and from an AI-enabled process improvements. For example, AI has helped us to automate reviews on things like prior auths. It's improved our claims processing time and it's increased the accuracy of our risk adjustment submissions. All of these things are collectively small parts of the business, but when we aggregate them and track them and make sure that we deliver on the commitments, what we're able to do is to offset trend and drive margin year in, year out. And so I'll go through the next few slides with a few more details on these topics. So as you saw in my first slide, we've made significant improvements in our MLR over the past several years really. And we're beginning to approach our target of 80%. The key levers to improving our MLR include disciplined pricing. You'll hear that from us over and over again, but also through medical cost management and doing things like perfecting our network performance, improving the core operations and creating clinical innovations. Technology obviously plays an incredibly important role about -- in that process. You'll hear from both Janet and Mario some further examples of how AI is today reshaping the cost curve for us. particularly in areas like claims automation and fraud, waste and abuse. And the benefit of all these efforts is that as we see rising cost trends, if we can drive down costs through these affordability initiatives, we can really pass that on to our members and make sure that the product that we're selling is affordable for them, too. So we've made significant progress on SG&A. We're not done there. I think we've got a long ways yet to go. Many of the factors that have driven our historical success will continue to do so in the future. First, we've seen significant operating leverage that compounds as we scale. In fact, our fixed cost base has gone from 30% back in 2024 to around 25% this year. Our technology and AI have already enabled savings in our cost structure that's been allowing us to drive down variable costs, improve our operational workflows, and we expect that this trend is going to continue. And I would just note that a lot of these enhancements and improvements are serving us well by driving down costs, but we're also being able to create better member experiences. So this isn't just an effort to remove costs out of our system. It's an effort to improve the experience for our members and realize value for ourselves as well. So let me pull up on Lucie. Mark talked a lot about the opportunity that's in front of us. The first thing I want to point out is that are greater than $4 of EPS in 2029. That target assumes a modest contribution from Lucie. We're not relying on that new business to hit our targets. Secondly, the economic opportunity that we see for Lucie is very significant. So right now, today, Lucie is allowing us to capture new fee-based value pools, 70-or-so carriers that are currently able to sell ACA and supplemental products through the Lucie marketplace. And as Mark talked about, if we're able to capture 100% of the potential value across Lucie's platforms, for every 100,000 people that transact in each of these 3 buckets. In total, that's a $150 million revenue opportunity per 100,000 members. So that's the size of the prize that we're working against. Now we don't expect that we're going to get 100% of all of the fees that are running through this entire marketplace. But the magnitude of that spend is so significant that if we're successful in getting a portion of that, we think this could be a significant business for us and one that's very meaningful to our results. I just mentioned that this business has several desirable financial traits. Number one, it's a service business, so it's very capital efficient. Number two, it has at-scale margins that are 5 to 6x greater than our core insurance business. And then lastly, this is a technology business. So we think that we can power it with AI and that we'll be able to scale it very efficiently. So let me turn to the balance sheet and our cash profile. So today, we're certainly operating from a place of strength. We've got $460 million of parent cash and more than $10 billion of cash investments in total as of the end of the second quarter. Our insurance companies are well capitalized. We've got almost $1 billion of excess capital at this time. The plan that we've outlined today is expected to generate significant amounts of capital in the range of $4 billion to $4.5 billion across the parent and our insurance subsidiaries. And so what are we going to do with all of that capital? Well, our capital priorities start with organic growth, reinvesting in the business. These are the highest returns that we can generate on the capital investments that we can make. We'll also continue to optimize our quota share reinsurance programs. These programs are a cost-effective way for us to manage our entity level capital requirements. We expect to continue to use them. We also expect that over time, we will use them to a lesser extent. We would anticipate that around 60% of our capital generation will stay within our insurance companies; we would seek to have that sent to the parent as dividends, and that capital that comes up to the parent would allow us to do things like pursue opportunistic M&A and manage share dilution. So I'll leave you with a few takeaways. Number one, we have already demonstrated that this company has durable earnings power and that we can grow revenue by more than 20% a year. Number two, we're well on the way to realizing that 5% target margin that we set for 2027, up to 7% by 2029 and generating $2.25 of EPS or greater next year and that we have an achievable path to getting to $4 of EPS by 2029 or greater. So overall, this plan is going to generate strong cash flows. We think that we have plenty of opportunities to put that capital generation to work to improve shareholder returns and improve the future performance of this company. And the targets that I just described to you are based on the business really as it stands today, and we can see significant incremental opportunity beyond the scope of these targets if this Lucie marketplace picks up and gain speed over time. And so with that, I think we're prepared to take a 15-minute break at which point, we'll welcome you back to the meeting. Thank you.

We'll now take a short break. Please be back in your seats by 10:25. [Break]

Please welcome Janet Liang, President of Oscar Insurance to the stage.

