All right. Good morning, everyone. Looks like we are officially on the clock for day two here. So welcome to day two of the William Blair Growth Stock Conference. For those of you whom I've not yet met, my name's Ryan Daniels. I cover the HCIT and healthcare services space for the firm. So it's my great pleasure to have our speakers from Augmedix here today. To my far right is the company's Chief Financial Officer, Paul Ginocchio, and then Manny Krakaris is to my near right. He's the company's Chief Executive Officer. Matt Chesler is also in the back. He works with the company on investor relations. Really happy to have the team here today. This is an exciting growth story. A small cap company for sure. Certainly has been a bit of a noisy space with a lot of competition, but a really unique asset, especially for the size of the company. Some great partnerships with Google, a relationship with HCA, which is the largest hospital chain in the United States, and a solution in our channel checks that really stands out for increasing provider efficiency and satisfaction, improving workflow, really driving the patient experience, enhancing revenue cycles. So a true HCIT technology in healthcare that actually increases provider satisfaction, which we don't see a lot, quite frankly. It's usually a burden when you implement technology, so really unique. It's also one of the few assets that's really harnessing artificial intelligence, again, internally and with the Google partnership, to drive this. So I think a really interesting opportunity to hear the story in a bit more detail from the team. With that, I'm gonna turn it over. Just two quick reminders. Number one, on our website at williamblair.com, you can get our disclosures. Number two, after the formal presentation, we'll stop, and then we will go upstairs to room Burnham B. Again, Burnham B, where we will host the full Q&A session. So with that, I'll go ahead and turn it over to the company's CEO, Manny. Manny? Whoa! Excuse me. Thank you, Ryan. Let me just get this on my belt. Okay. For those of you that aren't familiar with the company, Augmedix. Actually, let me just ask a question. How many people have actually seen a doctor during their lifetime? Some of you, right? Well, when you see a doctor, the doctor is required to document the encounter in your electronic health record. That task consumes two to three hours of a doctor's day, every day. When you take that and combine it in a situation where the capacity of the industry to provide care to a growing and aging patient population, you have a combustible mix. Doctors are burning out because of the burden of trying to maintain this documentation and see increased patient flow, and patients are dissatisfied with the level of care they're getting because doctors are oftentimes going through a perfunctory process as they have these encounters with their patients. They're checking boxes, looking at a computer screen with their backs often turned to the patient because they have to manage these multiple tasks. 11 years ago, we thought there was a better way to handle medical note documentation. We decided that repurposing the conversation that occurs between the doctor and the patient, using that as the input to create the medical note, would be a better way. It would save doctors the extra step of having to spend more time on medical note documentation and would also afford them the opportunity to interact directly and exclusively with the patient during the encounter, as opposed to splitting their attention between a screen and the doctor. That was 11 years ago, and people thought we were crazy with the idea. Fast-forward 11 years, and now the entire industry is focused on ambient medical note documentation. The advent of large language models has made that much more easy to do. And so there's a lot of noise in the industry regarding the application of LLMs in medical note-taking. Let me move forward here. So the idea is, of course, to harness the conversation between doctor and patient and use that as the input. It's not perfect because a lot of things that happen during an encounter are not spoken, and that varies based on the specialty and the care setting. And I'll talk a little bit more about that, but it's a good foundation for creating the bulk of the note, and then you have to do other things to fill in the necessary gaps to avoid placing burden on the clinician. So, as I mentioned before, we use a variety of technologies. The AI products that we have in our portfolio essentially take a recording of the ambient conversation, run it through automatic speech recognition models, convert that into a transcript, a written transcript. The written transcript is then pushed through a series of natural language processing algorithms that we have that identify the key elements within that transcript that we believe are relevant for the medical note. And from that, each one of those elements is prompted through our prompt engines to ask questions of the large language models that we employ to generate the appropriate content in a structured medical note. Most of the notes that we generate are problem-based, which is very different from the rest of the industry, which are SOAP-based, subjective-objective notes, which are essentially summaries of the conversation. That is not terribly useful to most doctors. They