Good afternoon, everybody, and welcome to the Babylon Capital Markets Day. So pleased to see so many of you here at Soho House in New York. We've actually chosen this location very deliberately because it encapsulates actually a lot of Babylon's values in itself. So, for example, great value for money, number one, very important to me. But actually also a great British export to the United States, but also incorporates a lot of diversity and creativity and entrepreneurialism as well. So thank you all for coming. The password is House7, S-E-V-E-N, for those who need to use the Wi-Fi. And apart from that, I just need to give a brief intro from our lawyers around some forward-looking statements. So before we begin, we'd like to remind you that certain statements made during this meeting will be forward-looking statements as defined by the Private Securities Litigation Reform Act of 1995 and as further described in the presentation slides for this meeting, which are posted on the company's website. These forward-looking statements reflect Babylon's current expectations based on the company's beliefs, assumptions, and information currently available to the company, and are subject to various risks and uncertainties that could cause actual results to differ materially. Although Babylon believes these expectations are reasonable, the company undertakes no obligation to revise any statements to reflect changes that occur after this meeting. Descriptions of some of the factors that could cause actual results to differ materially from these forward-looking statements can be found in the risk factors section of the company's annual report on Form 20-F, filed on March 30th, 2022, and its other filings with the Securities and Exchange Commission. In addition, please note that the company will be discussing certain non-IFRS financial measures that they believe are important in evaluating performance. Details on the relationship between these non-IFRS measures to the most comparable IFRS measures and reconciliation of historical non-IFRS financial measures can be found in the presentation slides for this meeting, which are posted on the company's website. Before we also begin with the main presentation, we've got actually most of our senior leadership team here with us today. We'd love for you to interact with them during the break and also after the presentations, so you can ask any follow-up questions. And in the meantime, though, I'm just going to ask everybody to introduce themselves and give 20 seconds on their background as well, just so you can identify who does what. And just because he's the closest, I'll hand over to Steve Davis. Hi, I'm Steve Davis, Chief Technology Officer at Babylon. I come from Austin, Texas. I've been with Babylon going on probably almost two years, I think a year and seven months. More? Yeah. Okay. Your background? Oh, background. Oh, yeah. Your story? Yeah, my story. I come to Babylon from Expedia Group. I spent 15 years in travel, helped build a company called HomeAway, now called Vrbo. Many of you may have used it in your travels. We sold that to Expedia, actually Dara, who's now the CEO of Uber. I heard Uber has a glitch today. I need to text him. But yeah, we did the deal with Dara, and I stayed on board with another exec and helped to run Expedia Group for the past five years. I was the SVP and GM of the data and platform throughout all brands across Expedia Group globally. And like I said, Ali convinced me to come join them and having a fantastic time. My colleague Jon. Hi, my name is Yon Nuta, Chief Product Officer. I come from over 20-plus years building great products, many of whom you and your kids have probably used. I spent seven years at Microsoft with Xbox when we first started, helping them grow with Microsoft Office and Windows launching Office and Vista. Since then, most recently, I've come from Gaia, which is the largest wellness streaming company. I'm sure many of you have used them before. Super excited to be here. Hi, everybody. I'm Darshak Sanghavi. I'm the Chief Medical Officer here at Babylon. I just came in this morning from Boston. I'm a pediatric cardiologist by background. I've taken care of hundreds, if not thousands, of kids with complex congenital heart defects. I came to Babylon because I love this notion of affordable and accessible healthcare. It's been a theme throughout my career before I came to Babylon. Here, I'm responsible for our care models, quality, and really, essentially, the clinical delivery of all the services we offer here, but my background, I spent time with a number of international organizations working in Kenya, working on tuberculosis control in Peru, eradication of cysticercosis in Central America, and I also worked for several years on the Navajo reservation. I've been particularly interested in how we can bring technology to address the problems that affect large populations. An example is if you had a baby in the past 10 years, you probably saw their baby was checked for a congenital heart defect. That's through the program I worked on, Universal Pulse Oximetry Screening for newborns. As a result, congenital heart defect deaths have fallen dramatically in the U.S. I also, for several years, worked in President Obama's administration designing. I was a Director of Prevention and Population Health at CMMI. I created the Million Hearts program, expanded the Diabetes Prevention program, and created the first program on social determinants of health for the U.S. called the Accountable Health Communities. I then spent four years as an executive at UnitedHealth Group, the first half at Optum as the Chief Medical Officer at Optum Labs, directing our big data platforms and a number of programs in the opioid crisis and maternal health. And then just before I came to Babylon, I was the Chief Medical Officer for UnitedHealthcare's Medicare business, $90 billion P&L, the largest MA program in the country. But Babylon is where it's at. So that's why I'm here. Thanks. Good afternoon, everyone. My name is Paul-Henri Ferrand. I go by PH. I'm the Chief Business Officer of the company. I oversee the go-to-market function, sales, marketing, business development, as well as the relationships. I came to Babylon a year and a half ago. At that point, I think we're a $15 million MRR business. In the past year and a half, you've seen the results. I mean, we're growing by a staggering pace. I joined Babylon because I ran into Ali, and he wowed me with his mission and technology and talent that he had assembled around him. I came from Brex, where I was. This was my first foreign startups where I was their COO. Before that, I'd been with Google for about seven years, where my last job was to be President of Google Cloud. Before that, I was in tech with Dell, many, many years, and Nokia and a few other companies like that. Great to be here. Hi, everyone. Samira Toxiri. I am Babylon's Chief People Officer. I have the honor of partnering with the rest of the SLT to make sure we bring in exceptional talent into Babylon, retain them, keep them at Babylon, and help them grow. Over the last couple of years, last year specifically, we hired over 1,200 new Babylonians, which was absolutely fantastic. I had the smallest or shortest commute because I come from Westport, Connecticut, so didn't have as far to go as many of my colleagues. Prior to joining Babylon, I've been with Babylon for about a year now. I came from GE, specifically GE Digital, where I was the Vice President of Talent Management and helped run a few different transformations in the digital world for GE. Prior to that, American Express, and before that, about 14 years of HR management consulting with the legacy Towers Perrin, now Willis Towers Watson. So glad to be here and see all of you. Great. Thank you very much. And finally, I think I know most people here, but just for the benefit of those on the webcast, I'm Charlie Steel. I'm the Chief Financial Officer at Babylon, been here just over four years. Joined Ali in 2017. I have never looked back. It's been quite an experience and helped run the listing last year. Prior to that, I was at CMC Markets, where, again, I ran the listing for CMC Markets as Head of Corporate Development onto the FTSE 250. And then before that, I did investment banking at Lehman and Deutsche Bank. Before we begin, I just want to quickly show everybody sort of a perspective really from the patient and how they'll see things to give a little bit of background to that before introducing Ali. I'm 35 years old. I have six kids: a 15-year-old, a 9-year-old, an 8-year-old, a 4-year-old, and a 2-year-old. I live in the Bootheel in Missouri, and I've lived here my whole life, actually, so 35 years. My husband and I have been together since I was 15, and we just never broke up. And so we've been together almost 20 years now. My husband works on the river, and for the most part, I'm a stay-at-home mom, but I do work a little bit in home healthcare for a family member. My life is pretty chaotic, so every day is a little bit different. I have four kids that are in school full-time, and my two boys that are my smallest will actually be able to attend preschool in the fall. So I don't know what I'll do with my time after that, but for right now, it's just taking care of the boys at home while the other ones are in school, running around, doing all of the appointments that we have, and making sure everybody's kept well. There's always something going on day to day that is either mental health-wise or just behavioral-wise, or that's aside from boo-boos and things that you got to do that most people see doctors for accidents and things like that, or something just doesn't feel right. But in our house, healthcare is an everyday thing because we struggle with different kinds of healthcare, well, different kinds of health issues, especially my oldest child, because she's almost to that adult age. And so I watch her go through so many things that I didn't go through just because of her mental health. It's hard for me to be like, "I understand where you're coming from," because sometimes I don't. So I can't really say that to her. And so when she's struggling, it's just like you're trying to be there as a mom. You're trying to fix something that you don't know how to fix. I had always understood anxiety because my mom always had a panic disorder. And so I always knew anxiety, and I could tell from a very small age. Everybody used to say, "Oh, she's a drama queen," or, "You just baby her because she's your first one," because between my first and my second is almost six years apart because we had the infertility. And I said, "No, this is something else." So as my first child, I'm sitting there like, "I'm not parenting her right. Something is wrong, but it's got to be me because I'm 24/7 with her, and it's my first kiddo, and I don't know what I'm doing as a mom." So it took me until she was nine to reach out for help from a psychologist. And then when they got her in and they started doing all of the questionnaires and all of the tests and stuff, they found out that she definitely has all of these problems. And so it was relief. I think a lot of people say that. It was relief to have a diagnosis, to know that I wasn't crazy or that it wasn't something that I did as a parent. Sometimes you feel punished as a parent when you have a child facing mental health because you thought, "This has got to be something I've done. This is my genes. I just haven't done enough for them." People may call me crazy because you've got 10,000 doctors or something, but I want you to have access to every single person so that you can get this under control and you can take care of what you need to take care of. We are actually set up with a behavioral therapist through Babylon right now that she's been seeing. She's only seen her a couple of times so far, but she's really great with her. She's super sweet and understanding, and she knows her stuff. So it's been good to let her get on. She talks a little bit with me, and then she'll get on there with my daughter, and then she switches back to me, and we just go over maybe some things that we could try before our next appointment. But yeah, because we have to go to doctors everywhere. And in our area, we can't access doctors just right in the same town as us. So we have to drive 40 minutes, an hour, just to get to a doctor. If you've ever been in a doctor's office with children, you will know that they do not, they just don't have that ability to sit, be quiet, hold on, you're next, and that kind of thing in the waiting and everything. So telehealth has been amazing in the sense of they're in their own zone. They're in their home. They can let me talk to the doctor while they play or watch TV, and then I bring them over and let the doctor assess them or talk to them, and then they can go back to what they're doing. We haven't left the house. We haven't had to spend gas money, and everybody's in their own comfort zone. Hey, Ali, over to you. Yeah, thank you. So thank you, Charlie. I'm going to, we're in a tiny room, but I'm going to use this just because of the webcast. Look, the healthcare that Jessica gets is this. It really isn't healthcare. You know better than most. It's sick care. It's reactive. It's episodic. It's physical. It's one size for everybody. It's entirely provider-centric. Jessica, with six children, needs to go to a doctor, find time, figure out what to do with everybody else. It's all about what does that doctor know. And it's all about, and frankly, the model, it's fee-for-service. It really is designed around a different model. A good friend of mine, Daniel, created this slide, although the credit for Daniel seems to be missing from here. He kind of talks about a healthcare model that is proactive, continuous, virtual, personalized, patient-centric, is AI-enhanced, and is value-based care in order to align incentives. Who's building this? Which company do you know that is genuinely thinking about how do I rebuild healthcare from a sick care industry that is paid in order to deal with people predominantly to a healthcare industry who gets paid in order to keep people healthy? If you follow the money, that is not where most of what we call healthcare gets paid. We think that we are in the middle of a transformation of an industry. Companies who will do this will create some of the most valuable companies on the planet in one of the most valuable sectors in the world. And to do that, I think the winners of those are going to be the people who don't just tinker on the edges, but fundamentally re-engineer the value proposition for customers. Everywhere in the world, value, where you go, I mean, value in general is defined as quality over cost. And everywhere in the world, we kind of define value as quality as accessibility and clinical quality, and if you want, cost as affordability. That equation of accessibility, quality, affordability is what every country is dealing with. In Europe, we're very good at clinical quality and affordability. We have a challenge with accessibility. In the United States, we're very good at accessibility and quality. We have a challenge with affordability. In many of the developing countries that we're in, people have problems with all three. In Babylon, for us to fix this, we believe as long as we can deliver most of the healthcare most people need on devices, on mobile phones, most already have, we deal with accessibility. In Rwanda, one of the poorest countries on the planet, financially poorest, not culturally, but financially poorest on the planet, we now look after a third of the population, give or take. They can make a quick phone call. They don't have smartphones, but just a basic feature phone and talk to a healthcare professional within minutes. Everybody in this room is among the top 1% richest people in the world. I bet many of you don't have that ability. That's some of the poorest in the world can do now, right? One of the proudest moments in the history of Babylon was when President Kagame made universally available for free primary care for his nation by giving Babylon to everybody. Secondly, it's clinical quality. The way we deal with quality, our belief is that if you can unify, standardize the delivery of care, you could have a much more quality care base. Currently, most healthcare is delivered on whatever it is that your doctor knows. It's a lottery. We all think that we know the best doctors in the world. We've approached them. We know them. It takes 17 years for best practice in medicine to become common practice. Last year, I think 11,000 papers got published on dermatology alone. How do you know that that doctor that you go to in that fancy office, how up to date are they? I used to run in my previous job, CIRCO, a partnership of 1,700 specialists in Europe, the largest in Europe. One day, we sent a survey. We asked people, "Where do you think you are in the continuum of your colleagues?" Over 85% of them put themselves in the top 5% among their peers. The reality is that there is a massive divergence between the knowledge of people for what you do. And unless we have some kind of platform to unify that knowledge, we are always at the mercy of the single brain that you could see. So what we believe in Babylon is that we can standardize so that every patient goes to the right expert with the right treatment and the right rehabilitation, and we'll come and show you what the work that we are doing on this. And then finally, it's about affordability. Again, on affordability, we believe that there are two things that fundamentally change the cost in healthcare and affect the cost in healthcare. The first one is about two-thirds of the cost on average, but in the United States, it's more like 50%. In Africa, it's more like 70% of the cost is in people. Doctors, nurses, healthcare professionals, salaries, this group of people are among the rarest and justifiably among the most expensive assets of every country. If you don't automate what they do, it really doesn't matter what else we do. By the time we fix the other 30%, inflation has taken care of all that cost. There is no solution in which you can fundamentally change affordability without dealing with automation in the things that people do in healthcare. There's just not enough of them. Today, we had a group of people from PSI, one of the world's largest charities, visiting Babylon's offices in London, and we had our standup where 1,000 Babylonians get together every week, and they were presenting, and they said something like, "In their estimates, the world is short of 18 million clinicians." Without automation, nothing changes. The second part of this cost, of course, is about two-thirds of all the cost. 50% is what my lawyer says that we can verify, so let's go with that. 50% of all the costs are in predictable, preventable diseases. Nobody needs a root canal. It only happens because nobody's watching. The most expensive thing you can do in dentistry, if somebody was watching, if you were monitoring our teeth, it will never happen, so these are what we built Babylon on. How do you create accessibility by giving most of the healthcare most people need on devices most of them already have? How do you create quality by standardizing? And how do you deal with affordability by tackling the two significant costs in healthcare? To do that, we needed to create an integrated digital-first model that is built from scratch. That takes time. Our team, Steve and Yon and the rest of the team will show you some of the technologies we're building on this. At the core of that sits data. How do you collect all the data from every member of our organization and as much as possible? My car used to break down five, 10 years ago. Which one of you has had a car that has broken down? It just doesn't break down anymore because we buried so many sensors, and we collect so much data, and we analyze it in real time that we can stop that. Who's doing that with our body? We believe that there will come a day we can collect so much data on you that you will never have a surprise. You'll have an accident, but you should not have a surprise. Steve Davis, my colleague who just introduced himself, will come to you and talk about the health graphs that we're building on this. And why him, by the way? He introduced himself out of Expedia. Think about the data in real time that that industry collects. Every hotel, every airline, every car rental company. We know their data all the time. Who's doing that in healthcare? We think that you can fundamentally change that game on its head by taking the data, continuously showing it in real, getting real-time insight out of it, and then being able to set up a plan for everybody. 