Welcome back, everyone. Now Mark started off this morning, kicking us off talking about Oscar's vision for the consumer marketplace and how it's a model for our country's health insurance system. And then Scott followed up with our strong financial performance and positive outlook going forward. I'm really excited to share with you the expertise and the execution model that is led by our teams at Oscar to deliver the results that you see today. I'm Janet Liang, President for Oscar Insurance. Now the ACA is the fastest-growing segment in the health care system in our country. And we are seniorly focused, right, on leading and growing in the consumer marketplace. Oscar has significant runway to expand our reach and set the pace for innovation in 3 really important distinct areas: products, network and technology. And these 3 assets are specifically designed for a marketplace where individuals can choose the plan that they want to belong to. So I'm going to start back at our Investor Day in 2024, where we have, and I can show you that we have delivered on the commitments that we made to you 2 years ago. So we have a track record of profitable growth that reflects superior disciplined execution. We have nearly doubled our membership. We have doubled our in-market share and doubled revenue in 2 years, right? And this is all in a context, I want to take you back, you said it all in a context where there was tremendous uncertainty in the marketplace, right? We had the expiration of the enhanced premium tax credits. CMS was issuing new guidance for payment integrity and eligibility, right? And we also saw a rise in market morbidity that became -- everyone became aware of towards the latter half of 2025. And the entire industry, all carriers were impacted by this, right? And so in a year of great uncertainty, we had a choice to make and we chose to meet that moment. And let me tell you how we did that, right? So we deeply understand this marketplace. So individuals were losing their subsidies and/or seeing significant increases in their cost share. And they needed options. Otherwise, they were going to have to drop from health insurance completely. So we took a step back, and we designed affordable bronze and gold plans that they could move to, plans that were priced with discipline, that generated margin and helped us grow. The second thing that happened was because of all this uncertainty, carriers literally exited the marketplace and/or raised their prices so that they could take a step back because they weren't sure what was happening in this marketplace and instead we laid in, right? And brokers at this time were very confused because they had large books of business that they had to move. So what we did was we said, "Okay, we're going to reach out to these brokers." Some of them who had never signed an Oscar Life before, right? And we said, "Look, you're looking for a stable carrier partner dedicated to the ACA, then that's Oscar, right?" And because of our early outreach and early education we grew our distribution by 60%, right, unheard of 60% growth in our distribution of people who wanted to now sell Oscar Insurance. So this strategy gave us a first mover advantage during a major market reset, right, resulting in year-over-year 60% membership growth, 60% revenue growth and 7 points of margin expansion all in 1 year, right? Today, like we are a stronger organization because of it. We now have the scale to operate with much greater efficiency and have much greater influence in the markets in which we operate. Combined with this pricing discipline, we have shown you that we can expand margin and that we can grow membership at the same time. And I'm here to tell you, we are going to continue to do that going forward, okay? So we're really, really excited. As we look ahead, we see a 65% growth in our addressable market, with Oscar reaching national share by 2029 of 18%. Now Mark walked you through the tailwinds that will grow the individual market to 23 million lives, right? And that's the choice adoption through a growing gig workforce and changes with AI driving labor shifts in our economy. But within that, our TAM sits at 9.6 today and is moving to 16 million. This reflects our footprint for Oscar, right, within that national 23 million. So by 2029, we'll have a TAM with 16 million. And there's 3 important ways that we're going to get there, right, three: So number one, we are -- we have an existing footprint and in that footprint, we are going to mature our market share from 30% to 35%. And that's going to bring 2 million -- add 2 million of addressable lives to our market for us. Second, we're going to expand into an additional number of counties, roughly 400 to 600 new counties. We actually expanded 150 counties last year. So we're going to double that each year in terms of a growth rate and expansion, and that will add 3 million lives. And then lastly, we are accelerating our choice options and our choice strategy, right, to add another 1.3 million lives in our footprint. And I'm going to double-click a little bit on Choice because Choice is a new growth avenue for Oscar and new -- some of you in this industry. So this year, I just want to share with you that we are already seeing 2x increase in membership from last year. And also this year, we've been busy. We refreshed 250 products for off-exchange where employers and employees can find their choice options, right? So that means that by 2029, right, we anticipate that we will be at 3x market share in choice. And our focus is going to be in high priority markets. So we expect to get to 20% share in these high-priority markets. So what's a high priority market. A high priority market has attributes, right, economic attributes. The first is that the state has passed favorable legislation in terms of tax credits for employers who are moving to choice. So tax credit, immediate savings. The second is that there is a very obvious difference between the average premium on the exchange and the average premium an employer may pay. That represents a substantial discount when you move from your current insurance into the marketplace. So there's an attractive savings for employers. And lastly, we look for markets where there is a good mix of employers from small local employers, but also large employers with multistate footprints that have a diverse workforce from part-time to hourly workers, to full-time exempt office workers. When you have that sort of mix in your population, you can quickly take advantage of choice to take a slice or category of your employee base and move them into choice of having to move your whole group to give you some early experience. And so there's early adopters willing to try with certain groups of their employee workforce. So together, right, this offers us an opportunity of a new TAM in our future. Choice is growing today. It's -- I know we talk about it as a future opportunity, but it is growing today. As we talk to our employer consultants, we talk to the chambers, we're meeting with HR employee councils, we are hearing that over 62% of large employers today are already exploring a shift. Our quotes for Choice options has doubled from last year. There is more quoting activity and more employers interested now that they're [indiscernible] of where the real benefits to move for both their employees and employers, right? Our value prop delivers the choice that they expect with significantly more products at affordable fixed prices for themselves and for their employees. So I just went over 2 things: One, right, our execution generates profitability; and second, we are growing our TAM, right, in market, expanding and through choice. Now let's talk about 3 really important assets that we have that are unique at Oscar. And I believe that these 3 assets put us on track to become the #1 consumer-preferred individual market carrier and those 3 assets, and you hear them from a lot of carriers, but you're going to hear why it's different at Oscar. That's our networks, our products and our technology. very different approach when you're designing for individuals in a consumer marketplace. So let's start with our networks. So we were fortunate that Oscar was forming at the time the ACA was forming, right? So we had the opportunity to build our networks from the ground up, understanding the ACA right, and making sure that we were addressing the needs of the people in the ACA. Unlike other carrier competitors who had already had very large wide commercial networks for all their different lines of business, and they had to retrofit those networks for ACA, right? So the ACA gives us a very broad supply with 73% of providers in the country participating in the ACA offering ample choice for any individual who needs their care needs met. So when you overlay choice, this level of choice with Oscar's advantages, right, it helps us to cultivate the kinds of partners that we're looking for in terms of providers that are going to work with us for the needs of the individuals and the ACA. Now together with our partners, our technology, our expertise, our understanding of this consumer segment, we were able to deliver $3 billion in affordability during the last 2 years, right? $3 billion in affordability that goes directly to margin expansion and competitive pricing. And we do -- when I talk about a high-performing network, this is what you're going to see, right? So you're going to see that we have improved the financial accuracy of our claims payments and timeliness, and we've reduced friction for providers by simplifying prior authorizations and really focusing on anticipating care needs with patients for their patients, our members and to take the guesswork out of what's going to be covered, right? That's what you see in a high-performing network that drives affordability. The second thing you're going to see is that we have care teams and care guides and our Oscar Medical Group, providing navigation, helping you find the doctors in your network, helping you to get care now. We are there to help you stay healthy and to make sure you're getting to the right care at the right time in the right setting. And lastly, right? What we depend on in terms of driving affordability is going to be our technology and our AI insights, our data platform. We have AI-driven insights that help to prompt whether it's an automated prompt or whether it's an AI agent prompt, whether it's a human being calling to