like to have the note organized by problem because in many instances, in fact, most instances, patients come in with more than one problem that needs to be documented. If it's not problem-based, you're gonna have symptoms and other matters related to a particular problem interspersed and interchanged throughout the note, and that's difficult from a review process for a physician to review the draft medical note that the technology generates. Some highlights here. At the end of the first quarter, we had an ARR of $53 million, net revenue retention, which is essentially like same-store sales of about 143%. We focus on large enterprises. We're in seven of the top 20, five of the top 10 U.S. healthcare enterprises in the U.S. today, and we've documented more than 10 million encounters, from which we have gleaned a great deal about what is important in a clinical note based on a particular specialty and the preferences of a clinician in that specialty. Some of the benefits I already talked about. So, there are 1.1 million practicing physicians in the United States today. We estimate the TAM for the portfolio of products that we have on market today to be about $8 billion. The enterprises that we have contracts with today represent an addressable market of a little over $1.5 billion. So we have direct line of sight to a sizable market opportunity. Within that, $1.5 billion-$1.6 billion TAM, there are roughly 150,000 doctors who are employed or directly affiliated with those 20+ healthcare enterprises with whom we've contracted. Some of the enterprises that we've contracted are presented below. This is just a subset. We started the business with a synchronous full-service solution 11 years ago called Live, and that's where we pair a medical documentation specialist one-to-one with a clinician throughout the clinician's shift. That resonated deeply with the healthcare industry, and it resulted in substantial growth, historical growth for us, since its inception. I, in fact, got inspired to join this company by my doctor, who was a customer, a Live customer of Augmedix, still is, at one of our larger healthcare systems. The reaction that she gave me as a patient, I was not affiliated with the company at the time, was something that I remembered when I was brought in by the board to talk about the company, and that is, she said this was life-changing for her, and I didn't understand that because I've never been in healthcare before other than as a patient, like most of you. I remembered that reaction when I met with the board, and they were describing the company to me. I did some research after I met with the board and heard similar anecdotes from other doctors, how important this was. To me, what I decided was, Look, you've got a company here that's got a great product market fit, a huge market opportunity, and not a lot of, not a lot of resistance to acceptance. There's not a lot of friction in the process. So I felt really inspired, and I decided to join the company five years, a little over five years ago. Well, that product survives today. It was responsible for about 90% of our revenue last year. But we knew, I mean, since I joined the company, a little over five years ago, that that was not the ultimate solution. When I first joined the company, I spent the first couple of weeks sitting with people, our medical documentation specialists, observing them doing their job, and I concluded after a couple of weeks that I could never do that job. You could train me for 10 years, I'd be really lousy at it, 'cause it's really hard cognitively to observe something in real time and document it, like a stenographer, a court stenographer, for example. That's really difficult to do, and I just felt that trying to scale a business based on that high level of cognitive skill would be very challenging. So we set about building tools that made that job a lot easier. And, the next step was to automate as many of those tools as possible. So the ensuing years saw that happen. The idea was, look, if we could create a set of tools that were automated, where we could take someone right out of university in India, for example, and spend three weeks training them. It is not a far stretch to create a client-facing piece of software to enable the customer themselves to do the work, as opposed to our people, and then offer that at a much lower price. That was the vision, five years ago, and it's culminated in the next two products that are on this slide: Go Assist and Go. Go is a fully automated self-serve product that we launched recently. It is fully generated by our ASR, NLPs, and LLMs, generates a draft note in seconds that the doctor reviews whenever they want to review it, right at the end of the encounter or whenever they have time, and sign off on it after they made edits. They could edit it on the application itself, they could edit it in the EHR, they can edit it by typing, they can edit it by voice command, whatever they want. There it's multimodal, and we made that available, as I said, a few months ago, generally available. It is a self-serve product. The latest product that we've launched is Go Assist, and that is the same as Go. It's got the same back end as Go. It generates a draft medical note, but the clinician has the option to push a button and have that note directly transferred to a medical documentation