20% of people, as you know, are responsible for 80% of all costs in healthcare. If we can predict who those 20% are and prevent them or manage them, we should be able to significantly cut costs. We'll show you the AI programs that we're running already on the health graphs that is enabling us to start that process. Now, I'm not going to pretend that we got this all sorted, but we're well on our way. When we give people a plan, we monitor them continuously, and we reward those who are doing well. We alert those who are in danger, and then we intervene early for those that we need to intervene. We intervene first through 24/7 care assistants, so everybody gets almost their concierge. We connect them to the virtual multi-specialty team. It's not just primary care. It's secondary care. As you saw with Jessica, it's behavioral care. And when necessary, we'll send them into clinics. In London, we run our own clinics. This month, we opened two. But in the United States, where there's quite a lot of clinics now, we partner with people who have existing clinics, and we just send our patients to them fee-for-service. When they go to a hospital, we bring them back. Our acquisition of Day to Day gave us access to capabilities that allow us to manage people at home. If necessary, we send their drugs home. We can manage their drugs. We can make sure that we use drugs that are economically and clinically of better value. And for those who are really in need, we give them a full care team to look after them. And you heard that again from Jessica and her daughter. And then that end-to-end will allow us to do all sorts of chronic care management. Gone the days that you have one single company doing one single chronic care management in complete isolation and silo from everybody else. Why do they do that? They did that because they used to sit on these relational databases that had no AI, and it could say, "If that, then that. If that, then that." Then there is only so much you could do. With the infrastructure I described to you, you could do both chronic care management and complex care management. Imagine what we are now trying to do with this, which is put all of that into our AI and allow our AI to watch every interaction. Every consultation we do, we tape, and our AI can learn from it. Every chat we do, our AI can learn from it. We're building and rationalizing a new infrastructure that doesn't just collect data, give insights, set up goals, plans, monitor, and reward, but does so first with a digital layer that is fundamentally run with AI, with no human interaction. You've also the statistics of around 40% of our members that now every interaction they do with us, about 40% of which has no human in the loop. We pass people on to care assistants who are not clinicians but are well-supported in order to provide the support that they need. We even launched our AI now in Rwanda, where our care assistants in there can actually have access to the best information and have a conversation with the members on the other side and have the AI to advise them. And then, if necessary, we'll send them to human doctors, first virtually, and only one in 10 times or 15 out of 100 times to a physical place. In 85% of all cases, we can deal with you without you leaving your home, and then, of course, to complex care. So how do we make money? We make our money by either licensing our technology. We are in a threshold in that business, and we'll talk about it. We took a technology that worked for us in the U.K. and, frankly, was kind of figured out to work in a number of other countries. We now are rebuilding that technology almost from ground up so that it can be international, multi-country, and be SaaSable, frankly, so other people can license it from us. We've given it to the National Health Service in Britain, Prudential across Southeast Asia, TELUS in Canada, so on and so forth, to give us a series of clients with whom we can use. If you want, these are our beta cases before we make it SaaSable and give it to everybody. We, of course, do fee-for-service when necessary. We do it for Bupa, Centene, this and that. We cover about 4 million people in the United States. But really, when we sell our technology, we're getting $1 or $2 or $3, whatever it is, PMPM, a few cents per person per month. When we are doing our fee-for-service, we're getting tens of dollars. But really, what we're interested in is to be able to take the entire budget where you get hundreds of dollars PMPM, thousands of dollars a year, because even a 5% margin on that, a 10% margin on that, is a much bigger money than all of that thing above. Lots of people tell us, "Well, yeah, but technology has 80%, 70% gross margins." That's true, but if you run your business on cash, because at the end, it's cash that pays salary, 10% of thousands of dollars is a lot more money than 80% of a few dollars, right? And also, it allows us to do what we want to do, which is to align interests. Our business, I mean, we've been doing this in the U.K. Our business is very well in the U.K. We showed in data provided by the National Health Service, NOTAS, the NHS's data, that we can save up to 35% of the cases of costs between 15%-35% depending on the complexity of your work, which is really interesting, actually. If you look at that thing in here, that the patients that came to us are much more complex than normal patients. So actually, the person who switches to Babylon, because if you're kind of very happy and you don't have any needs, why would you switch? But the people who switch to Babylon are 22% more likely to use A&E and 62% more likely to use NHS 111, which is our emergency phone line, and only after a few months, 12 months, I think, maximum with Babylon, they were well below average. The cost saving was significant, 25% reduction in emergency room, up to 35% reduction in total cost. Now, we then thought, "Okay, we bring that into the United States." When we brought it into the US, we needed to choose the business model we want to use. What we saw in the US was that you have a lot of physicians single-handed, brick-and-mortar businesses, single neighborhoods, so on and so forth. We saw then exactly, I mean, I'm old enough to remember when the digital transformation happened to retail, right? We saw that the franchise model exists in here. People go to physicians, ask them to partner together, maybe join them. They take their back offices over. They take their finances over. We saw the retail chain model or the concept chain model where people build the same shop in every neighborhood. We thought nobody's doing an Amazon in here. Nobody's really creating a digital-first model that can scale massively. That was the attraction. A model that is mostly virtual, that it employs these doctors rather than partners with them. It doesn't work on neighborhoods. It's regional. It can be national. It's continuous. It's leveraged automation, and it's high growth with medium capital needs. We brought that in the U.S. two years ago. I was just telling somebody in here. In January 2020, I was here. We had in San Francisco for a conference, and we didn't have a dollar of revenue, give or take in the U.S. January 2022 was the next time I could come to the U.S., and we almost were running at a run rate of $1 billion of revenue. And every one of my leadership team that you saw, we hired during that period remotely without even meeting people in person. So people say, "What are the results?" Just think about what we had to do in two years' time, almost virtually, and compare that with the speed at which everybody else has been growing. We now have over 4 million people with whom we provide clinical services. We grew our value-based care, which is the business we really care about, 16x from 17,000 people to 271,000 people in Q1, the latest results we announced. Our revenue growth from $29 million first year to $228 million, and now it's a billion-dollar run rate. We went from two states to eight states. We'll be in 20 states shortly. And throughout it all, our quality score has stayed over 95% for an five-star rating. I challenge anyone to find a model that can deliver that if it's not digital-first and it's not based on our model. That level of growth, that level of customer satisfaction, that level of thing. And people kind of tell us, "And okay." And by the way, our growth has not even started yet. If you think about 271,000 people, we are PH and his team are in discussions with 153 various payers, right, PH, like in various one way or another. If you look at only on the payers whose name you saw before, and you think that we do a tiny percentage of their business, and even among them to grow, this business can grow. Growth is not our challenge, our issue, as most of you know. Six months ago, if we were sitting here, everybody's focus would be on this slide. How fast is this business growing? But we understand the market has changed. So what people care about is now how are we performing. We always cared about what's how we're performing. The beauty of our model is we can learn at a much faster rate than any non-digital model can learn. So if you look at our penetration acceleration of just sign-ups and registrations, in our very first model, whose results you will see now, it took us almost 80 weeks to get to 29% of the sign-ups. In Georgia, that number kind of got fast-forwarded four to five times, only almost a year later, four to five times, and in Missouri, 8-10 times. The speed at which we got almost to the same percentage numbers massively grew because we can collect the data, we can analyze the data, and we can learn from every interaction. Today, in our standup, people were showing the difference between two SMS messages. One that says, "Remind me," one that said, "Come and see us to do this," versus one that said, "Come and see this person because she wants to talk to you," like come and see a nurse or come and see a therapist versus come and do a consultation with us. We can learn very quickly in a way that it's super hard for non-digital models to do. And the results speak for themselves. So what we are seeing, and this is done over the same time period, is that as penetration goes up, you're seeing, for instance, inpatient numbers coming down, an average length of stay coming down. And that leads to people ask us, "So how are your numbers doing?" In first year, in Medicaid, some of the most challenging areas, we're already seeing that our cohorts of people are breaking even, 2% give or take, right? The reason that the totality of the book is below MLR of -0.5% is because many of these contracts, as you know, are very, very new. They are almost a quarter old. We hardly have had a chance to look at those contracts. But when you look at the older contracts out there, then people, when you look at our Medicare Advantage, which is where often people compare us, we say, "Yeah, but this company does that kind of Medicare Advantage, medical things." You see that for our contracts that are only a year old, we're already getting 14% margins. That's a year-old contracts already. And then contracts that are commercial; we're already at 20% margins. I think that is already as good as anybody. And it's interesting. There are not that many people in Medicaid, but where people are in Medicare and in commercial, we are in line with people who've been here 10 years already with all the experiences they got, all the shops, all the physicians, everything else. And Charlie will explain that more. So what we are also seeing is that our cost is significantly falling. You've seen all of that compared to our revenue. We're seeing a 921 basis points reduction in adjusted EBITDA on the revenue, and SG&A and technology costs falling as a percentage of revenue significantly too, and that continues to drop. And fundamentally, if you follow the model, it has advantages that will keep accelerating, right? Our scalability will just keep getting stronger. I talked about AI. I don't want to take the thing away from the team. They'll show it to you more. And the digitizations of manual things and the platform leverage will work. And at the end of the day, it's all about the scale team. We think that when we get to around $3 billion of value-based care revenue, even if the average numbers is around 7.5%-10%, because you adjust for the fact that we will always be having new cohorts coming in that will bring the average down. And also remember that even today, about 50% of our revenue is already Medicare. And that is even before we began really the balancing between Medicare, commercial, and Medicaid. A year ago, almost everything was Medicaid. Now, about 50% is Medicare, and we will also build commercial on top of that. So these margins will, if we just replicate the margins from below, those margins will work. And then by the time we sell about $150 million of licensing, you get to a point that you should break even sometimes around here, $2-$3 billion of revenue, let's say $3-$4 billion if you're at lower margins. And then after that, every dollar you almost make on your gross margin kind of falls through your cash because your operating costs is give or take constant. So that's why every platform digital company, once they become, every time, all the time, people argue about when are you going to become profitable. And when they do, they become a massive cash engine. I've seen it in all. I've seen it in Amazon. I saw it in Tesla. And you just need to go through that hump. Scale will give you margins. Margins will just go straight into a cash engine. But you don't have all the other costs. And I'm not shy about the fact that I fundamentally believe that our story in the US is not known well enough, and the market is not treating us the way that it deserves. On almost every measure, we should be at multiples of the revenue valuations we are at. And that's just looking at where we are. And our job, and today's part of this, is to try and engage with the investment community more to explain our story in greater detail and try to persuade people, to try to show them, share with them significantly more. With that, what I'm going to do is pass it on to Steve and Jan, who are going to, starting with Steve, who are going to show you, share with you some of the details of our technology. Unfortunately, our Chief Science Officer, Saurabh Johri, who's been with me over five years, has caught COVID, so he couldn't pass the COVID test to make it to the flight, although he's feeling good. Is Saurabh going to join us remotely? He will be joining us remotely, by the way. Okay. All right. Thank you. If you see me trembling, it's not from nerves. This isn't my first rodeo. It's freezing cold in here. So just heads up. A couple of things. One, I talked a little bit about my career. I think it should also be noted, one of the reasons I came to Babylon is I truly believe that humans are the most valuable asset in the world. My daughter's a trauma nurse. I talk to her every day on the drive home, and I was just so inspired by just the stories she would tell me on the impact she was having on lives, and I looked out the window one day and thought, "Here I am spending the past 15 years kind of shaping how people travel. Why don't I try to shape how people get care and really make an impact?", and so when you look at the companies that have come before us, many of you don't know, but over the past 20 years, how you shape and how you do things have really been influenced by these companies. Whether you're streaming, whether you're driving, whether you're shopping, whether you're traveling, all these companies had a specific purpose-built strategy in mind, and that is they built a foundation to run those companies on collecting as much data as they could. They deeply integrated those platforms into the overall technology strategy, and they really invested heavily into AI and how that influenced the behavior and what you did. An example is I've owned a Tesla since 2012. I think my VIN number was some 1,000-something. So I've been a huge believer in Tesla, and so if you look back since 2012, they have 10 years of data on how I've driven my car every single day of the year. When you think about traveling, everywhere I've been, when I was at Expedia running the platforms there, we collected on average about 100-150 billion events just in an afternoon, so that was just an afternoon, the amount of data that flew through that ecosystem and how we leveraged that to actually shape the experience and how you actually used our products. Amazon. I think there was a stat one day that was published that