prompt, like we help our members like understand your benefits, use your benefits, ways to save money, right? And also, again, we want to make sure coverage is predictable. So we are really focused on how do we integrate the care that you need with the coverage that you get and make that a seamless experience. So these high-performing networks, here's what the great news is, that these networks, right, have produced this level of affordability. And these networks we built for the ACA, it's the same chassis that's built for our Choice products off exchange. So what does that mean? This means that this design strips 20% of the waste that comes from large generic employer plans and employers and employees can stop paying for breadth that nobody uses, right, and can use these dollars to customize benefits and plan options, right? So that leads me to our second asset, our products. right. So now we have these networks that are core, right, as a foundation for our products that allows us to put money into product design. So we customize our products for high-growth segments that we want to attract through a new category called lifestyle products. And I'm going to walk you through what that means. So many of us are offered. We think of products as I have a PPO, I have an HMO. I have an HSA with my high deductible self-funded plan, right? And if you're in the ACA, it looks like I get a bronze, a gold, silver choice, maybe platinum, but not usually, right? And so we're here to say that we really understand the individual customer. And I want to tell you that individuals want their plans to resonate with what their needs are at the time in their lives, right? And our lifestyle products, you can see this, have a Net Promoter Score of 71 versus an industry average of 12, an industry average of 12. Guess what? People don't really love their health insurance, but I'll tell you they love our lifestyle products. And members saved nearly $1,000 on average each year on out-of-pocket costs. If they choose one of these plans that's tailored for them versus if they bought a generic plan. And I'm going to give you 3 examples of products that do just that. So what we're really good at is zeroing in on what are the health patterns that people experience, what are preferences that different cohorts have, right? And what are buying behaviors of consumers for each of the cohorts that we serve. And then we build benefit designs, care navigation and rewards right, around those distinct needs. Now we have targeted 4 high-growth segments for our product pipeline, 4 high-growth segments. And last year, 3.4 million new consumers came into the ACA. So when people say the ACA is not growing, not true. 3.4 million new consumers came into the ACA. And many of those new consumers are individuals that fall in one or more of these categories. So the ACA is growing with consumers that we want to serve, right, a great match for us. So what we do is we take this knowledge of how we want to grow, right, where are segments that make sense for us and who's coming into the market, and we build solutions around that. And so I want to talk to you about 3 of them. And you know our Buena Salud, those of you who have been with us, our Spanish first experience and product and also our condition-specific products, one of which is diabetes. But three new products for you, just to give you an example of how we have really risen the bar on how we think about consumers and how people buy health insurance. Number one is HelloMeno. So HelloMeno is our very first life stage product, right? And this is for women navigating menopause. So we take our core medical benefits, our core medical coverage. And we put on top of that for $0 out-of-pocket, let me say, 0 out of pocket. Any care that you need that's related to accessing a menopause clinician, hormone replacement therapy, bone density scans in some new medications and anything else that's common and related to the treatment of menopause, right? 0 out of pocket, we take the barrier of money away so that you can get the care that you need. If you're somebody who says, "I have menopause," and that's my primary focus right now. Now because of the stigma associated with talking about menopause for women and for their broker distribution channel, we use social media and influencers to reach women directly. And guess what? We saw a 2.5x increase in direct enrollment versus our traditional plans. That means unaided by a broker, right? So the good news is we understood there was an unmet need, and we met that need. And we were rewarded by it, right? We win for everyone. The second product I want to talk about is Hy-Vee Health. This is an example of our signature choice product designed for the employee marketplace. Hy-Vee Groceries, if you don't know who they are, they are a very large kind of grocery, regional grocery leader in the Midwest with over 40,000 employees and over 500 stores. And they happen to have a very strong health division where they own their own primary care clinics, they have pharmacies, they have dietitians and a lot of different clinical services for their customers. We partnered with them and we work with them so that we created a medical product where essentially 90% of the care that you need, which is primary care, right? 90% of the care needs that you have. If you go -- if you use their concierge primary care facility, it's 100% 0 out of pocket. Again, seeing your doctor, basic labs, basic prescriptions, right, any time, unlimited access to your primary care office, 0 out of pocket. This combined with, sort of, what they're really good at, which is discounts and groceries, access to their food as medicine programs as well as just -- they have local dietitians, so they're all available to the members in this product, right? Great example of a partnership that expands our reach and also a great brand for Oscar. The last product that I want to share with you, is super excited, and this is our focus on creating, again, affordability and options for people that are buying health care that are buying are the health care consumer. So you'll see this product in its filed and it's ready to go for sale in 2027. And we have partnered with Allstate Insurance to create this new bundled products or product bundles, right? So it takes Oscar medical insurance plans, combined with Allstate's supplemental cash benefit plans. And together, they really help individuals, mix and match or how they want to take their risk and their cost shares and their deductibles. So greater flexibility, greater coverage depending on how you know you're going to utilize health care. And this new partnership gives Oscar access to their distribution network of 38,000 brokers, right, who have not -- some have heard about 20% overlap with our distribution. So about 30,000 new brokers who have not written Oscar medical plans previously, right? So this is a big win for us in terms of both offering affordability to our members in the marketplace, but also to expand our distribution reach. Now all of this is not possible without technology, sort of, the third asset that I talked about earlier, right? So technology, it's in our DNA. It's in our founding DNA, it's who we are. Oscar was formed because we believe that technology and an integrated tech stack could reduce friction in the hassles on the American health insurance system, right? And we are doing just that. And it makes sense that we're an early adopter of AI, right? Because we are a technology organization. So at Oscar, technology really underpins every business transaction right? It drives our intelligence platform. And now with generative AI, right, we're enabling automated solutions and workflow at scale. We are essentially operationalizing AI. So I want to give you an example. We have and we are building a very dynamic member profile. And we -- because we have our own tech stack, we have all of the information, real time from providers on all the activities that they're driving for a member, all the utilization, right? We also have our 24-hour -- 24/7 AI agent, Oswell, where members are querying from symptoms to coverage for how to find help and resources. We have all of that information as well. We track your prescriptions. We know what benefits you've used. We know the plan you selected, so we know what you're interested in, right? And we take all that information, it's a living profile that's updated instantly. And that profile and that data helps us to develop prompts. Prompts for you to -- you should go get care now, right? We need to help you navigate to a lower-cost setting to help you save money, right? We're going to help you maximize your benefits because we noticed you're not using some of the out-of-pocket 0 benefits like your annual wellness visit, right? If you're a diabetic, you're 0 insulin program, right? And of course, Above all else, we want you to stay healthy, right? So we are hyper personalizing experience at scale. And in addition to that, we take that feedback, right? And we use that to update and to design our future networks, our products and our tech, right, capabilities. So I am incredibly bullish on the business value that gets generated with an integrated tech stack and the model that we have at Oscar and Mario is going to talk about more of that with you shortly. But I want to say from the business side, the way that we work together is unlike anything I've ever seen in my 30-plus year career in health insurance and in care. It is truly phenomenally in advantage. So together, we have -- remember what I said, network, product and tech advantage. So in summary, I just want to share with you that we are on a clear path to deliver profitable growth with 18% national share by 2029. And the 3 key drivers. We have an addressable market that is growing by 65% to $16 million by 2029. So contrary to what you may hear from other industry analysts, we are growing and the market is growing. Number two, networks, our networks are on pace to deliver continued affordability, right, gains. And we are going to continue to provide exceptional provider experiences at the same time. And lastly, we are going to win new customers because of our relatable meaningful products that deliver real value. Our expertise and commitment to the consumer choice is unmatched in this industry. And we are on the path to become the #1 carrier in the ACA by 2029. Thank you. And Mario will now take us through our technology platform.