specialist on our staff who does the last mile, does that last bit of lifting, editing, if you will, that the technology misses. There could be a break in the transcript, for example, because the, you know, doctor and the patient were speaking over each other, so the ASR gets confused. So there's a gap in the transcript. The MDS has contextual understanding and can fill in that gap on behalf of the doctor so that the doctor receives a more complete draft of the medical note than Go itself. So the doctor has an option to choose that particular product, should they, that his needs or her needs warrant it. We have a product for every price segment of the market at different price points. Live is priced at roughly $2,400 a month per doctor, Go Assist at roughly $600 a month per doctor, and Go at roughly $300 a month per doctor. So every price point, different value propositions for each product to address the largest possible TAM in the market. This is the landscape as we see it, in our space. On the far left, you have the legacy tools, which are dictation tools dominated by Nuance and M*Modal, and then you have the ambient products to the right of that in different categories. And as you can see, we're positioned in each one of them, with the exception of the in-person category. That's not a model that we've ever pursued, nor wish to pursue. That is a difficult model to scale. It requires you to place your assets, if you will, your people, in the same location as the customer, and that's becoming increasingly important and difficult to do in the United States because the cost and limited availability of qualified candidates to do that work. There's a lot of noise in the market, especially in the pure AI sector. I'm sure you've heard much of that, and what is interesting about all that is that no one, everyone's putting all of their eggs in one basket, which is a pure AI self-serve product. And we know from feedback we're getting from our customers and initial really strong demand for Go Assist, is that the pure AI product doesn't meet all the needs of all doctors all the time. That's something that people need to keep in mind. There are certain specialties, and there are certain times during the course of a day where a clinician needs some help, our job is to make sure that we have a solution that addresses those needs, not just the self-serve needs, but any other needs for greater service. We have the infrastructure to do that, nobody else does, and therefore, we think we have superior market positioning relative to anybody here on this page because we can address all those needs. It's not as if we require the customer to make a binary purchasing decision. They can choose on demand what they need, given their circumstances. That is a huge amount of financial flexibility we offer our customers that nobody else can match. Some of the competitive advantage, I just noted some. We've learned a lot about data sets and workflows from the 10 million+ medical notes we've documented. We cover over 50 specialties, with more than 35 EMRs that we work with. We have the broadest product portfolio in the market. We've built the infrastructure we need to actually scale the business, meaningfully. We have a bidirectional communication channel to the point of care. The point of care is ground zero in healthcare. It is the most valuable piece of real estate in healthcare. That's where everything happens, and having the ability to not only pull information from the point of care but push information to the point of care, that we can curate intelligently through the use of AI is very impactful, and we're just starting to do that, harnessing AI with, in the emergency department, with third-party data sets and rules engines that focus on risk and safety within the ED. There's gonna be more to come, in terms of clinical decision support systems that are governed by AI. The whole idea here is to harness all of the disparate information that exists out there that can be of value and use during an encounter to the clinician. The challenge today has been that it's very difficult to access that information during an encounter, and if it's not presented during the encounter, it loses its value. The ability to know when to do that and deliver just what is necessary is really valuable, valuable, and that is something that HCA has tasked us to do, and that is something that we are in the process of doing right now. Some of the technology of what we do, this is the back end. I won't go into the details, but primarily what we do is we take content from the encounter, the audio file. We combine that with information from the EHR, the electronic health record of the patient, and other sources, feed that into our process, which is multilayered. We do multiple prompts. What's not shown here is we do multiple prompts of the large language models. That's one of the, one of the unique things about our approach. Large language models are general-purpose tools. They're blunt instruments, very powerful, but if you ask a general-purpose tool like GPT-4 a question such as, "Okay, from this transcript, deliver to me a structured medical note," you're gonna get a decent summary, but you're not gonna get exactly what you need. So our job, from a technology perspective, is to narrow the aperture of the prompts you ask the LLM as narrow as possible to get a