the Amazon homepage is one of the most valuable, basically, websites in the world, that homepage. That homepage is directed and influenced directly by an AI team, and it all comes from that data that's coming in that helps inform how you might want to shop, how you might want to discover. The reason I tell you this is because the fundamental belief we have at Babylon is that we took that very same approach as there weren't vendors out there that actually provided platforms that allow us to deliver on this value-based care mission. To do this, we spent the last year and a half or so building a pretty incredible foundation, which is purpose-built, that is basically rooted in the strategy of how do we collect as much data as humanly possible from the disparate systems that have traditionally been decoupled. I am going to talk a little bit more about more of that as we go through this. We have also integrated deeply our AI technology into every aspect of our product, both how the data comes in, how our products work, how we deliver care. AI is just not a cool piece of technology that sits on top of a data lake. It's a platform that's been deeply coupled into the services that we build. If you open the hood at many of those companies I just went through, you would find the exact same reference pattern. In fact, you'd find many of the same engineers and architects that have built many of those platforms that have come before us are actually at Babylon working with me now, building this platform to deliver on this vision. And lastly, each one of those vendors that I spoke to spent an incredible amount of time and energy into basically a scientific approach about how they delivered amazing member experiences. And a member experience doesn't start with just a cool app that looks great, that might be easy to use. It's an application that allows for itself to learn, to change, to adapt. As Ali said, we had a story this morning of just the subtlety of a simple piece of text changing drove conversion by an astronomical amount. Our ability to have a member application that we can push out to the app store every single week of the year and actually adapt and inform to those changes, actually change the design, change the interface, and really drive member delight. We believe that's the secret sauce and engagement as well on top of these other three fundamental principles on how we build software, and again, I challenge anybody to look in kind of the ecosystem of the market. I think you'll find lots of vendors that are really good at specific areas, but there is no clear winner on who's really shaped the way that we actually navigate and then care globally that you could actually point to. There's nobody that's really cracked those codes. We believe this is an opportunity for somebody to win, and there probably won't be one leader. There'll be many, but we clearly want to be one of those. I think the other thing that's different today, and I actually debated with Ali about this, is this is going to be a bit of a different review for us because we're actually going to demo code. I'm not going to show you a bunch of pretty PowerPoints and talk about the cool stuff we're doing. We're actually going to show things working. And so my colleague, Saurabh, is in London. Unfortunately, he couldn't be here. He's actually going to demo some of our models being deployed, some of our algorithms running. I'm actually going to go through code working in real time. These are in our production environments. This is code that we've written. These are platforms that are built, and this is data that is streaming through, but we thought we'd actually represent how things work in a bit more detail, so we're probably going to get a little bit nerdy on you, so I apologize, but we actually want to show you a little bit different under the hood at Babylon today, and like I said, we're super proud of what's been accomplished. At the heart of our innovation is really what we call the health graph. If you think about healthcare data, and by the way, I started my career in healthcare literally almost 25 years ago. I was telling, I think it was Dave at lunch, it's like not much has changed. It's still a lot of data in very disparate systems that is very difficult to access. It's both unstructured and structured. Some people follow the standard. Some people don't. Some people publish to the Care Quality Network. Some people don't. Some people let you gain access to their data. Some people don't. And so our ability to really be the best in class and how we collect the most amount of data that we can, bring that information together, put it into a Firebase standard, take unstructured and structured, and pull the wearable information along with that, which is the IoT devices, which have really evolved over the past two years. We've seen a hyper-acceleration of that. In fact, my old boss I just saw is the CEO of Oura Ring. Imagine how much data now is being collected by the hundreds of millions of Oura Rings out there all flowing through the ecosystem now, and so we believe this is a fundamental difference in what we see with the competitors in the landscape out there, and I'll show you why. The next piece is, once we have that platform and we have that data streaming in, do we have a platform to actually take action and do something in real time, and the answer is yes. We've been hard at work at building this functionality for over 12 months. We've also built a tightly integrated AI platform. We call that Pottery, and so that's deeply coupled with our data stack. That's not just a job that runs at night that actually looks for some cohorts. It's actually inferencing and decision-making that's happening in real time as that data is flowing in and changing. And then we not only use this as just an analytics platform. This stack is actually powering our product that Jan's going to go through today. It powers our AI platform, and it powers our analytics. So it's a modern data platform with AI tightly integrated, actually taking a very different approach in how we actually build modern software. And so I'm going to walk you through an example. Unfortunately, I can't pause. I might let it loop. But on the left, I'm showing you our application. So this is really to represent what we call our streaming fabric. So this is how our data flows in, whether it be in the app or in Apple Health. I'm going to show the data flowing in. We're going to show kind of BMI changing in real time. This is actually a snippet of our health record. So you can see this is the health record in the system. These are the large streaming events. So I'm just going to click play and kind of let this go through. And so on the left, they're going to enter in the weight into our app. So this is the My Health aspect of our app. So if somebody's entering their weight, that could be pulled from a scale as well in an automated fashion. So we're typing in 187. We can immediately see in the middle the BMI calculation is immediately updated. So that's all happening in real time. That's the same record a clinician would look at. Okay. And so we're now going to go into Apple Health. Inside Apple, we're making that adjustment to 190. We will immediately see here near real time, the 190 has just changed. We now have basically both Apple Health and our product streaming in real time. You see the information changing in the middle. The right is just the, these are all synthetic IDs, by the way. This is just the number of events that is flowing through. There's some snippets here on last claims data at risk for hypertension. These are some of the things that Saurabh, my colleague, will go through as well. I just want to show you, this is actually our working product in production today and how it works and how that streaming platform is basically integrated into our overall stack. Okay. And so with that, I'm going to turn it over to my colleague, Saurabh Johri, who's in London. So we're going to try to do a handoff here. He's going to take control of the Zoom. I'll let him introduce himself, and then he's going to basically go through our AI stack and some of our predictive engines. So Saurabh, can you hear me? I can hear you. I'm just trying to get set up with the Zoom. One second. Okay. Okay. Let's get back to the slide which we're on. Go to slideshow. Okay. So let me just actually step back for one slide. Just give a brief introduction to myself. Let's see if you can all hear me. Yep. That's affirmative. Thank you. So yeah, as Steve mentioned, my name is Saurabh Johri. I am Babylon's Chief Scientific Officer. I've been with Ali for six years. He mentioned five, but six is the one year I'll give him. I've been building AI systems for healthcare now for the last 20 years, solving some of the most complex challenges there are from drug design to genomics to pandemic modeling, and I'm really excited to speak to you today about some of the work which we're doing at Babylon, building on that complex data which Steve showed you in the health graph. So if I click forward, what Steve showed us was the kind of architecture for the health graph. But the true value of the data from the health graph actually comes from our ability to extract predictive insights about our members through the development of predictive AI models, which help us to better understand how to take care of our members, providing effective clinical care and decision-making. So to do this at scale, as Steve already mentioned, we've developed a proprietary plug-and-play AI platform for healthcare, which we call Pottery. So it's deeply integrated with the health graph, as Steve also mentioned, and it extends this kind of familiar paradigm, which is train, test, evaluate, deploy, and monitor. But it extends it importantly to cover the specific use cases that we have at Babylon. And the specific use case there is to enable the development of so-called causal models. And those are really important for us to deliver interpretable, explainable insights and recommendations, which can be actioned by our care teams, which can also provide enhanced monitoring of our models because we're able to incorporate real-time clinician feedback into our models. Pottery also does a lot of the heavy lifting to pre-process the various sources of data which we build AI models in healthcare with. So you'll be familiar with claims, medical records, audio-video, medical images, labs and pharmacy data, and so on. But the real benefit, I think, when all told, is the increased productivity that these AI platforms bring to our scientists and engineers, allowing them to focus on building high-performance machine learning and AI models rather than having to deal with complexities of DevOps and engineering. So this is all rather abstract. So if we go to the next slide, I'll show you what's happening under the hood here. So I'm going to pause in a few spots during this video. So if we play the demo, what we can see here is the kind of simplicity of creating an environment on the platform. So what I've simply done there is, through a simple command, I've provisioned a large amount of cloud and compute capacity. That's really important to provide us with GPU instances, which are these kind of high-performance chips, which allows us to train machine learning models. What's also happened here is that we've automatically set up the tooling and infrastructure to build these large-scale AI models. If we then just clear that and go into the environment, which has just been set up, what we're doing here is importing a model, which we call DPU. We're going to spend a little bit of time talking about this model. It's actually a 300-million-parameter AI model, which we call DPU. DPU is Deep Patient Understanding for short. I'm really going to geek out on you here. We've developed this model really to leverage some of the major innovations in AI of past 18 months. These major innovations are in a technique or in a type of neural network called transformer models. And those have really created significant breakthroughs, as I've mentioned, over the past 18 months, particularly in the areas of language understanding. And you will have heard about models such as GPT-3 and more recently, DALL-E. So you can think about DPU as this foundational model, which boosts the performance of any other model which the authors offer it. So whether that's a risk stratification model, a disease-specific risk model, or even an NLP model, which we'll see later. So if I play the video again, here we are. So if I play the video, we've imported that DPU model. And now what we're going to do is to select a cohort of high-risk members from the health graph. And that's exactly what we're doing. And so here, what we're going to do next is to quickly train a baseline disease-specific model, in this case, a 12-month diabetes risk prediction model that leverages DPU. So it's taking in DPU, and it's actually going to, we're actually going to build a diabetes model straight there. I'll press pause again. And what you'll see is we see some kind of impressive results. This is just a baseline model. But the key thing here is how quick it is to actually build these models. Once we've got this baseline model, DPU already bakes into the AI platform. Now what we'll do is quickly train another model. In this case, we want to know about another disease. In this case, we want to know about hypertension risk. And so this just goes to show you just how quick it is to develop machine learning models on this platform, while also leveraging the power of this foundational model, which I call DPU. What we see is actually when we've removed that model from this, we see massive drops in performance. So this clever piece of technology gives us a massive leg up in terms of the accuracy and the precision of the results which our AI models drive. So if I click forward now to the next slide, there's going to be not just code, but a little bit of fancy animation as well. Next slide. Okay. Oops. Here we are. So here you can see an illustration of one of those trained models, which is part of this, which is part of Health IQ, which is our growing library of disease and risk prediction models. And here you can see a single member, which is being passed through one of those abstract disease-specific models, which I showed you, giving you a specific recommendation, in this case, classifying that member as high risk. But as I mentioned, this is rather abstract. What we want to do now is to look under the hood. So if I go to the next slide and play this demo, I'm going to pause here as well in certain cases. So if we run the model, or if we run the demo, rather, what we see here is one of our Health IQ models. In this case, this is our 12-month risk prediction model to identify the future risk of a patient admitting into the emergency room. And again, this model leverages DPU. Just pause it. And so what I'm doing here on the left is loading in 180,000 members under our care. So I just want to stress, be very explicit and stress that these are synthetic IDs. There's no kind of patient information in this demo at all. So if I continue with this and we wait for the execute button to click, what you'll see flowing through on the right is a series of predictions for every member. All of those predictions are going to stream back in FHIR format straight back into the health graph, which ensures that those results can actually be leveraged by other applications within Babylon, whether that's the member experience, the clinical portal, and so on and so forth. Just pause it there. Okay. So on the left, you can see some of the results of these predictions. You can see the predicted probability. In this case, it's a very low predicted probability at 7%. And what we're also able to do, because of the explainability of our models, is able to identify the key risk factors which are driving that particular risk prediction. So if we take a closer look at one of the patients, so I'll just wait for it to run through. So here we have an example of another one of our members. Here we can see this member has a risk of around 24%. And again, because of the explainability of the models, we're actually able to identify the key drivers for the member's risk. In this case, we can see the member has a history of hypertension. We can see they have a diagnosis of type 2 diabetes, and we can also see that they're a previous smoker. Going further, again, leveraging the explainability features of our AI models, we're able to actually provide a recommendation. That's actually what's going to be shown next. These are relatively simple recommendations for now, but they'll get much more complex. These types of recommendations can be provided to inform the member's care plan. Those care plans can be developed either internally by our own team at Babylon, but we're also integrating third-party care plans from other providers as well. Sometimes what I've just kind of talked about here is that sometimes we kind of present these results from back-end systems, and it's not always immediately obvious what's going on. It's really important to stress here that the results here are from predictive models that allow us to understand the future risk that members have about admitting into the emergency room. And once we have this kind of predictive measure of risk, this provides us with a means of assigning proactive interventions. And the reason I stress that is because the majority of other approaches, so-called threshold-based risk stratification approaches, which are used widely across the industry, only provide a view of the historic risk. And that's a really key distinction to understand. What we're developing here is predictive models to allow us to actually understand how to intervene early. And what I should say also is the flexibility of these approaches also allows us to incorporate additional population-specific data sets, such as social determinants of health. Again, the types of risk stratification approaches typically considered, which are those threshold-based approaches, do not necessarily allow for the integration of those data types. Okay. So I'm going to skip ahead. Let me let that demo play through. Okay. So again, going to skip. Okay. So just another illustration of kind of painting a picture of what's happening on the back end once we're running these models. That previous demo actually showed a 12-month risk prediction deployed on the AI platform. But really, what I showed previously in the previous demos was the beauty of the AI platform. And that is the ability to rapidly build and deploy not just a handful of models, but going into the 20s, 30s, 50s, and potentially