Please welcome Mario Schlosser, Co-Founder and Adviser to the CEO, to the stage.

Thanks, Janet. Excellent to be here. I am Mario Schlosser, the co-founder of Oscar and adviser to Mark on technology matters and other matters. I had -- I was sitting here in the front, I had to take a picture of Scott standing in front of his now-favorite charts. The improvements of Oscar performance since 2021. And it really is astounding to reflect on how far the company has come, how well I think we've done the work of becoming just a top-notch insurance company and health care company. And that is really -- was really the original founding vision for the company that's building a unified technology platform will enable us to deliver higher margins at scale. It's exactly, I think, what we've seen here. Of course, it's a team spirit and team efforts to make that happen. But that is something we have been able to show in the last couple of years here. Before I even jump into this, I have -- I just realized this morning, I've been writing code for almost 40 years at this point, which is a crazy number to say. And I have never seen anything like we've seen in the last 10 months now. The way in which agentic coding has entered the bloodstream of companies or should be entering the bloodstream of the companies is totally insane, 1 thing I'm going to try to do here is to give you a real visceral idea of how we're using leveraging that at Oscar and how you have to have a very different mentality as a company, different technology [indiscernible] as a company to even make use of all the stuff that's happening on the AI side. Okay, why are we able to make use of what's happening in the AI side? Well, we have 1 platform, quite simple, claims service, clinical workflows under experience. It's all the same cohorts, the same foundation, the same mono repo. That is a big reason why in the last couple of years, we have often been the first, I believe, to launch certain AI tooling with the first launch, what we like to call, super agents, benefits chatbots. We're the first to launch as well. Janet just mentioned him or her, actually don't exactly know how the [indiscernible] I should pick that one out. Her, let's just call it, which is the clinical chatbots. And that just works because we have this foundation, we can leverage over and over again. Within Oscar, the health plan, that will continue to lead to more operations efficiency, network efficiency, things like that, better member experience, better navigation, lots of ways the insurance folks can make use of that. Outside of Oscar on the Lucie side, that we're going to apply these techniques to better shopping, better comparison of information across the industry and things like that. This is bit of a grounding charts I'd throw on here. What does it mean? What do we look at when it comes to the peer group of how we benchmark our technology and AI efforts versus others. The peer group isn't other insurance companies, you might not be surprised here, that's the peer really ought to be for all of us for enterprises nowadays, AI native companies, companies like the Foundation Labs, for example, that really, at this point, say my engineers don't write code anymore. It's all agent-written, things like that. And to really live that kind of AI application, you have to have a different environment. And the environment is shown here. You're going to build codes and build agents and so on one foundation, really on repo and things like that. You put that all inside of expert design guardrails. If you have -- if you deploy your agents in that way, you can then just watch them work and watch and produce results. You get feedback from real outcomes as quickly as you can, and you feed that then back right into building something better. He is a very nice example for how we've been applying that theory in a sense in practice in the last year or so. When we launched Oswell last year, midway till last year, I think we had the first alpha members in using it, the clinical chatbots. It's a multi-agent architecture. So there's a pharmacy agent. There is a medical record summarization agent and so on. These have all been laboriously fine-tuned and get the instruction written and things like that and then test it over and over and over again. That architecture are the points where we know so much about what goods delivery of experience looks like that we have so-called evals all over the place, that let the agents modify their own prompting. So there's almost a living system now that improves on itself as it goes along. Of course, again, within expert design guardrails. So clinicians are there to watch this. Our operator out there, they watch this. But that's really what he wants. We went as an industry from agents that are in some workflow a year ago to agents that can write codes maybe even themselves. Now to increasingly agent them improve themselves. And we got to ride that ladder, claim that ladder up further and further. If you don't have that foundation, if you don't have that mindset, AI is just going to be another point solution for you. You can hire a bunch of vendors. They will stitch you something together. It's not all that exciting, I think. That's when you're eating out some gains, perhaps a lot much more. If you have this mindset foundation, that's how it becomes a part of how the business operates really. And so yes, this is how we build Oscar from the very beginning. We like to call ourselves the full stack health plan on the top right, then a little bit of help and down here. You got products, creation, network, billing, claims, care, all in that one platform that gives you a lot more insights into what's actually happening on the platform. Pretty much, I would probably say every operational metric stream now has some agent looking at it and I'll have a slide later on where you see what that means. We have a pharmacy agent that just watches pharmacy claims and can see spikes that we should be looking at. And with some humans involved as well. health expertise in coded, I think, is the next frontier here, and I will have some videos for you by the way later on as well. One of those videos, if you hold that thought for a second, will be about one of our claims operators, not one of our engineers, one of our claims operators who rewrote the software tool he is using to take claims. And that's really the promise of AI, I think agentic coding that you can have non-software engineers build software very, very quickly. And then four, of course, you have to put that in this closed loop and keep improving things. It now took us a couple of weeks more recently to replace a vendor we used to have. That's a so-called itemized bill review for us. So bill comes in as a rule basically gets supplied, [indiscernible] complex rules. You want to pull that in-house. And really in a couple of weeks, we can do something like this now. Benefit is, of course, we save some money on the vendor cost, but we also get much more information in real time as to which rules fire how and what does that mean for providers, what we have to manage going forward and things like that. So that's the kind of platform you would want that we have been investing in from the very, very beginning. That's what's exciting about the company. A little bit about, again, software developments here with AI. AI is changing the calculus of how you think about tackling which projects. They all these I always like to say Oscar was ever bottled by ideas, ideas were always flying around more than we could handle really. Maybe that was in the early years, a little bit the problem as well. So we focus on also having good numbers. And -- but it was always in the execution. You had to actually somehow figure out what to do next in -- in agentic coding, one thing you can certainly now do is you can check much more quickly if an idea is a good idea. You can build a prototype more quickly. You can get it out there, get it connected. Almost