tighter amount of information into the medical note that is accurate and comprehensive. And we do that because we have natural language processing algorithms that I talked about earlier that identify the key elements within the transcript that we believe is relevant, are relevant to the medical note. Well, those elements are what we prompt the LLMs for. Tell me more about this medication. Tell me more about this particular symptom, and structure it based on the problems that are presented by the patient during the encounter. That is a very different approach from what everybody else does, and we do this in parallel with everything else, and we do that in seconds. Just some examples of what the application looks like to the clinician. You've got the transcript that is converted in real time from the audio file. Again, that's not common in the industry. We have a clinical data summary page, which is unique also. This is structured data that is very organized. It's not a flat file like the medical note, so that data today, we feed that into HCA's data lake, which they use to mine for a variety of purposes, including facilitating their RCM function. It's very difficult to mine a flat file like the structured note itself, and by organizing the data in something that can be digested through, you know, typical, you know, data analytics engines, you add a lot of value to what you're doing in the process of creating a medical note. It goes beyond the medical note itself and provides a lot of insights to the healthcare provider organization as to how they're doing from an efficiency perspective, but also how to eliminate or minimize denials and pushbacks from payers on reimbursement claims. The economics are very compelling. In fact, about two weeks ago, we got some data from HCA, some of the initial users of one of our AI products, that showed even more compelling, economic data than this, which is really encouraging for us. It's not just that we're facilitating or reducing burnout of doctors. We're actually creating a powerful economic incentive for healthcare organizations to adopt this technology, these particular solutions. So we're really encouraged by the initial results and expect that we will continue to lead the way in delivering very high ROIs alongside technology that sits in the background. It's very hard to develop a powerful piece of technology that is unobtrusive. That's been the key to Apple's success, and we try to follow that model. Everything that we design is intended to minimize the amount of interaction a doctor has to have with the technology, allow them to practice medicine the way they were taught to do it. Some financial statistics here. Growth has been robust. It continues to, we continue to grow the business. We are transitioning from a revenue mix perspective into away from that $2,400-a-month product and more into the $600 and $300-a-month product. So you need to have roughly 4-6x, sometimes 10x, the revenue, the volume, to equate to the same revenue if you're gonna have that substitution or transition. So we're going through that period now. That is why we adjusted our full year guidance down to $52 million-$55 million revenue this year. But that is where we want to be. That is a conscious decision on our part. That is where the market is, but we will continue to offer a solution that addresses each price point in the market. More financial statistics, consistent growth in revenue, gross margin. This is the team. I just want to point out one recent addition that's noteworthy. Dr. Alex Stinard came to us as the Chief Clinical AI Officer from his most recent post, leadership post at HCA. We had worked with Alex for the last year and a half. We invited him to join our AI Advisory Council last summer. He represents the cross-section of the clinical world and AI. He's an expert in both, and he provides valuable insight into our thrust into clinical decision support systems. We're leveraging his knowledge and expertise to further develop that capability, which we think is gonna be very distinguishing and differentiated from what our other, our competitors are doing. Our competitors are focused on basically adding features to product. We're going much deeper by harnessing all the data, third-party data, that's out there that is inaccessible or not practically usable by doctors, and delivering it, acting as an intelligent curator of that information, and he's helping us realize that. Investment highlights. As I mentioned before, broad product portfolio with a compelling ROI story for our customers. We're in five of the top 10 healthcare systems in the U.S. We have great partners in, as, as Ryan said, Google and HCA. There are others, and there's some significant ones that we're working on right now, to augment those two. We have this bidirectional communication channel for the point of care, which is really powerful as long as you can harness it properly, and we have a scalable business. We have an infrastructure that we've built that we can grow on. Some of the new competitors in this space are, have developed products. They still need to build businesses, companies around them, the infrastructures. We've done that. We've done all the hard work, and now we've got a product portfolio that we can build on top of that or layer on top of that infrastructure. That is the end of our presentation.