hundreds of models, which are delivering these deep predictive insights across our populations based on this kind of real-time updates to the health graph. So that's the overall kind of vision. But it's not just a vision. This is actually how we think about Health IQ. Health IQ, as I mentioned, was this kind of growing library of disease and risk predictions, delivering these kind of near real-time predictive insights based off of streams from the health graph, and what we're able to do is once those predictions are pushed back into the health graph, it's available for downstream products and services to consume, for example, the member experience, which Yuan is going to show you shortly, as well as our clinical decision support systems and also population health management tools. So the strength of the approach is that we're actually able to leverage Health IQ models, not just those developed by our team, our world-class team of AI scientists at Babylon, but also the platform actually supports the integration and development of third-party models as well, such as those from our partners, but as well as from our collaborations with leading research centers, such as those that we have with the University of Oxford, Imperial College London, as well as a growing list of others, such as NYU and Columbia there in New York. So if I go to the next slide, what you'll see is an example of another AI model, this time driving part of the clinical experience. So I'm going to put the volume down first and then just show you what's going on here. So just changing gears now, because previously I showed you a risk model, but now this is an AI model which has been integrated into our product. And so what we see here, if we play through, is a consultation taking place between a patient and a doctor. And what you'll be able to see is that there's an NLP model under the hood providing these near real-time summarizations of this consultation. So if I put the volume up. I'm asking some questions. I can see that you have diabetes. Do you struggle with your blood sugar? Yeah. Based on your data that you've provided, it looks like your blood pressure has been increasing. Have you had any chest pain? No. No chest pain. Has there been any increased stress over the past three months? Yeah. I've been a bit stressed. I'm going to put the volume down. What you'll start to see is the application of this NLP model providing these kind of near real-time summarizations of that consultation taking place between the doctor and the patient. This really goes beyond the simple types of transcription of conversations, which is the approach taken by other vendors. It's able to capture, as you can see, and summarize the dialogue in a way which kind of closely resembles the way in which clinicians actually write out their consultation notes. I'll just play that again while I'm just finishing up here. The results that we see from this kind of initial rollout of this system in the U.K. is actually this reduces documentation time that our clinicians would otherwise have been consumed with by around 23%. It also results in additional data being collected about the consultation with fewer inaccuracies as well. I'm happy to say that actually today, this work is actually being presented at a leading NLP conference, reflecting this kind of rapid translation of research work straight into production systems. That's it for my section. I'm going to hand over, I think, to Steve. Yep. I got it. Hey, thanks a lot, Sara. Yeah. So let me try to quickly summarize what you just saw, because clearly we went into a lot of nerd talk here for you. But to summarize, we've really invested in a platform that allowed us to onboard data scientists and really start to make an impact quickly. I came from a world where I could literally look at almost every single line of a P&L of a $12 billion company and tell you how many models are running against that company, basically running the show. We see that same vision here, as we need a platform that allows us to rapidly iterate and build on basically that platform. And we're trying to do that with custom-based tooling. The second thing you saw is just a massive investment in our DPU, which is basically a unique way to use transformers to create a deep patient understanding model that actually solves for gaps in data and actually makes a more accurate model as we run additional models on top of that one. And then thirdly, you can see the integration of this technology into our platform allows us to do things like note summarization. Note summarization, while speech-to-text, is pretty novel. We actually use multiple third parties, whoever has the best quality. But the complexity is, how do you now summarize that into clinical notes that you can serve up to the clinician to save them time so he or she can either edit, delete, or update on what we may have missed? Those same models can now be applied to basically our intent and understanding as we start to roll out conversational AI experiences, which we believe is really the next step. We really want to push that experience out to the member through that member experience. My colleague Yuan will actually show you kind of where we're going with the product. But with that, I'll turn it over to my colleague Yuan, who's going to walk through some of the things we're building on the member side. Thanks, Steve. All right. I'm going to try to show an example of our app on my phone, and I'm going to try to get it up on the screen. So give me one second while I switch over. That worked. All right. I don't know if we can get rid of that little box. I'm going to move it around. Pretend it's not there, but this example is going to start with our patient, Donna. She's quite ill, and what you're going to see over and over here is our ability to assess patient risk, then to activate them, then to monitor them, and then to intervene over and over and over. Before I get into it, though, this email highlights our new look and feel of our brand. To some of you interested in art, it's inspired by Henri Matisse, who in the late stages of his life got sick and couldn't paint anymore. And so he resorted to cutouts. And in many ways, that metaphor applies to what we're trying to do. We're taking all these fragments of information, fragments of data about our members, and combining it together to form one cohesive picture where we can then assess, activate, monitor, and when appropriate, intervene with the right level of care. And so Donna here received an email. It's super easy. It's personalized to her needs. She can click Activate Account. Eventually, she downloads our app. It dropped her onto our home screen, which is already personalized to her. Click Activate Account. There's a couple of pieces of information we have to collect, make sure it's really her, like her birth date. Some IT challenges. This is the problem with live demos. Let me restart this share. Sorry about that. All right. Let's see if it's broadcasting. There we go. All right. Sorry about that. So I'm going to enter her birth date, July 30th, 1980. Very quickly now, we're collecting all the data available on Donna. We're connecting to appointments records. We're connecting to pharmaceutical records, her previous interactions with doctors, and creating this personal health graph. All of a sudden, she's been activated, and in the background, as Saurabh and Steve showed, we're really risk assessing everything we know about Donna, and so congratulations. Her account's been activated. The first thing that we do, though, is we want to onboard her family. Many of our users, their parents, there's lots of children. We want to make sure we can deliver care to everybody. What we detected in building that health graph is that Donna actually has children. Just as easily, she's able to add them and then lands on our homepage. This beautiful new homepage should be pretty straightforward. We are highlighting her health score. We'll get into that in just one second. There's a rewards section where she can accrue rewards. It's a place where she can easily get care now. That could be both via text, which we will show, video, which we'll show, or telephone. First, let's just go start looking at the data we have on her. Let's click All Data. We're collecting a lot on Donna. Her activity levels, her heart condition, sleep, blood, body. It goes on and on. It's quite the long list. What's important, though, is that our AI already did some work on behalf of Donna. So we think she might be at risk for diabetes, and so we've really personalized this page to really highlight just those pieces of information that really matter to her. Likewise, on My Health, she'll be able to see her lab test results, her prescriptions, and if she has any active referrals, active referrals as well. For each of these measures, she can see how she's doing in terms of red, amber, and green indicators, so she knows exactly what to work on. In fact, we can go into one of them, like active time, look at her activity over a month, over the day, or even over a year. We can set personalized targets or use recommended targets. We want to set her activity, set the target, and then right away, she can see how well she's doing to hit that target. But through the magic of demos, time passes. She gets a notification. Remember, we're constantly assessing her risk. We're constantly monitoring her. And this time around, we notice her glucose levels aren't within normal range. And so we want to intervene. She can jump into our chat, which is led by our Babylon advisor, and right away starts interacting with a bot. She's asked how she's feeling. She said she's feeling okay. And in most cases, this is a false positive. It's a false alert, but we're still monitoring Donna. So there's no action on her part. We want her to continue her medication. We want her to continue her healthy lifestyle choices. In other cases, though, other things may go wrong. So again, this time, Donna receives a notification. "Hi, Donna. Your glucose and blood pressure levels aren't within range. Please speak with your care assistants." Our AI jumps in, connects her with a personal care assistant, and we're escalating the care now when it's most appropriate, and this time a human takes over. She asks how she's doing. Donna complains about not being able to get her blood pressure levels in range, and so we are able now to escalate from an AI to a care assistant to then scheduling an appointment directly from the app. It's super convenient. It's super easy. Donna can click Confirm Appointment. She goes to a confirmation screen. She hits Done. Time passes, and then she can get her video appointment. Now, the outcome of this appointment for Donna's case was a new prescription, and we wanted to make it as easy as possible to get that prescription delivered to her. And so we send her a notification once that prescription runs through our system. She can select Delivery or Pickup. She selects Delivery. We confirm it. She confirms her address and then confirms the prescription. And one of the things we've done here, from a design perspective, we've actually included a photograph of what the medication looks like. Prescription compliance and adherence is complicated. Many of our patients require multiple medications, and it's oftentimes very difficult for them to understand which drug corresponds to which pill. And so we're trying to set that connection early. And then we'll go into adherence in a second. So she confirms the prescription. She has that handy DoorDash-like look and feel, so she knows when it's going to be delivered. Eventually, she'll get a notification with a two-hour window when it will be delivered. And within 24 hours, she had that appointment with her doctor, and she had that medication delivered, and it all happened straight through her phone, so now time passes. Donna opens up her app, and this time, we're asking her not to check out her health score, but asking her if she's taken her medication, so we're trying to gamify this experience, and it's just an example of where that's happening. She can answer yes. She accrues some points, and you'll see here the points counter also went up, and what this does, it starts gamifying the experience. We can then start providing rewards and other incentives, badges as appropriate in certain markets, so that we can really drive that adherence. If we prescribe a drug and people aren't taking the drugs, it's hard to make those drugs work. And so we really want to make it a fun, enjoyable, and convenient experience to train people to develop that habit. But in Donna's case, and for the purpose of this prototype, her blood pressure is still too high. And so she gets this notification, and she decides to start a chat. We show her what her blood pressure is. We ask how she's feeling. And then our AI jumps in and asks her a bunch of other symptoms. So we'll go through this. She's asked if she has a foot pain. The answer is yes. Has she injured her toe recently? I don't know. Have you recently had a fever? Yes. Do you have a headache? Yes. And so, based on all this information, our triage is looking at it and looking at it and making a decision, "Wow, we need to escalate this." And so, right away, it gets escalated to a care assistant. Jane picks up again and asks for some more information and allows Donna then to quickly upload her photograph. And then, Jane, the care assistant, makes a decision to refer her out to a specialist, in this case, a podiatrist. So, very simply, using our network and scheduling technologies, we book an appointment this time with Dr. Mary Smith, who's a podiatrist, and they can determine their next steps. All right. So, because of that podiatrist visit, Donna is recommended to go to the hospital. She can just click a button now from the app to get directions. She can call a ride if necessary. The hospital's chosen based on what her insurance provides, her geography. To try to make it as convenient as possible, we've shared her records to the hospital in advance so the hospital knows that she's coming, and she can quickly jump in, and then after some minor surgery, we refer her to our sister app, Day to Day, which we've just acquired, to help her on her recovery, and from this app, she can monitor her surgical site care, take photos as needed, describe the wound, and eventually, she'll be well on her way. Going back to our app, though, time has passed again, and this time, congratulations, the homepage has changed again, and we're showing how, for once, we can monitor and update our risk assessment for Donna at all times. In this case, her blood pressure levels are within range, and we're congratulating her on taking her medication. She went from 67 to 77, but because she has a chronic condition and she has needed a lot of care, what we've decided to do was to assemble a personal care team for her. Now, on My Health, you can see that her metrics are now in a much better place. Her blood pressure is healthy. Her blood sugar is healthy. She's well on her way to recovery, so time passes again, and again, we are monitoring her so we can intervene more appropriately. In this case, her mood score is low. Her active time has gone down. We recommend that she speaks with a therapist. So I'm going to click Book Appointment. This will be the last piece of our demo. And now, our AI is recommending she speaks to a therapist and knows what time Donna is most likely to be available based on all the previous appointments she's had. And with that, I'll wrap it up. Charlie, I think you're up. Yep. So thanks very much, Yon, Steve, Saurabh, and the rest of the team. We've got a short break now. There'll just be a 10-minute break, and then we'll come back for the super exciting part with Darshak and also in conversation with one of our doctors, Misty. Sorry. And then on to the financials. Thank you very much. Great. So thanks, everybody online, for remaining with us. And I hope everybody else had a good break. And without further ado, I'll pass over to Darshak. Thanks. Thanks. Everybody, so I'm going to take some time to walk you through our clinical model. Just step over to the side here, so I'm going to block the view of you. So we've had a little bit of a view too into our technology, our AI, and our product. One of the things I'm fond of saying is that healthcare is often something which is, there's really a human element to it. And so what we've talked about is really what I would refer to as the anatomy of care. Now, what I want to talk about is what is that physiology of care? In other words, we've got all the architecture, you've got all the tools. How does that really all come together in order to take better care of our patients and ultimately lead to value? What I'll talk about here is I'm going to give you a little bit of a sense of where we're going. I'm going to spend some time giving you a little more context to what our clinical model is. In other words, how does this, everything we've taken care of, how does that come together in the care of our patients? I'll then dive in a little deeper because one of the questions that's often asked is, "Look, we love the technology, the product looks good, we get the clinical model. How is that actually going to come together to reduce the total cost of care? And what is your strategy to actually create margin, particularly in some of the markets in which we're at?" I'm going to give you a peek under the hood at our broader strategy and how it's all going to come together. That's a critical part of sort of the underlying thinking and the progress we're making towards getting to margin this year, and then finally, I'm going to be joined by my colleague, Dr. Misty Zelk, who is one of our practitioners, and I think we're going to be talking about real cases in an anonymized way, of course, that'll give you a little bit of a sense of what does this look like actually on the ground as well, and then we'll move on from there, so let me take a few minutes to walk you through what I like to call a clinical model, so we'll go ahead and go to the next one here. So you saw a little bit of this picture with Ali, but I think it's important to sort of say, "Well, how will this actually operate?" So think about the last time you saw a physician or had a healthcare interaction of some type. It might have been a primary care visit. It might have been a time when you suddenly had an injury and needed an urgent care visit, or it's possible you even had a hospitalization you wanted to deal with. And think about how digital tools might or might not have helped in that situation. At Babylon, our motto is affordable and accessible