any 2 will we now build, we'll have several clickable prototypes that get put together very, very quickly and leads to better design from the very beginning there. On the engineering side alone, and this, by the way, also shows the insane speeds that the world is changing at right now. At the beginning of the year, only about 16% of the codes that was shipped in the production systems was written by AI agents. And now that number is almost up to 2/3s of all the lines of code written by coding agents. That is astounding to me, as I said at the beginning, I don't think I would have thought that we would just, sort of, hand off this craft, we've been honed over decades as humans to AI agents and be better off for it. And so that gives us also about a 2 to 3x faster time to market. Health care is not known. I think for its time to market ensures even less known for time to markets. So many more of these plans on a [indiscernible], so many more of these ideas around the actuaries have around which we look at, how should we get the costs here now become possible because this now really works quite well. This is a bit of a service slide, I thought. We always like to -- in every earnings call, ensure is now I can talk about the AI use cases they have going left and right, get through that a little bit. And what I always listen to is, do they also talk about things that they try that didn't work out. That's often a good sign as to whether what you say works actually did work as well. There are plenty of things that don't work. And so we're going to have to continue to be on that journey of figuring that out and learning that for ourselves. Just a couple of examples. Provider data is an old pain of insurance companies, getting that right, right phone number, right specialty, right provider for the right time. The models are good at if we have several sources that sort of agree that need some nuance in the interpretation in giving you that nuance and interpreting that. The models can do quite well. And that's 1 reason why we've been improving our provider data accuse in just the last 12 months as well. They're not very good at sparse data and making that up. There's still a lot in health care, which I think is a bit of a metaphor business for where you have to go into a practice into a hospital and figure something out in the physical space in the real world. So that's the model soon to get better. Fraud [indiscernible] abuse is an area where we benefit a lot from very systematic, very deterministic application of rules again, not really what the models are that good at. And so when we experiment with fraud rate and abuse, often we see -- get build a model that is -- or build a system that is rules-based, not one that is, sort of, -- don't have the agents try to apply interpretation in the moment in time. And the final one, incredibly important as well, a nice example here with Lucy. We've been experimenting there with how do we get best gets people the best information, recommendation and when we test this with our brokers, we have so many brokers that love working with Oscar. We use them as a sounding board oftentimes. Then we realize they often like simple, clear search drop-down fields more than, sort of, Blackbox.ai recommendation because that makes them themselves feel like they can stand behind this. And it just really matters for the transit of trust that they then deploy towards the prospective members as well. Now that is going to keep coming back. How you build AI tools people actually like using enormous craft. I think we're ahead of the curve there in healthcare, certainly health insurance. Obviously, those are not failures, those are really good learnings as long as you can work them back in into what you do next. Listen to the user and then scale reliably when it makes sense. Okay. We're at the first video here. One thing I would say ahead of time is this will run fairly fast, fairly quickly here. We're going to share these videos out separately here. I hope I'm not whole promising here. I'll tweet them out. And after the talk and what we -- can you be pause for one second actually? I start talking anyway. So Teddy, one of our guys who's been with Oscar for 12 years, I think, at this point, started on the phone, then went to the claims team, runs one of our claims SKU. He runs the provider disputes claims SKU. So obvious provider experience matters a ton for Oscar. He builds at now for himself started over a weekend saying, "I have to look at too many different systems here when I challenged -- when I look at disputes and providers. I would have build a new one." What you saw happening here just now is same popping into our coding agents, saying what kind of tool he wants. That agent then goes to and build all this. This is now the real tool here. We really reenacted this, okay? This is not made up. And in that you see the tool pull together all these various documents that go into the resolution of the claims disputes and nicely put this together here. Again, quite fast. We're going to show it to you afterwards again. Of course, when you build something with the ground up, you can build data visibility right into it. And so for free, so to speak, you now get all this dashboard stuff here that tells you about tells Teddy about how his queue is now working. Now he pops back into the coding agents. Again, coding agent here, going back and forth saying, I'd like this difference, I like that difference. And what he's building now here is even more mindblowing. It's an AI within the AI. So the coding agent build tool for Teddy. And now he built himself bots that looks at all the same data, where he can now talk to that bots or his folks on his own claims SKU team can now talk to the spots and the bot starts interpreting data differently here. Again, an AI agent builds within a larger AI coding agents. All this stuff is only possible because we have an environment that's production, of course, with our production data and in the same environment with lots of synthetic data behaves the same way and Teddy can go against this synthetic environment and test of this as if it was the real world. And so you can see him nicely go through here, pull together all kinds of different PDFs and different instruction manuals that he would have had otherwise or his folks, we look at many and quite manually here, and we did speed up the coding agent working here. You saw that text come pretty quickly here, but that was him reenacting how this works. And this was the real tool that then came out of it. Now it took the engineering team another probably 2 months or so to take that to that he builds over the course of starting on a weekend and another 2 weeks or so. So there was a lot of kind of connecting of the dots and things like that, but that's just enormously much faster than it would have been otherwise. This would have been, I would estimate a 6-month project with 10 engineers or so in the past and product managers around it and everything else. So no 1 will speed up in actually building this. And on the other hand, this was literally one of the people who's adjudicating these disputes every day doing it. You cannot get any better at getting this expertise from the front lines. into tooling, then if you have it work in this kind of way. I think that is the absolute future, both of Oscar and of industry overall. And you got to have that's kind of set up to really be able to deliver that. A few more examples. This is now starting in operations. I mentioned before, we got all these operational queues we can look at. One of them is pharmacy claims, and we just have an agent look at these claims every single day and say, is there something here in some configuration that look strange. There's an example here that popped up a few weeks ago where in just 1 thing that looks strange that the investigation team looked at as well. I was able to confirm about EUR 28 million or so in drug costs that we probably shouldn't be paying for. So that came directly out of this real-time enumeration here. Prior authorization, a great example, of course, as well. Then hundreds of CPT codes in the Oscar prior auth base where we do completely automated approvals. We don't do automated disapprovals ever, right? That is always done by human beings, but ultimately it will pole better for everybody. provider gets it more quickly. Our team is less frustrated. And so some of the highest volume CPTs we have, high 90s at approval right now because AI can collate medical records and things like that. So quite a bit more in terms of leverage we can get there in some earn. This is my experience. And you can look a little bit on the right side there as well. This is a bit of a video of as well, and we have another video coming in just a second, so we don't have to parse us through too much here. Oswell now solves about 30% to 40% of all memory messages in receives right of the bat and also we very purposefully clinical and nonclinical. And an AI analyzes 100% of every interaction we have. So last year, we had some turns out that we -- people were calling us about changing PCPs because it was confusing on the ID cards, that stuff. You've got to realize very, very quickly. And it's small enough for people do not realize they just look at dashboards, but it's big enough to really matter to folks. And so it can help with this. This change in the role of care guides, right? What's the role of the human and all of this? Well, care guides have a different job. Their job becomes more about empathy, about some handling very complex judgments required questions, advocacy matters more things like that. But we can automate so much routine work when the answer is clear and bringing the right person when the decision requires judgments. Clinical finally -- and then I have one more video for you here as well. On the clinical side, 1 thing we do quite a bit, and we have a large team of nurses doing it is we reach out to members proactively about issues they might currently have a might face before they go to a hospital for surgery when they come out of a hospital, we don't feel like they have a good PCP, they can, sort of, catch them when they come out. It takes a while to do the research-only member for a nurse, about 20 minutes. You have to look at medical records. You have to look at discharge documents, things like that. That's a great AI application. It was on the right side here. You see a tool that's live, where the tool just looks at all these documents. You might have seen the footnotes there a second ago, 6, 7 different PDFs. The nurses don't have to read anymore, gets all put together in 1 place and nicely compiles in this preparation time goes from about 20 minutes or so down to less than 5 minutes, which gets amplified by the fact that you don't reach every member every time. So you really per member, you reach, save so much time in preparation and the nurse is so much more educated about the member, just quite a bit of opportunity here in managing our clinical affairs much better as well. Okay. One more video here. I'm going to roll that in a second. This is a workflow we're putting together at the moment there's a mix of things we've already launched and things we're launching, knee replacements, okay? That's a primary authorization. Oftentimes, insurers do this thing where they approve that little thing, but it's a step in a bigger process. What we're now working on here over the next few months at the moment already launching in the next few months is let's take that journey and authorize it all the way out. and say, okay, you will need that surgical procedure, but you will also need medical equipment, nursing care, outpatients, physical therapy, things like that. When we authorize this, in a sense, we earn the rights as well to then guide the member to the right place at the right time. So he got the physical therapist now. It's all real stuff. We get these pictures. I don't even know where that is here, but oh, yes, in Florida, of course. And then we can route you to these PTs that have capacity much more easily. Along the way because Oswell is there in a helpful manner, you can ask even clinical questions, right? Maybe you knee is hurting still 3 weeks after. We won't shy away from answering these clinical questions because we have the confidence that we know enough about the members and about these conditions and build these agents well enough that they can actually do this reliably, Oswell can pop back into the original workflow we authorized. It can look at that. It can pull in discharge nodes from the various physicians you encountered and some -- this is multimodal. So you can upload pictures of you need there as well. And of course, ideally, we are the connector, not just the delivery of care. So we can get you back in with a physician that shows up and our beta is a really good position for these kind of issues and for you as a member there. So really quite astonishing to see how much more of ability I think there will be in the future already is in the current Oscar system for having us orchestrate your care and not putting that burden back on the member, right? [indiscernible] of members are asked to be the advocates and the managers and the accountants and whatever their own care, we need to be able to take that away from them. And I think half and that's a big reason why brokers love us, why Oscar grew and out retains and now grows other insurance companies. Lucie marketplace, so much more to be done here as well. Separate from Oscar, as we talked about before, we don't mingle data. We don't mingle sort of recommendations there. But we do know a bunch about how the make members feel comfortable with, again, with AI and with technology and things like that. And so we have a team in the company was building a great tool here. we saw in the number of spike as we started talking a year. So it looks like people really need to buy better health insurance. Lucie the experience of building towards the area is one where you will tell us what your health care needs are, we can upload medical records, right? Intermobility is getting a bit easier now. I think we're the fourth one of that as well. Get these medical records in reason over them compare across the marketplace what works best for you as a member that does not have to be scary very clearly, right? There's so many other good health plans out there. I love the statistic of 75% of all doctors in an ACA plan, but 57% are in just 1 plan, hey, you got all this choice. You just got to pick the right land for a no family. That really is an algorithm problem we ought to be able to solve. And then, of course, bundling this with GLP-1s, bundling this with all kinds of other things is incredibly powerful. Forgot if we have another video here, we? Okay. [indiscernible] many videos here. Yes, enormous opportunity here still, I think we've just gotten going to have gotten to this point with a platform that is not scattered that is not vendoring out all kinds of stuff in fragmented is very powerful. I think we've held on to that to throw agents in the mix has been powerful already in the past 2 years. It's really taken off as you saw from these technology numbers here just in the last 10 months, that will keep layering on itself. There's growth and margin opportunities. We can do much more member experiences and retention and things like that. There's medical loss ratio purenesclearly in care navigation and actuarial insights affordability programs in terms of SG&A opportunities as well still, greater automation, self-service few money interventions. So again, it's a team sport, we have gotten these kind of results here in the bottom left, but the team is incredibly aligned in how they're able to use that technology across the board, and that is for someone like me incredibly powerful and fun to see, and I'm looking forward to so much more of this. And now we are at the Q&A. Thank you.