care. In other words, we want that care to be at the right place, at the right time, and the right quality. So at every step of the way, what we've done at Babylon is taken the healthcare system through experience. Recall that Babylon has been doing digital health since 2013. It is very important to keep in mind the amount of learning that has occurred over that time. In other words, digital health built from the ground up, you can't just take a system that currently operates and suddenly just give them tools and output that and expect that to change right away. It's a process that occurs over time. How does one practice efficiently in a digital-first environment? How do you know what you can actually address just by talking to patients, when do they have to come in? That kind of experience, again, it's been learned over the past seven or eight years. That information has been codified so that our physicians have learned to practice as a group. They've standardized those practices. You probably heard that when new innovations came out, you heard Ali talk about the 11,000 papers in dermatology. On average, it takes 17 years for new technologies to sort of diffuse fully out. We at Babylon, because of that digital-first nature, as well as the ways in which we get information spread and standardized, have cut that cycle time dramatically. So when we talk about the pyramid of care, what we're always trying to do is to take care that's being delivered in highly sort of artisanal settings, so to speak, which is how healthcare has always operated, and bring that down to make that more and more accessible or more digital and ultimately something that can be self-served. So what that means in the digital age is now, suppose you have abdominal pain, a headache, we actually have an AI-powered symptom checker. You sort of go through it. It gives you a sense of maybe what are the potential diagnoses one could have and even recommends that next level of care. That's digital self-service. We've also, in the past, worked with tools that actually perform and actually start your visit even before you see a clinician. That kind of digital self-care is what we're all striving towards. That's how we get to scale. And then when there's a signal which is seen, which is sort of more complex, we sort of escalate. The personal care assistant, we talked about that. And then when the personal care assistant needs additional help, and you saw Yon work through that with the demo, we get to higher and higher levels of care. But behind that is the thinking that at every step of the way, our teams are always analyzing those processes and saying, "How do we make them more digital? How do we push them down and make them more scalable and more on the self-serve continuum?" And I'll show how that occurs with a couple of concrete examples. But this is our care model. Now, what does that actually look like? So Ali went through a little bit of sort of our virtuous care cycle here. But I want to give a couple of concrete examples of what that actually looks like today. Let's take the personal health graph. That actually pulls in all of your data. But what does that actually mean? Well, that means that in that initial health screening, it might ask you questions like, for example, it'll perform a screening of a PHQ-2 and PHQ-9, meaning it screens you if you're appropriate for the presence of anxiety or potentially depression. We'll find out if you have high blood pressure as well. Now, that information generates a real-time insight. You saw the AI-powered engine. So that insight can be something very simple, which is, "We're concerned you might have anxiety." Or it could be more complex to say, for example, "You've been on an ACE inhibitor, and your potassium level is a little bit low, and you're at risk of a cardiac rhythm disturbance." Those are the types of insights that are surfaced, and then they go through and power that health goal and care plan. So that assistant may reach out and say, "For example, you are at risk of anxiety. You're engaged in your care right now. We're actually going to start you on a digital anxiety care pathway." That's actually something we have that's actually operational right now. If one is at higher risk, on the other hand, it's going to shuttle you into a different appropriate care plan, which is an actual set of virtual visits with our therapists. One of the things I'm particularly proud of is that at Babylon, we place equal importance on mental health as well as physical health. In fact, fully half of our visits that are done by virtual care providers today involve some kind of behavioral health because we recognize that that is an equivalent part of healthcare, and it's all powering the data and the care journey that's going along. We can go along this route, and I'll give a couple of demonstrations of that. So when that signal comes through, we say there's remote patient monitoring. So through Bluetooth enablement, we can then collect, for example, blood pressures, glucose readings. But proactive health and remote patient monitoring isn't simply those kinds of devices. We can then push out surveys. For example, after you start that care journey for, say, anxiety, over the right period of time, we'll rescreen you to make sure that that anxiety is getting better. If it's not, the appropriate escalation pathway can take place, either with a specialty visit. For example, we have a national network of asynchronous specialists. In other words, if you actually have a condition that requires more intensive medication management, dermatologist input, endocrinologist input, psychiatry input, we can perform today asynchronous visits with specialists to make sure that we get that care back and inform your care at that point. So what I'm doing is building the case that this is data, it's technology, but it's powering a comprehensive series of complete primary and secondary care throughout that full continuum of care. It's a critical, critical part of our care model. You see, we then go through hospitalization support. We can currently, through ADT fees or admission/discharge transfer fees at hospitals, determine when Babylon patients are admitted to the hospital and ensure that they then get the support when they leave through daily monitoring through their application within care manager and care assistant support as one of the many types of examples. We also, for example, partner. One of the big areas of interest in the markets we serve in Medicaid is newborn intensive care. Through partnerships, we actually provide evidence-based support and care for NICU hospitalizations through a partner that we have there as well. And then moving through there, I'll talk a little bit about our chronic condition management as well. The point I'd like to. Oops, I've lost my clicker here. It is not. Oh my goodness. That was an amazing magic trick. Whoever has the computer, could you just advance the slide there? Thanks. It's hard. It's funny. I can do surgery on little kids' hearts, and I can't keep track of a clicker up here. It's a different skill set, I guess. The entire care journey I talked about and that broader care model, the goal here is to demonstrate that it is comprehensive and it provides an enormous suite of services that are coordinated by this digital brain and digital backbone. But at the heart of it also, I think I'd like to emphasize is that there is people-based support. Ultimately, care at the right place at the right time, what it does, it allows us to best use extremely specialized resources in a much more thoughtful way. So the thing I want to emphasize here is that those resources are all there. You heard about our specialized care teams for our most complex patients. This is part of the team that our patients benefit from: mental health specialists, nursing, dieticians, primary care, and of course, our 24/7 care advisors. Again, this envelops our patients in that comprehensive care program that then escalates or de-escalates intensiveness depending on their condition. Now, what does this look like when we put it all together? So this is an example of one of our real patients. I sort of walked you through this journey. One of the reasons I want to do this is that in contrast to some sort of avatars that you might see walk through, patients on Medicaid and those with complex conditions, particularly those at the highest risk, they are challenging in terms of the number of medical issues they have and also the complexity and the unpredictable nature of the course of that clinical condition. So I want to give one example of a patient that we've taken care of. I want to emphasize we're also going to spend some time talking with Dr. Zelk and reviewing sort of real-world experience of at least four additional patients. But I want to put this all together. So we're going to call this person Amy, but please do know this is based on a real person. Amy's not the real name. Lives in the rural Midwest. This person is in her late 20s. So first thing is that accessible care is a massive innovation. Just think about the last time any of you needed a clinician. How easy was it to actually see somebody right away or make an appointment, particularly if you're looking for a primary care physician? Did you actually do it right away, or did you kind of wait for a little while? As Ali pointed out, think about in Rwanda. We made care accessible even in that population for less than $1 a day. We made care accessible in Missouri to this population. These are individuals, by the way, who lived in the Bootheel. You saw a video of one of them. These are individuals who actually had no access to care in many cases. They'd have to drive an hour away. This woman, actually, even before she was placed in this program, had access to Babylon and used it for just a simple concern like a cold or a cough. But love the idea about digital healthcare. This is a little bit of detail. We were originally just offering our services part of a, "If you need us, call us sometime," a fee-for-service arrangement. However, we later, thankfully, and this is what's truly exciting about Babylon, entered into a full-risk contract. This is so important to understand in the sense that our financial incentives now, we're accountable for the total cost of care. Roughly 20,000 patients in Missouri were assigned to us, meaning we had to find these individuals, offer them care, and then ultimately, if their total cost of care was lower, we benefited. Many of you are familiar with this type of arrangement, but full risk, in my view, is the engine of innovation because it allows us to invest in the longitudinal care to actually take care of these members. So when this person came into a full-risk contract, we reached out to them. They knew who Babylon was, but they hadn't sort of engaged us proactively. This is what's so critical, is that we got to know them ahead of time. So we did an initial digital assessment and found that this woman actually had chronic low back pain and also complained of some chronic gynecologic pain, which she felt had been caused by prior childbirth. She had not mentioned this at any of her prior urgent care visits. So again, showing the importance of that digital proactive outreach and making that care accessible and easy to use. She was booked with her virtual PCP. So that's an in-app sort of app experience. And she talked to that person, and this is what also makes a big difference, is talking those things through unlocked all these other things that were on her mind. She had PCOS, which is polycystic ovary syndrome, some food sensitivity she didn't understand, and then this other thing. She actually mentioned, "You know what? I actually think I've had my thyroid checked, but I never checked on those lab results." It's really, really remarkable. And she mentioned that she had had childhood trauma. Now, none of that came to anybody's attention before. Even though she technically had health insurance, she technically had sort of care of some type, but that's what a digital model can do. That's what proactive engagement can do, even in a Medicaid population that's highly remote from care. That health assessment, that virtual PCP, she built up a sense of trust. Again, in addition to full risk, the other thing that we do at Babylon is we develop a relationship with these patients. We see them over long periods of time. We're actually moving towards a fully employed primary care workforce. That means the kinds of physicians we hire and train are those that truly practice digital-first medicine. I want to emphasize how important that is. To take one example, in Kaiser Permanente, which is another value-based care system, it takes them four to five years, by their estimates, to take a physician and train them in the Kaiser way. They're so challenged by that that they actually opened their own medical school recently. That's how hard it is to find individuals that actually practice in this way, and those are the individuals that we're bringing into our system. I want to emphasize that's important because when they reach out, they know how to build relationships. They know how to win trust and so that people even talk about these kinds of things with them. You'll see that when you speak with Misty as well. So then what happened over time? Impressively, through the health graph, we're able to then consolidate those records together. We found that she had had a thyroid issue that hadn't been treated. What does hypothyroidism cause? Many of you probably aware. Low mood, problems with energy, menstrual irregularity. So this was a key part of something that had been missed in this individual over time. Again, how was that treated? Home delivery of medication could occur. Then look at what else happened. Long-term care. She was able to get into a virtual care program for treatment of anxiety, met with a virtual dietician, and had a virtual specialist GYN visit. This is what digital health looks like when it's fully enabled in its form. That's the kind of care we can deliver. Then look at what monitoring data does over time. We push out reassessments. We make sure that we quantify and monitor her mental health over time. Again, this gets beyond, I challenge many of you to think about when we talk about monitoring, what does that mean? It's not just the same old blood pressure and glucose, but it's actually evidence-based key types of pieces of data like the GAD-7 and PHQ-9. I'll say one last thing, which is that it actually turned out, again, through a health graph, we learned that she had had an abnormal Pap smear that had never been followed up on as well. When we think about what does preventive care look like, that's prevention. Finding out that somebody had a test that was abnormal, it turns out that's generalizable. Some of you are probably aware that a substantial number of individuals that actually undergo cancer screening have an abnormal result and never follow up on it. This is the kind of work that we can do at Babylon, again, because we have that longitudinal relationship. So I hope that's a little bit helpful in painting that picture, and we'll come back to that in a little more detail with Misty. I want to spend some time now saying, "Okay, that's wonderful, right? You saw our, like I said, you saw the product, you saw the technology, you heard the high-level vision, and now I've talked to you a little bit about our care model and this notion of what truly digitally enabled care can look like and how we're constantly innovating to make it more and more digital to make it scalable. "How on earth is this going to lead to margin?" That's what I'm going to talk about now. So this is our clinical playbook. And when I talk about margin, I want to give a little bit of sense of what that means. We are proud to now have some of the really most talented individuals in U.S. healthcare that are now working with us here. I'm referring to my colleague, Dr. Nirav Vakharia, who's our U.S. General manager, formerly the head of value-based care at Cleveland Clinic, joined by a number of other extremely skilled clinical operators and clinical and healthcare experts that have joined our company recently. Those individuals are working tirelessly in order for us to make sure that we achieve our margin frameworks. What I want to say is I'm going to go through the margin framework here, but I want you to also be aware that these are not just high-level concepts. We have mapped the potential savings in each of these areas in a data-driven way, laid out a specific set of metrics that we are tracking regularly at monthly and even weekly business report levels, and we believe that as those track, that's what's going to get us to our margin goals. So we are taking this very seriously and we're doing it in a very quantitative way. So these are our six levers. And I want to emphasize a little bit about why we at Babylon feel that we're going to succeed and a little bit also what distinguishes us as a digital-first provider from other organizations you might know. So first is on the left-hand side, which are provider-driven programs. In other words, this is when you proactively, you as a patient, reach out for care. This is when people think about telehealth, which is a word we don't like. We use digitally enabled health. That's what we do. But this is standard telehealth, meaning the first one is, "Yep, we just reach out. I've got a little cold, ear infection. I'll call up. The provider I see has sort of Uber-like doctors that are just kind of working on the side a couple of hours a week." That's typical urgent care. That is the first pillar, but we is way beyond that. When we talk about 24/7 primary care, I want to emphasize that this is longitudinal care with people that are trained in digital-first methods. This is critical. This is what we spent the past seven years honing. Ali, and you heard him reference this, so I want to point out this very important data point. In the NHS, this 24/7 care alone significantly increased the number of touchpoints that patients had over a period of time. At the same cost, I may add. Simply making that care more accessible through the protocols, all of that care and that highly skilled primary care digital workforce reduced the total cost of care by anywhere from 15%-35%. That was a peer-reviewed paper published last year, by the way. What that means is it's simply offering 24/7 primary care. I want to emphasize, this is primary care that's digitally born and enabled, very different than the typical episodic PM primary care offered by some of our competitors. That is associated with reduction in total cost of care. In addition to