Please welcome Mark, Scott, Janet and Mario to the stage for Q&A. [Operator Instructions]

All right. So we are ready for Q&A. Who wants to ask?

Andrew Mok from Barclays. I appreciate all the color this morning. When we consider your comments on stable 2027 industry enrollment alongside your targets on revenue growth and market share this morning. It looks like you're expecting industry ACA growth to accelerate to very high single digits, if not low double digits in 2028 and 2029. One, is that right? And can you break down the components of that industry growth, including how much choice is reflected in that?

Scott?

Yes. We reflected on the fact that we have seen 3.5-ish million new lives coming into the ACA this year. A lot of those trends that are driving that, we expect to continue. Things like more and more people who are working multiple jobs they used to be in an employer-sponsored plan. Now they're in working part-time in 2 jobs. We see more and more evidence that, that is continuing. We think AI is going to continue to accelerate that. We think that just a core gig economy, individuals who are not part of a large organization, that's going to continue to grow. We see evidence that, that's been a big driver of recent growth. Expect that to continue immigration, while maybe not as high as it's been in the recent past will continue to be a driver. We see that as a fundamental. And then choice, we talked about the acceleration that we're seeing in that marketplace. I think, as Mark talked about this, that's a market that's kind of doing a little bit of a drip, drip, boom. And I think what we're starting to see is the real acceleration of the J-curve with choice where more and more companies are exploring it, and we are confident that, that's going to lead to more and more of those companies joining in the market. So I would say -- we're not going to go through each 1 of those as to the specific drivers, but those are the cumulative factors that are driving growth in the ACA.

And there are 2 sorts of phenomena. One is employees getting displaced for a part-time job or whatever and ultimately going to work -- going and getting their own ACA plan because there isn't any structured way to do it. But one of our anticipated approaches is to create an hour banking system within a wallet so that if I work for multiple employers, the notion would be the employer puts so much per hour work into that pool. This is something we used to do back when I was a union organizer back in college, create our banking opportunities for people to amass the money based on multiple employers and then go buy their policies. So there isn't a structured way for people to get it, but we now believe we can build structured ways for people to get coverage that are displaced by the employer-sponsored workforce.

Great. And just a follow-up on the choice. You mentioned that there would be a tipping point at some point even though unclear what are the barriers today? And what do you need to happen to unlock that growth.

Janet, do you want to cover that one?

Sure. Well, I think choice that what we've been talking about is to actually address and sort of build the, I would say, the highway for employers to come on to the marketplace, right? So when Mark talks about Lucie and what we're doing, it's in fact, to address some of the friction for employers is simplifying administrative connections between the employers, employee list, the ability to do the selection and buying for their individuals and to ensure that the payment transactions work. So there hasn't been an elegant solution in the industry, and that is what Lucie is about, creating that marketplace and putting together the infrastructure and also the coalition that we formed with ICRA in order to agree to standardize some of the connection points and integration, so there's interoperability. So it's really been understanding choice, I think, is the first part that it's an option for employers. And the second is how do you make it easy for the employers to make the switch? There are 2 major barriers in the thinking of employers and in consultants and brokers. Employers are worried about network access. Will all the people that work here have access to a network. We solved that with [indiscernible]. The second is, will my employees be able to keep a a competitive plan that doesn't have me left with just figuring out what defined contribution as every year and having that as the argument. And again, using the broker community to get people into the right plans based on their current lifestyle needs stabilizes the underlying risk of that population. On the consultant and employer side broker side, brokers hate the CROs because they lose the commission on the group. But what we've designed is the wholesale sale converting the employer and then turning all the employees over to the broker to convert them and getting paid commission by whatever carrier they place them with, gives them a lot more opportunity, giving them the tools to do it easily. The consultant side, I think, is going to be a fight over time. I think it's extraordinarily expensive for consultants to convert an employer. It's up to $80 per employee per month. We think there's a cheaper way to do it. we're investing any opportunities to do that. But that will be hand-to-hand combat. And obviously, they're going to advise large employers to stay where they're at until they can figure out how they make money from this.