that, as I said, integrated behavioral health. This is the second pillar. Equally important to the care of the body is the care of the mind. Many of you are aware of some of the challenges around mental health during the pandemic. We have grown to be a provider that actually, on the same platform, again, this is another Babylon advantage. We don't send them off to multiple vendors on multiple platforms. This is a true team-based effort. The behavioral health specialists operate side by side or screen by screen with our primary care providers and actually provide that integrated medical behavioral care. That is really critical. Some of you are familiar with collaborative care models. Significant evidence showing reduction in total cost of care. In other words, when you treat somebody's anxiety or depression, it has the effect of reducing their total cost of care. I'll say something else on that. I used to work for large payers. Typically, they subcapitate their mental health and have other organizations handle that. The P&Ls are separate. In other words, the capitated mental health provider wins simply by cutting down the cost of mental healthcare. That is not the way you want to treat mental health. Sometimes you have to spend more on mental health because you get the reduction in total cost of care on medical costs. We eliminate that disparity here. Just want to point that out. These are sort of wonky things, but they're important structural features that allow us to align the incentives to improve total cost of care. And the third piece is the specialty offerings. I just want to give one example. It's a personal one, but I have a son who has celiac disease. He had a 15-minute, I'm sorry, 20-minute telehealth visit with a local specialty provider at a fancy hospital where I live. The actual billed amount to me was $720. Now think about that same care is provided now by our asynchronous Babylon specialists, and that cost is tucked into the, there's no charge for that. That is the ways in which simply these sort of parts on the left spine can drive value. But that's not it. I want to again emphasize each of these. We have set specific targets. For example, if we engage this many patients in 24/7 primary care, we believe it will translate to this much savings. And we use that to build up and stack to get to that margin target. The same for behavioral health and specialty offerings. Again, it's a very quantitative, very thoughtful framework. But that's not it. Again, many of you in clinical medicine realize that an enormous amount of work is actually done in value-based systems when you're not actually face-to-face with a patient. This is work that's done offline. If you've ever talked to clinicians who manage complex patients, this is the work they do at nighttime. I'll give one example. In Missouri right now, you're probably all aware, as across the country, there's a shortage of baby formula. If there's a shortage of baby formula, who does that impact the most in terms of total cost of care? If a baby has, say, a milk protein allergy, they're often treated with highly specialized formulas, and those are also in short supply. If you go see a pediatrician at a regular fee-for-service arrangement, those individuals, nobody is watching and calling and proactively notifying individuals about those kinds of shortages. We can go through that. That's just one example. We can go through 10 or 15 things like that down the list of signals that we can be aware of, either through our AI technology, through our clinical information, but through our data architecture. That's what chronic condition management means. It is being aware of all of the conditions, and I'll be talking through that in a little more detail, that could actually be impacting total cost of care and quality of life and continuously surveilling that and then making sure that we put them on digital pathways and digital care journeys to address them. I'll show a couple of ways of how we think about that. That's what real digital-enabled population health looks like. We then talked about episodic management. This is actually in hospital, recent before or out of hospital. Think about it this way. How many procedures are canceled because patients accidentally eat before they're supposed to go to the hospital? Or they forget to discontinue the aspirin that increases bleeding risk. Last time you had a procedure surgery, think about were people really kind of helping you, guide you along the way? We have programs in dozens of surgical specialty areas that guide them before and after those episodes. I mentioned our NICU elective program as well, and then finally, our health assistants that actually can reach out and be available, contact people, make sure that the things you're supposed to do, preventive care, are actually addressed through, again, primarily digitally-enabled footprint. What I'm emphasizing here is that this six-part plan includes not only the traditional sort of clinical-driven programs, but in our B360 programs, this is how we create value. It's digitizing this and making sure it's clinically enabled, and there's a clear clinical strategy behind it. Again, emphasizing, there's a quantitative piece with each of these. We know exactly how far we have to go to get to the right margin target. This is the formula we use. So we want to think about, again, what some of our—what we do. I just went through all the things we do on the claim savings side. We're not risk-adjusting our way to value here. We're not just coding and documenting. We're actually making people's health better. And that's all the six things that do that. And I talked about the rigorous way in which you do it that's clinically informed. That doesn't mean we're totally naive. We want to make sure we're doing all the things that we should do to be good stewards. If patients have more complex conditions, we want to be sure about that. So we obviously have our gap closure program and risk documentation programs as well. But that, I just want to emphasize, this is one part of a much larger ecosystem of care. This is one of the reasons I'm proud to work at Babylon. We are about, and most fundamentally, about improving the quality of care in people's lives. Then we also, as Ali pointed out, and this is the thing I'm always struck by, I just want to emphasize a stat that he said. Just want to make sure this is sort of people heard this. The primary care capitation rate in the U.K., a fully developed country, was, Ali, what was it? $150 per member per year. That is optimizing a skill mix and cost of care delivery. We use that same way of thinking to make sure that we eliminate waste. We improve and use our talent to the best of their ability, and we take care of those resources well. Again, you heard about the data. So this gives a little bit of sense of just the comprehensive nature that we use here. I'm going to try to accelerate just through the last little bit so that we can get to talk to Dr. Zelk. I'm excited to do that. When we talk about data, and this is just a little bit of a deep dive just to show, again, just as you saw the demo that Steve and Sarah put together to show the complexity and the depth of what we're doing in AI, this is some of the, gives you a sense, again, the complexity and depth we use when we think about our total cost of care. This is real data from our populations. And I've restricted this for now just to our Medicaid total cost of care businesses. But you can see here we have used our expert analytics team using all the data we've gathered, sorted by chronic conditions using a well-validated algorithm. And we see, first of all, what is the% of spend in that area, as well as the% of our members. And this helps us prioritize and strategize for the specific programs in which we invest. What do we prioritize, in other words? If you're going to do three or four chronic condition programs and really spend time, how do we figure that out? This is how we do it. See, our number one expense, particularly in this population, is childbirth, then masked claims. Masked means that those are almost always due to sexually transmitted infections and conditions with either alcohol use disorder or opioid use disorder, anxiety, hypertension, depression. It gives a sense of what the spend is, the number of members, and then we add in impactability. And that helps us arrive at how we're going to make that pathway to margin in some detail. Childbirth is one example in particular. Let me just give you a sense of how our team thinks about this area. Yes. Oh, again, time to move on. All right. Well, over drinks. Love to talk to you about that. Or any of these areas. So again, we use that to sort of think about, well, where do we focus in those different areas? I'll let that go. But I will say that one of the things we're particularly proud of is our chronic condition program. In 11 specific areas, we're deploying a number of strategies to monitor and impact that course of that condition. Again, thinking about our markets and then usually selectively dialing up or down the amount of effort we put into that. Again, all in the service of using all this fancy technology to get to that all-important thing, which is we have to improve margin, which I read as better making people's lives better. So with that, this is a list of some of the conditions we'll be looking at. Again, I'm showing this to you just to, again, in the service of completeness. I believe you all have access to these slides. Again, if at any time you want to talk about these, but I want to just give you a sense of, again, we are doing all of this this year. All of these programs are going to be put together and be made available to our members. This is our goal for 2022. It shows you, again, the level of complexity that we're tracking over time, clinically informed. So with that, I'm going to move on from this. This is another example I was going to show you. Why don't I move on to our conversation with Dr. Zelk? Because I think sometimes speaking with one of our clinicians that uses a number of these tools to impact the lives of the members we serve, again, brings to life exactly how our clinical model works on the ground. So I'm just going to ask, is Dr. Zelk available on the phone or by Zoom? While we're waiting for Dr. Zelk to come, I'm just going to tell you a little bit about her. Oh. Hey, Misty. Can you hear us? Hi, Darshak. I can hear you. Oh, great. Well, thank you very much for joining us today. We've got a terrific group here. I hope it's okay. I'm just going to brag a little bit about your background before we start talking. But Dr. Zelk is one of our Babylon clinicians, and she really embodies the characteristics and the passion that our clinicians have in taking care of the members we serve. Dr. Zelk attended medical school in Arkansas. She then did a residency program in both medicine and pediatrics and is board-certified in medicine. She's also served for over 20 years as an officer in the U.S. National Guard, having served in the Middle East, as well as in a number of humanitarian missions worldwide. She currently lives in California and practices on our platform as one of our primary care physicians. And I've had the privilege of spending some time with Dr. Zelk to talk about some of the patients that she's seen. We've chosen maybe four scenarios. Again, these are all real patients that Dr. Zelk has cared for to show different parts of our model. Misty, I hope it's okay if I start with the first one, which is you mentioned that you had taken care of one patient who works nights. And as a result, it's challenging for that person to get appropriate healthcare. So I thought this really touches on the theme of accessibility as being a critical service we can offer, particularly one that's digital. Could you tell us a little bit about that person and a little bit about how you've helped that person on their journey? Certainly. So yeah, so this patient found me promptly at 8:00 A.M. one day. It was her first time visiting Babylon. And she had quite a laundry list of issues that had built up because she worked overnights. And so she had a physician, but she hadn't seen that physician for years. She would utilize urgent care because she was sleeping during the day and couldn't get in for visits. And so we sorted out the laundry list and prioritized it. And currently, we've done 14 visits together over the past two years, just working through those various issues and picking up on some things that had been dropped over time, thyroid issues, for example, started some new medications, optimized them, but developed a relationship and really got to work through some issues that had built up over time. And the epiphany that really struck me while I was working with this person was that this was the person, this was the type of patient in the clinic that would just frustrate you sometimes because they had issues. You wanted to touch base with them periodically. You wanted to go over issues, do some preventive maintenance, and they just never made appointments, never came in, and so like this patient, I've found there's this whole subset that are working multiple jobs or working graveyard shifts, and this is the perfect solution for them, and I've gotten to meet several of them and become their primary care doc through this platform. Great. That's a wonderful story, and I think, in particular, one of the stats that stands out is that this individual has seen you 14 times, and you've been using that time to gradually, through these multiple meetings, develop not only trust, but then help encourage those next best actions that then keep this individual healthy so they're not using urgent care visits. Terrific story. I wanted to switch gears to the next patient that we chatted about. I think what this gets at is the importance of accessible care that's also compassionate and accepting. Again, digital services, one of the wonderful things about them is that they can help you get care from individuals that are non-judgmental, potentially sort of don't come with prejudices as well. Could you talk a little bit about this individual who, I believe, is a patient who had been transitioning? Right. So yes, I met this patient. It's always a priority, and it's something that we stress on Babylon to find out their chosen name, to find out how a person wants to be addressed. It's a very simple human courtesy. Just from that start, building a relationship and trust, she was pleasantly surprised and actually tearful that I wanted to know her chosen name and then called her by that, that she had had some rough experiences in-person care because she was earlier in transitioning and appearances were somewhat mixed for whatever reason people would reference back to a male pronoun. Just being able to start out that relationship on the right foot and to get her to talk about her issues. We started to have a number of visits, and I would see her from time to time. Then there was a gap where I didn't see her for a good month or so, and then saw her name on my schedule, which was great because I was just thinking about her and thinking about reaching out to her, and found out that she had actually fled her home. Turns out she was a victim of domestic abuse and had gone to a different part of the state and was having trouble accessing care and resources, but knew that she could come back to Babylon and come back to me. So I was able to direct her to appropriate support services. The battered women's shelters in that area were not taking her. I was able to get her support and resources, a place to stay, connect her to appropriate support services for this situation. That was particularly gratifying. Great. I love that story because it's really about person-centered care, but also, again, finding ways to use digital resources at the point of care to help get them to something which will really make their lives better. And in this case, it wasn't necessarily a medication or prescription, but really a place you could go to be safe. Again, an incredible demonstration of the power of digital care. Switching gears a little bit, we were just talking a little bit about chronic condition management and the importance of evidence-based care. And I know one of the things we as clinicians know is that many individuals with asthma don't get the care they need. So I was hoping you could also talk a little bit now about the individual you mentioned who had asthma and a little bit about that person's journey as well. Oh, absolutely. Right. So there was a young gentleman in his 30s scheduled an appointment for a medication refill, which is a common way to get introduced into chronic management, and he just wanted his albuterol refilled, and that's in person, that happens all the time. Go into urgent cares, etc., and just get your refill, and so I asked him. I took a deeper dive into the asthma and asked him how long he'd been an asthmatic, how often he had to use the albuterol, and albuterol is a rescue medicine. You're only supposed to use it as needed, and if you're using it more than two times a week, there's a problem, so he was using it multiple times daily, talked to him about preventive maintenance and his triggers, and he'd never had that conversation before. And so we went through a few visits to try him on some maintenance meds to identify what those triggers were. Now I see him basically when he just needs a refill of those maintenance meds, hasn't used the albuterol now, even in the springtime, which in California can really flare up some nice asthma attacks, no visits. And the best thing was at the last visit, he told me that this is the best his breathing has been in his entire life. And that just made it all so gratifying and worth it to ask those questions. Yeah. I love that story because it gets back to some of the things we talked about here, which is just saying, "Look, we want a welcome visit or just come on in," wasn't what this person needed. But it was, they reached out because there was a service we could offer, convenient refill of a medication. We could have just stopped there, but that's not how we operate. We want to make sure we understand why they need that refill, get them into longitudinal care, and ultimately even potentially digital self-service tools to track their asthma symptoms over time. So love that one. I think we have time for one more scenario, and this is one that I was particularly struck by, was when patients need care and they feel like they need a second opinion or feel like they haven't gotten what they need from the clinicians they have. They sometimes reach out to us, and we can use that as an opportunity to sort of help them along and almost be a wraparound service or provider enablement. This is a really kind of a frightening but really deeply moving case. Talk a little about what this individual had. I think they came to you with either it was chest pain or back pain. Yeah. This was a particularly poignant interaction. An individual new to her insurance. She's in a remote area of California near Yosemite, not a lot of services. She gets assigned a PCP by her insurance, but she can't get in for a while. In the meantime, she's having horrible back pain and went into the local emergency room, and they see a lot of back pain. No doubt about it. It can be a common vehicle for asking for narcotics. Got sent home on some NSAIDs, went back a second time because the pain was just so severe. Again, got sent home on NSAIDs and no testing or further evaluation. And so she's kind of stuck in a rock and a hard place. She's not getting an evaluation in the ER. She can't get into the PCP, very limited resources. And then she realizes that she has access to Babylon. And so came on for a visit for this back pain. And the unusual part was that the back pain was thoracic, mid-back as opposed to lower back. And that's not a typical storyline for drug seeking or just hurting yourself lifting something. And so because the history was different than typical, ordered X-rays and did some preliminary lab. I was thinking about autoimmune disorders and various other things off of the normal. And then the X-ray came back, and the lab came back, and it was floridly abnormal. The X-ray had what radiologists like to describe as a moth-eaten vertebrae. And that is the classic sign for a type of cancer known as multiple myeloma. And this patient is only 45. So once I got those results, got the assigned PCP on the phone, was a nurse practitioner in this remote clinic working for a health center, government-funded, and explained the findings to the nurse practitioner and what these meant. The nurse practitioner had never seen a case of multiple myeloma before. And so I took the time and just walked through next steps, pain control. This person was in some severe pain and how to take care of it, resources, the closest institution that could handle this type of cancer, but just coordinating that care got that patient admitted same day, got her the pain relief that she was long overdue for, and got her into appropriate treatment. And yeah, that one was significant. That one made me cry because I understood that moment, how much pain she was in for weeks and trying to get care and being so frustrated. And that one will always stick with me. Yeah. Yeah. And the things I'll call out about that one, obviously, in addition to the compassion that you showed her, was the ability to make sure that she got you picked up using sort of this unusual type of pain, were able to get her the testing she needed in a way, even through a digital-first approach. And not only that, then communicate back to the referring team, building a relationship and that trust with Babylon, and then guiding them to that next site of much more specialized care. So I'll close with that story. Just think about that care pyramid in our model. This sort of epitomizes that, is that we work all up and down that model. Again, always strive to make that care more accessible and more affordable and higher quality. So Misty, thank you so much for your time. I think we'll move on to our next segment here. I believe that's Charlie. Sorry. I didn't lose the clicker. Thanks, Josh, and saving the best bit for last, obviously, around the financials. So look, Ali covered it a little bit earlier, but just sort of really want to go over. See if I can click forwards. There we go. I saw something move. No. Okay. Yeah, so as Ali pointed out earlier, we have three revenue lines. And these have been the same for the last few years, albeit value-based care is very much the newest for Babylon. The reason why that's important, I think, is when you see some of the results we've had, I think it's really important to remember that that's in the context of effectively operating in value-based care for only a period of 18 months. There's also, as those of you who've been in healthcare for a while know, in particular value-based care. It takes a while to get some of the data through as well from the payers. So when we sort of think about the data results that we've got, actually, a lot of the contracts have been running for under a year, which actually means we've only got a few months' worth of data. So therefore, what we also do is use a lot of leading indicators, for example, engagement and the level of contacts that we have with patients in order to assess future operational expectations. At the top here, though, as Ali mentioned, we do software licensing. We've been doing this for a number of years now, extremely successfully with leading global players. But however, the problem with this is that while the contracts in themselves can be quite large and at great margins as well, you basically only get a few dollars per patient. And the way that we're thinking about the business over the next few years, and I'll come on to exactly why we're doing that, is really actually about growing scale in order to cover our OpEx costs. We have huge levels of scalability in this business. We've shown that we can do that quarter on quarter. And therefore, really, that's sort of where the value-based care up to Babylon 360 at the bottom here comes in. As we've articulated, I think sort of many times before, seen massive, massive revenue growth. And I think sort of also contextualizing this a little bit is also important. I think this time last year, we were forecasting $321 million of revenue for 2021. And actually, for those of you who we spoke to this time last year, you'll remember that most of the discussion actually went into our ability to grow and also deliver on that revenue. And I think sort of a lot of people saw that growing revenue around 4x from 2020 to 2021 was going to be very, very difficult. And we achieved that and exceeded that expectation as well. For 2022, we have already upgraded guidance from where we were originally. So we originally forecasted $710 million for 2022. Again, a huge amount of upside from even the 2021 number. And now the guidance that we put out at our Q2, sorry, our Q1 earnings was over a billion dollars. So again, huge, huge amount of revenue growth. But at the same time, though, that revenue growth needs to be tempered. And that's sort of the reason why we're also seeing a massive amount of patient interactions year on year, both on total appointments and also digital-only interactions. And when you sort of look at the graph there, you see that that mix is basically staying proportionately the same. The reason why that's important is that you see increasing use of our digital tools. And then that's augmented by human consultations at the same time. I think it's very important to note you can't entirely replace one or the other. Both will be an important part of the mix going forwards. And you've seen how the product delivers that during the course of today. Again, sort of clinical utilization, I think the reason why this is really important is that it shows what we can do when we're at scale. One of the things that you saw earlier, as Ali said, is that we've delivered very, very high-quality care. High-quality care also comes at a cost, and when you're scaling very, very rapidly, we've made sure that patient safety is the one thing that we do not make any sacrifices for at all. What that also means, though, is you do have underutilization of clinicians early on because you want to get people onto the platform using it, and then you start to get critical masses, so what you see in the U.K. and Rwanda, for example, in the more mature markets have been going for a number of years, have hugely high levels of clinician utilization. Whereas in the U.S., it's sitting at around 58.9% at the moment. There are two ways we're solving this. So one is that we're having cross-state licensing across all 50 states. So basically, everybody can be in the same pool. At the moment, if we start in a new state, we basically have if you think about the first second that we start, we basically have a doctor who's effectively doing nothing for a while. And then as we get that engagement, that utilization increases. Once we've got the cross-state licensing across all 50 states, we can hugely improve that level of utilization. So there's a big cost drag from Babylon's own costs at the moment on two things. Basically, one is that scaling cost, but also secondly, the cross-state licensing, which we're enabling during the course of this year. What also goes with this, though, is the clinical care delivery expense margin is also going down massively, right? So at the start of last year, when we had only a few contracts going compared to now, huge improvements in the level of Babylon's own costs going down to 9% as of today, and we expect to see that continue to improve over the course of this year and also next year. Ali touched briefly on this slide, but I think this is actually also hugely important in the context of the overall market today. So starting at the far right and on Medicare as well, what we're actually seeing is that our longest cohorts, we're exactly in line or better than others in the market. So there's no outliers from Babylon's position at that point. On Medicaid, again, the engagement on Medicaid, as you know, is also significantly harder and takes a little bit longer. We've articulated in our results previously, in our Q1 results, that we intend to increase the mix of Medicare and commercial. We've been doing that actually already during the course of Q1, and we've shown how the patient numbers have improved in that mix. But I think it's also really important to contextualize this. As we've mentioned a number of times before, first of January 2020, we basically had no business in the U.S. from a revenue standpoint. And therefore, it takes time to earn trust from payers in order to demonstrate our ability to deliver on some of our promises. And that's the reason why we started with some of the most difficult populations possible in Medicaid. We've done extremely well with that, delivering high levels of clinical quality and patient satisfaction, which is also then a gateway to enable us to win further contracts. I think sort of one of the things that we also articulated during our listing process last year is that we actually saw fee-for-service as an entry point into value-based care. One of the things that we're very much seeing this year is actually we can skip over that largely. A lot of our new contract wins we've had recently have been with payers where we have not had a previous fee-for-service contract. That is also hugely important when we think about medium to long-term profitability because we don't need to take that short-term hit on fee-for-service revenue in order to win the value-based care revenue in the long term. So again, when we look at SG&A costs and technology costs, substantial revenue hugely declining. And basically, if you look at our P&L on a quarter-by-quarter basis, our OpEx costs, so i.e., not including our cost of care delivery costs, are basically staying flat almost quarter on quarter. So what you can see is we can hugely, hugely grow revenue while at the same time keeping our SG&A and tech costs back. I think that demonstrates the big advantage of having a digital-first business over a physical-first business. And we can't reiterate that point enough in terms of just the business model differentiation that you see versus others. And that basically continues to deliver that steady margin improvement to just below 30% of revenues guided for this year. So where our guidance as of May 21st was sorry, May 12th was $295 million of negative adjusted EBITDA on $1 billion of revenue. And I think sort of the other thing here that's hugely important to emphasize is the importance of scale in order to deliver that profitability. One of the things that you'll notice is that compared to our projections that we put out at this time last year, in terms of absolute number of dollars, the adjusted EBITDA losses are higher, but the revenue also is almost 50% higher with that as well. So we also, though, very much take a disciplined approach to growth. I think the market environment, as Ali said, and as we all know, has massively changed in the last few months. But at the same time, though, while we have a huge level of focus on delivery of those margins, and you've heard that from the team today, that takes a few months to change, right? So we had a huge, huge pipeline this time last year. We've executed on that pipeline. We delivered on the promises that we delivered and actually, frankly, surpassed them as well. And then the market environment changed. So we have to change with that. But at the same time, operationally, you can't just completely turn on a six-month on these things. But what I really want to emphasize is that with that massive pipeline that we've got, you actually don't need a sizable level of percentage point margin in order to get to that EBITDA break-even from a profitability standpoint. We've guided, as you know, to delivery of that by the end of 2025 or sooner. That's sort of one of the key areas of team focus over the next few months and years. Just in terms of guidance, I just want to talk about the guidance we put out on May 12th. I'm not commenting on whether it's an update to that. That this is the guidance that was at May 12th. Basically, full-year revenue guidance of at least $1 billion. Then with that, an adjusted EBITDA loss of -$295 million. That's basically continued delivery on those contracts and also preparing our delivery for 2023 at the same time. The break-even, as I've just mentioned, and then also on the funding guidance, I just want to make sure that this point is completely clear. The funding guidance that we've put out has said that the cash that we have on balance sheet at the end of Q1 is sufficient to fund 2022. That does not mean, and I just want to be very clear on this point, that we would not consider doing any incremental funding during 2022. What it means is that the cash on balance sheet, we do not need any cash on balance sheet in 2022, further cash in order to be able to get to the end of the year and beyond that. So that's really sort of a summary of the guidance we put out from that point in time at May 12th. Appreciate we're finishing a little bit late. So that's the reason why I've been running through my slides. I want to also make sure we've got a little bit of time for Q&A for Ali to wrap up. There is also some drinks afterwards. Should you wish to ask the senior leadership team any individual questions, or we can set up individual meetings as well. But I would love to take any Q&A from me or anybody else at this point in time. Thank you so much. Just a follow-up on the EBITDA guidance. A follow-up on the EBITDA guidance with the losses of 29 or EBITDA break-even is my question. Is that on an annualized basis for 2025, or that you will be EBITDA break-even by the end of 2025? It's by the end of 2025 at the latest. But as we say, we hope to improve on that guidance. But that's our drop dead date on the guidance. Okay. Thanks. Dave Larson with BTIG. So thank you for all that data. Can you maybe talk a little bit more about the margin profile by plan? It looks like here on the commercial side, am I seeing this correctly? You're basically delivering now a 19.6% gross margin on the commercial side, or is that what you project to deliver? And by what timeframe? And for Medicare Advantage, it looks like a 14.3% gross margin. Can you just give a little more color around those items, please? Yes, sure. So that's around the medical loss ratio. So that does not include Babylon's own costs, to be fair, that are overlaid on top of that, but does include the effect of the stop-loss insurance on a net basis. So when we think about delivery of the Babylon MLR, we incur in some cases some premium for stop-loss insurance, which we net off the medical claims costs, and then we get some money back when those stop-losses get invoked. So the way to think about that is exactly as you say, Dave. It is as of today on both of those cohorts, on Medicare Advantage and also commercial. But that is only the medical claims cost element of that. And what is the? Sorry, just to be clear, this is the 2021 results. This is 2021 results, correct? Yes. That's what we've done, that's the delivery. It's not where we want to go. Where we will go, this is what we've done in our first full-year report. So that's what you did in 2021 in your first report? That's 2021, correct. And what is the Babylon admin cost piece roughly as a percentage? So. Yeah. So that was on the previous slide where basically we show that going down. So that's 9%. So that's the 9% that you've got here, which is the Babylon cost overlaid. Now, the one thing I want to be very clear on with this is this also includes a recharge that goes between value-based care and fee-for-service, which is in some ways an internal allocation cost. That is something that we're looking to continue to reduce during the course of this year. And a lot of this cost is driven by this utilization point. So it's driven by this utilization point. So as you get the utilization up, the Babylon cost of care margin goes down massively as a result. Because if you think about it, we've effectively got physicians sitting around doing nothing half of the time at the moment. Okay. So that 19.6% for commercial, if we subtract, say, 9% admin costs, you're really looking at like a 10% sort of gross margin. Is that the right way to think about it? So if you allocated the cost to commercial in that way, yes, broadly across the board. Yep. Can I just supplement on that? So the way to think about this is the majority of those costs of care that you see goes towards fee-for-service. Can you imagine a small uptake on fee-for-service? I like that you could assume that the majority of that cost, the vast majority of that cost goes towards that you subtract the substantial. Okay. So the 20% commercial sort of margin that I'm looking at, or the 19.6, you wouldn't subtract 9%. Maybe you take out 4% or so. So you're looking at like a 15% gross margin for commercial business. And I think that you said that Medicare and commercial are what percentage of your total value-based care lives now? It's close to 50%. 