Yes. And the truth is when inflation becomes a point where it just becomes too great of a burden for employers they're going to make the switch, very similar to the change from pensions to 401(k) plans. It was not easy for employees to decide to give up our pensions for their employees or is this sense of what they needed to provide and when a long-range balance sheet impacts? They said, we've got to make the switch, right? And it's been successful. It's been healthy. It's burned the whole industry. Very similar, I think, very similar path.

Go ahead, Steve.

Steve Baxter from Wells Fargo. Thanks for the questions and all the information. Just to come back to Lucie and the economics like you gave us that slide that had the helpful framework on like the per 100,000 economics and the margins you're thinking on that. I guess how much of those economics are, sort of, known today, contracted versus kind of have to be borne out in the market over time?

I think they're known in their current state. So the commissions that supplemental carriers pay as a result of building of connecting people together. The parts that we're pricing out as we speak is the rents on the actual marketplace, and we're negotiating those carrier by carrier or having conversations. How much does Lucie get when we create these connections and allow them to traffic through the site.

And then I know that the commentary, I think, was a modest contribution to the EPS target. Like any any general framework you kind of want to offer beyond that? And I guess, how should we think about how you might report this business over the next couple of years, we can kind of keep you honest on all these targets.

Yes. I'll start with your -- the end of that question, which is we don't expect to have significant amounts of Los specific disclosure until it becomes a larger part of the business as we think about how it contributes to the $4, I would just say this, if that business is not successful, we still believe we're going to be able to deliver more than $4 of EPS in 2029.

It's Michael [indiscernible] from Baird. So on choice, as employers move employees from group coverage into choice, how do those members compare with Oscar subsidize individual members on morbidity, risk adjustment and retention, acquisition costs, margins in general. Does employer sponsorship structurally improve the risk pool for individual marketplace? Or does it create some adverse selection by carrier.

I think that one of the things that we observed is when we see new people come into the ACA regardless of where they come from. After a short period of time, they all start to -- we can definitely see that the performance normalizes and looks very consistent and similar. So we believe that bringing more lives into the risk pool actually stabilizes the risk pool for the remainder of the population. So we do think that if we have the ability to create 2 opportunities to engage that member, 1 in Oscar Insurance, where you could be a member of Oscar Insurance, if you happen to have a provider and in a network that is in our footprint, we would love to have you be an Oscar member. But if you move and you go somewhere else and you need a different network and with a different employer, with Lucie then can capture that life and retain you. So the whole business that we're looking at and trying to tackle this individual market is about extending our relationship with members, both through Oscar Insurance and as well with Lucie. And we think that by using both of these 2 vehicles, we're going to be able to have significant duration and significant lifetime value from those members.

And on the ill employee side, so if you have a sick employee that moves over as part of it, we have the risk clearing mechanism in the ACA to amortize that over larger numbers of lives. So it's just part of the normal routine. We wouldn't want groups putting their sick boys in and keeping their healthy employees on a self-funded or partially funded plan.

Got it. And Mario cited 33% improvement in operational efficiency since 2024. I mean with further efficiency still ahead, how much has that shortened the J-curve for your new market cohorts, both in time to breakeven and mature contribution margins, how do those cohorts launching today compare with those, call it, 2 to 3 years ago? And how could Lucie drive further improvement through acquisition costs, retention and admin leverage?

Well, as I spoke about, the goal here is to extend lifetime value. That obviously is a -- gives you more opportunity to incur potentially even higher upfront acquisition costs if you have a longer-term relationship. We don't necessarily believe that that's the outcome where you end up with higher acquisition costs. We think we can have very efficient ways of bringing lives into the Lucie marketplace. Just to pull up on how is AI influencing the business. Hopefully, from the presentation today, you have a sense of the pace of how change is happening at Oscar. And when I look at how many more projects we can be doing with the same amount of headcount, right, like just the -- our ability to deliver significantly more throughput in changing our systems and delivering more agents with the same amount of people. That is the efficiency that we're seeing. You hear others talking about spending billions of dollars to try to increase their AI performance. We're actually spending the same amount of money and getting massively more out of the teams that we already have. I think that's important about the types of -- the quality of people that we have in this company and our technology organization. So your question about the J-curve, I would just say this, we see an acceleration of profitability on all members. And so it gives us more confidence in our ability to enter into new markets to hit the margin targets that we've got for those targets. And we see the evidence of that every day that what we're building is working.

One more comment before I turn it over to Kevin to ask his question. The other part of this is AI is not just a cost reduction. AI is retentive and allowing people to get serviced quickly and get things done. So on the growth curve, we've had a lot of volume and scale growth. And part of that is because of AI and the way we manage customers, onboard customers and service them. Kevin?

Kevin Fischbeck, BofA. My understanding is that the plans that you offer in the choice market are basically the plans that you offer on the exchanges. Is that -- if that's true, do the plans in the choice market have risk from legislation? Like did the pricing go up a lot on the choice market this year when it went up a lot for the broader market? And I guess, if that's the way that it works, how do employers think about that regulatory risk as they think about moving people on to the Choice market?

We are doing both on- and off-exchange products. So we price it out based on what we think the mix will be.

Okay. So the off-exchange being more stable, you would say -- and so that gives the employer then visibility.

Literally, what the employer does is they take what they're spending on health care unless the employee premium gain share and they divide it across their employees and everybody gets the same amount. That's pretty much the standard approach.

But Kevin, if you look at the performance of the ACA on trend, I think the trend over time in the ACA has been more favorable than what we've seen in the commercial space. And so yes, there's been some near-term pressure on rates in the ACA. But we think that with a stabilizing market going forward, we'll see a more consistent profile. We think that's going to be better than what we will see in commercial. So I don't think that the short-term headwinds in pricing that may have happened over the ACA in the last year or so are an impediment for choice to continue to grow.