50%. Okay. So you earn the business through Medicaid, and you're increasing sort of the quality of the lives that you're bringing into value-based care. Exactly. And that will continue. So as you've seen with the DCE contract, for example, that's continued to increase through this year. Okay. Great. Thanks very much. Hey, Daniel Grossleit with CITI. Thanks for taking my questions. I'd like to stick on the margin question. Maybe if we can flip to that cohort slide. Just, I guess, more of a technical point. On the commercial margin, is that under the professional services capitation, or is that the global capitation? Sorry, I missed the second bit of that because I'm trying to fix the slide. That's okay. We don't need it. We have it on our screens. In the commercial market, is that professional services cap, or is that global cap? Because I know the Medicare business was going from professional. It's professional services at the moment, but we're trying to transition that to global cap. So as that transitions to global cap, should we see that margin fall? I think, look, we're still undertaking the commercial side of that transition at the moment. It's where we end up getting to on that. So we haven't got any guidance on that at this stage. Okay. And then on the Medicare side of things, I noticed that if you look at the Medicare growth from 4Q to 1Q, that was driven all by DCE, which comes in at 104% MLR. If you look at MA, that actually just dropped a little bit from 4Q to 1Q. I'm just curious how we should think about growth in Medicare over the next couple of years. Will it be driven by the DCE, which is at a much higher MLR, or will you be able to reaccelerate that MA growth? It's both. We've got some new MA contracts that are due to come online. I think the other thing that I emphasize about the DCE contract, in particular, getting to the 104.1% MLR, is that we've got very scant data on that at the moment. What we've done is we've taken a market trend, which is effectively 100%, and then we've incorporated some of the drags within that. CMS, for example, retains 2%. That's 2% of the drag that sits within that 104. We've then got some more, which is, for example, we don't accrue 100% on the star rating bonus, quality bonus either. So our view is that we've been quite conservative on this, particularly when we look at how others have reported in the market, where you've seen others who are reporting under 100% on their MLRs. But effectively, the way that we've got to 104% is largely structural drags that sit within those contracts. Got it. And then last one for me. If I think back to last year, it seems like decades ago, but it was just last year where you were doing your de-risking process, you mentioned that you could get to gross margins of around 30% eventually. And you thought maybe by three years, you could get to 30% gross margins. Given what you know now, is there any change in that assumption that you can get to 30% when you reach scale, and it will take you three years to reach that 30%? Yeah. So look, I think it mainly depends on the cohorts that you're taking on. So for example, as we've kind of shown, it's harder to get on the higher margins on the Medicaid cohorts than it is, for example, on the commercial or on the Medicare cohorts as well. As we've also said, and you've heard from Darshak, we haven't gone through any recoding, for example, which we also see as one potential further upside. The one thing I'd say, though, is that when you look at other comparables in the market, particularly with people who have established physical-first practices, for example, they are showing 30% cohort margins in some of their cohorts. We have a structural benefit in being a digital-first offering into where we think we can get to. Now, at the same time, though, our focus over the next few years is covering our costs and getting to that EBITDA break-even point. That is the key thing for Babylon at this stage, rather than necessarily focusing on getting a small number of cohorts, for example, the very, very early contracts, to those 30% margins. Our focus really is taking on the revenue that will cover the OpEx, and we demonstrate that through the scalability rather than necessarily getting to 30%. Yes. There are two 30%s. We had one that was on the 30% on the B2C side of the business, which we thought we could get to on some cohorts. Then there's the overall gross margin on the overall business, which is the medium-term target, which was based on, as Ali said, including the software business at the same time. Yep. Understood. Thank you. Hi, Brenda from Berenberg Capital Markets. I just want to make sure that I'm understanding one point a little bit clearly. You were saying that we don't necessarily need a large% gross margin to get to profitable EBITDA. Can you just walk me through that one more time, please? Yeah, sure. So unfortunately, I can't flick the slides forward to the relevant one, but I'll explain it, which is we've got a huge, huge pipeline of revenue growth. As we grow that revenue, and we've shown that we can grow the revenue from $323 million last year to over $1 billion this year without really increasing the OpEx level that much in the business. Therefore, our core focus is making sure that we have sufficient revenue at, frankly, whatever margin it takes to cover that OpEx base. The reason why we can do that is, in some ways, you are seeing in this business growth, even in the services side of the business, with margins and growth trajectories that look like software businesses, where the OpEx doesn't change. The reason why we can do that is we can take on tens of thousands of patients without incremental technology costs, or with very small incremental technology costs. Therefore, our focus over the next couple of years is delivery of huge revenue growth with margins that are sufficient enough in aggregate to be able to deliver profitability to the overall business. Got it. Thank you. Thank you. One question on the utilization in the U.S. That was a bit lower, around 50%. A, do you think it's possible to reach UK levels of utilization? And B, what's the path to get there? And also related to that, could you shed some light on the cross-state licensing? What does that mean? Sure. So the one thing that actually the data point didn't go back far enough. One thing you'd see those utilization levels in the UK, for example, when we launched GP at Hand in 2017, were actually very, very similar. Actually, I think even lower than the numbers we've seen in the US. So we've basically got a high level of forecasting accuracy in order to be able to deliver high levels of utilization. The thing that's been blocking or the few factors that have been blocking that during last year and also this year is basically number one. If I take on a new state, for example, take on the state of Mississippi, we need to have clinicians who are licensed in Mississippi, and as of right now, we can't necessarily use a clinician that's licensed in, say, New York in order to be able to deliver that. So when you first start delivering for those patients in that new state, you've effectively got 0% utilization, which when you look at it on a blended basis, contributes to that 50% utilization. What the cross-state licensing does is it unlocks that, so we basically license our clinicians across almost every state. Improvements in the product also means that if we've got clinicians who are not licensed in every state, you can pre-plan with the forecasting models that are very accurate in order to make sure that we've got the right clinicians at the right place at the right time. That's really what will be driving that utilization improvement during the course of this year. That cross-state licensing, is that on a state-by-state basis, so to speak, or how does that work? Some states are on a state-by-state basis. For example, California is on a state-by-state basis. Some pool their licensing. Say, 10 licenses. Sorry, one license covers around 10 states. It varies slightly. I just had one quick follow-up, please. On this Medicare Advantage margin chart, there's something that says full book 4.0%, longest tenured cohort 14.3%. In a previous question, there was a comment that the MA book had an MLR of 104%, which would imply a negative 4% gross margin on your Medicare book. But if I'm reading this correctly, what that means is, for the full book, sure, it's a negative 4% gross margin, but for the longest tenured cohort, you're looking at probably closer to a 10% positive gross margin because you've had more time to stabilize those lives. Is that correct? Am I thinking about it correctly or not? No. So the 104.1% drag is on the DCE business that started January 1st. This was going up to the end of 2021. What margin can you get that up to over time? On the DCE business? And the MA business, yes. Yes. So we haven't done enough analysis on that at the moment and had enough time to see the engagement with the data coming through to be able to determine that. But we don't think that it's going to be out of whack with any of the other MA books we've got. Because with the DCE business, you get an extra 8%. The plans aren't keeping their share of the admin fee, right? So the premium you collect should actually be higher. So it should probably be at least in line with your longest tenured MA cohort, right, or not? So you're correct. You get that after a while, but you've got the inherent drags coming in through that at the start. And then we've got our own costs that sit on top of that. All right. Thank you. I hope I don't get myself in trouble. If I do, just sit there and kick me. But think about that 104. That 104, frankly, we have no data, right? It's three months in, end of Q1. We got these lives right at the beginning. So we did something that is super conservative. We basically said we have to add 2% for the state. We get 4% on the quality things. Let's say everything goes wrong, we only get 2%. So our starting point is 104. And that's all you got, right? It's no reflection on the reality of the numbers we're going to get. It's just that we need to report, and we've been super conservative on that. And as you quite rightly say, there is no reason why those data should be different. What is super encouraging to me is that on year one, the data we've shared with you on our first cohort is in line with some of the people who've been in this market for 10 years, right? The other thing to keep in mind, much of the technology that the team have showed you has not been deployed in that cohort yet, right? A lot of what we're doing, we will see the results coming our way. Hey, Allen Lutz, B of A. One for Jan and Steve. There was a slide on the health graph. I think you went through really quickly a lot of the different KPIs that you look at: BMI, weight, blood glucose, things like that. But there were a lot of them, and you went through that very quickly. I guess, what are the major ones that you're looking at? What are some of the data sources that you have? And I guess, what are some of the most important insights you can get from some of that data? Yeah. One big change that happened actually at the beginning of this year when we actually signed on an account. There's a load and an integration that happens in member records that are created in our health graph before a member even registers. So that means we have a digital persona and a profile on that member in our health graph. And so that includes a variety of data sources. That could be the claims data. That could be the medical records data. We have basically a series of integration feeds that we require from each of our customers when we launch. Our preference is real-time. Not all of them can provide real-time, but that's our preference. We then go after basically some various aggregators in the United States, and we choose those aggregators based on the quality or coverage that they have. Redox is one of the aggregators that we use, and that allows us access to, again, additional information. We've also partnered with additional networks that allow us to get to discharge information as well as lab information. And then we have developed direct integrations into specific EMRs as well. And then on top of that, the wearables. And so both direct access into the Apple Health as well as Validic, which I think gives us 500-plus devices that we support today. So it's a wide, wide variety. We're casting a very large net. And so depending on what zip code you live in the U.S., we could have a very high hit rate on certain areas and low on others and vice versa. But really, the intent is, how do we have a robust strategy that allows for us to ingest the maximum amount of data, assemble that, and link it together and unify it? I think the other breakthrough that we didn't talk about because it's getting into the nerdy part a bit, and that really is, how do we take that data that is unstructured and actually codify it so that we apply knowledge to it as well? And codifying means we're actually taking things that haven't been coded, haven't been structured, and we're applying structure to it in a Firebase standard. So when we say health graph, that's actually not a data warehouse. That is a very structured knowledge set of information that's been codified and structured and linked in a way that we can now use across various prediction models. Great. Thank you. We have a spotlight shooting in our eyes here. Are these the most comfortable chairs you've ever sat in in an intensive health presentation? They are very comfortable. Glenn Santangelo from Jefferies. Ali, I just had a quick question for you. It's a question we kind of get a lot. I'm kind of curious. I mean, it seems like there's so much capability you have once you onboard these patients. But I think I'm kind of curious to find out a little bit more about what type of due diligence you do on these patient populations before you take them on, right? Because obviously, not all patient populations are created equal. And what's to prevent some of these state Medicaid agencies, for example, from adverse selection and handing you their most expensive patient populations, right? How do you protect yourself from that happening so you're starting from a better point? Remember, we don't get a set number. We get a percentage of the we get the MLR for those people. So we look three years back. We look at all the statistical data on that very same cohort of patients. So there is no selection in here. So imagine if I cost the system $10,000, Darshak costs $10,000, you cost $5,000, and PH costs $20,000. And we get Darshak and PH, we get their $20,000 and his $10,000. And my $10,000 or $5,000 and yours doesn't come in, right? So we are protected because we go on historic data of the individual and what costs it for the member. I hear that a lot. And Glenn, you're right. I mean, people say, "Oh, you must have grown so far because people must have given you their worst," right? Just not true. People gave us some of their people that they had issues with. We took those. But the prices we got, we got them right. Also remember, we showed a historical loss of minus 5%. That's people that we got for the full year. And a lot of those people, minus 0.5%, not minus 5%, minus 0.5%. These are guys that we had only for a few months. So we had actually very little ability to affect them yet. And once we started affecting them, we should start seeing the results. So I think the data speaks for itself that we're not taking bad cohorts in any way. And one of the things that is interesting, I mean, we announced that already. Our customers are seeing the data, even the kind of data that we don't share because contractually we can't share. And the fact that they're now coming to us and saying, "We're happy to do a kind of deal that says we're pretty confident that you can manage the downside, so let's just share some of the upside." So going to your question on why is it that some of our upsides we're now tampering with is because in some of our newer deals, in return for not taking any of the downside, we're sharing the upside with our customers. Does that make sense? Yep. Any other questions? Okay. If you can stay, I don't want to be the guy who sits between you and hopefully some refreshments. So thanks so much for your time. Thanks to the team for being here. I'd love to hear from you later over during how we can improve on what we're giving. Our approach to this is complete transparency. It has always been where we came from, where a lot of the money we are paid for was taxpayers' money and the government money. So we started with total transparency. We're showing you a lot of what was under our hood so you can see the plans and the systems that we are building. Not everything we showed you is built, but as we said, everything that we showed you will be built in the next few months. We're not showing anything that we are not comfortable with. We have a lot more in the planning that we haven't shown yet. But we think that when you add all of this together, when you add the fact that we can take data from 100 different sources now, 80 billion strings of data, and within days, integrate them into a health graph, that we can deploy AI models in order to predict who's going to go to a hospital in the next 12 months, soon who's going to have urgent care within weeks, and so on and so forth, where we can add on AI data from any model outsourced, where we can then continuously monitor people, where we figure out how to engage more and more of our people in more and more seamless ways. The model of healthcare we're paying, nobody has. You don't have it. Doesn't matter how rich you are, how wealthy you are, who you are, you don't have it. We're giving that to some of the poorest communities in the United States. The people that Darshak talked about are not the rich, the powerful, the well-to-do. Some of those people have access to the very same thing. By solving some of the hardest problems in U.S. healthcare, we will be solving some of the, we will be putting ourselves, I hope, in a place that is going to be very hard for others to reach. I think we are adding structural advantages to Babylon that it's hard for others to catch up to. That's our plan. We're going to go steady. Of course, the market has changed on all of us and on us, and we're very conscious of that, and we'll make the necessary adjustment to live with the current markets. But we also won't get deviated from building, but we'll make a fundamental structural change to the way healthcare should be delivered. Thanks so much for your time. If you can stay with us, we'd love to host you over a small amount of refreshment, and then the team is here to answer any particular questions you have. Thanks for the time.