We've got a couple of minutes. Everyone has a break that they're coming into, so we're going to get started right at 3.28. Everyone has a moment. Everyone should have been given co-host ability to turn a camera on as well if you want to test. Awesome. Good to see you, Ty. I forgot. Are we using our own? Are we going to share our own screen so our content can be our own? Same stuff we sent in, but. Yes. I'm going to share Samantha's for her because I believe that she is not able to at the moment. If for some reason you need me to share yours, I can. But the plan, I think it's smoother to have you share your own. If you want to try to practice that really quick, that you can pull your slide up. Go for free. Awesome. Matt. Hello. All right. As a reminder, I'm going to prompt everybody to introduce you all, and then everyone gets 10. As you're coming up close to the end of your quote-unquote time, I'll probably give you a warning if it looks like you still have a little bit of ground to cover, and that way we can keep somewhat on track. But this is the last session today as well. If we run a little bit past, not a super big deal, the goal is to end by 4.30, so Eastern. Hey, Jessica. Sorry if I missed this, but do I not hit start webinar? No, I'll go ahead and do that. Everyone's taking a break really quick that's been in the conference, and so we're going to get started in three minutes and I'll start the webinar. Awesome. Thank you. You're welcome. You want us to stay on with video until our turn? How do you want us to manage our time? What may be best is to have your video off. The thing is everyone's going to be sharing, so the videos will be at the top anyway. It won't be taking up the full screen while someone else is talking. If you want to keep your video on, that's totally fine. But it may be a little bit cleaner if you want to turn yours off until I prompt you. Order as a reminder, Samantha first, then I'm going to have Ty, then Pat, then Matt. Do you know if Laura is getting on or not? Did she respond to you? She mentioned that she was going to. Okay. I'm keeping my eye out for her, but if she is not able to jump in, I will go ahead and do your introduction. That shouldn't be a problem. Thank you, ma'am. Thank you. And then are you just going to say okay next slide for me, Sam. I was just pulling mine up to see if I could share it. Oh yeah, if you can share your own that's even better. See. I can see you. Oh wow. Okay. So, if you do the slide show. Let's make sure it works. So it's pulling the second screen where you see the notes. Can you try a new share. Yes, then share the PowerPoint slide show. It's still showing the other side. I don't know why it's doing that. Do you want me to go ahead and do it then? Because I've got it pulled up. That would probably be best. Okay, so let me do that. And I'm going to be starting the webinar now. Hello, everybody. Thank you so much for joining us. This is our final session of the day. So hopefully you are excited to jump into this last session. It's going to be a little bit different, a little bit more rapid fire than our other sessions, because we have actually compiled four different panelists who are going to be sharing a use case for AI in orthopedic practices. And so each presenter is going to be getting 10 minutes to present their use case to you. And then after everybody has presented, Kathy Lada, who is our chief marketing and membership officer, is going to be doing a little bit of a wrap up for today's event as a whole as well. So our first presenter for our use case presentation is going to be Samantha Towler. If you were on our keynote session, you heard a little bit about Samantha at that time. Samantha is with Tennessee Orthopedic Alliance. So I'm going to hand it on over to Sam to get us started. Thank you, Jessica. So if you want to go ahead and go to the next slide. Thank you. Oh, maybe the next one. OK. So as far as so we'll start this slide. So as far as our labor productivity gains in the past, we've been using Infinex for almost three years now. And we have no new FTEs since we started this. I still have the same amount as when we started in April of 2022. And as far as denials go, we are denial rate is extremely low. And then for physical education or physician education, this is huge for us because we had a huge issue with our documentation from the physicians. So that's been huge for us to sit with them and let them know what's going on as far as the denials go. Next one. Thank you. So increased productivity for us. Like I said, we've not hired any more employees, which has been huge over the past two and a half years. We've added more physicians across all of TOA. We have 22 locations. So we've added several physicians and not had to add any more FTEs. And with the automation for us, the HL7 integration has been huge because it takes less than a minute to upload cases. So that's been a huge win for us as well. The next one. So for our denial rates, one specific instance that I pretty much tell everybody is. So from December till April of this year, we had 900 cases that we had between MRI, CT, surgery, pain injections, all of those. And we have a less than a 1% denial rate for those 900 cases. So going back to our physician education, that's huge for us because they document more. Custom reporting has helped me a lot, especially in those instances where the denial rates, that was our physician's biggest thing in the beginning. So the denial rates, the customer reporting through Infinex is amazing. I can also pull productivity for my employees that way too, which has helped me considerably since we started sending everybody to work from home. And as far as the misconception that employees are gonna be replaced with AI, that has not happened. I've stressed that to my staff. That's something that was not gonna happen. As far as are my employees concerns go. So that's been a huge win for us as well as the employees know that this is not something that is gonna replace their job. So that's been huge. Okay, and okay. So, oh, sorry, I messed up my slides. So Infinex, honestly, Infinex has been huge for all of my staff and not only just my staff. We have several departments that are using it now as well as staff. We're actually starting to use it for eligibility and benefits as well. So not even from an authorization standpoint, but also from an insurance eligibility standpoint. This has, my entire team, as well as surgery scheduling and our pain injections, it's all the same process. So for us, that's huge for me because I like to train on different things. So for me, I can work in any department. If I'm using Infinex, all of the presets are the same. So that's streamlined that process a lot. As far as denials go, we have sat with several physicians if we continue to see denials increasing for them. And we all know a lot of it's six weeks of conservative treatment. So we've sat with those physicians and that's made a huge impact as far as our approval rates going up and our denial rates extremely low. So that's been huge. And then I think we can go to the next one, Jessica. Okay. So that's all I have. Awesome. Awesome. Well, before you wrap up, is there anything else in particular you might want to share? Like from using AI, being able to use this one instance, are you looking at other opportunities for AI in your practice? Yes. Not exclusively anything off the top of my head. I know some other departments are, but this has been the biggest for us. And Infinix has several different options as far as, I mean, not just AI with Infinix, but I know that there's others out there. But with Infinix, there's several different ways that we can use it across all of our departments. So, and we're continuing to grow with Infinix and using it across all of our departments. So. Okay. Awesome. Does anyone have questions for Sam before we move on to the next presenter? If so, go ahead and drop that in the chat. And if it takes you a moment to think of it, we can always collect some at the end as well. So I'm going to go ahead and stop my share. Our next speaker is Ty Allen. Ty is the CEO of Social Climb, and he is going to be presenting another use case for us. So Ty, why don't you take it away? Awesome. Thank you, Jessica. I appreciate the opportunity to present today. I was excited when I heard about a conference that was specifically addressing AI usage in orthopedics. I'm going to talk about a couple of, actually one case study, how it works, and then share a member benefit that you probably don't know you have as members of AAOE that we set up recently. So first question. Actually, really my only question. What if you could increase your high value cases on demand? So by high value cases, I mean, for most of you, it's probably total hip, total knee, spine, and then one or two others where you have a lack of cases to keep your subspecialty groups busy. Having now worked with hundreds of orthopedic groups, I see, I can understand a little bit more about the details of the actual value of certain cases over other cases. So what we set out to build a few years ago was the capability for you to, on demand, turn up or turn down case volumes of the types of cases that you want. So it's AI-powered marketing that allows you to target certain types of patients and pull them into your practice before they choose to go elsewhere. I'm going to walk you through a case study on this from Desert Orthopedics in Las Vegas. We'll talk a little bit about how it actually works and why it works, how you could actually give it a try as an AAOE member benefit. So the practice profile that I'm going to highlight is Desert Orthopedics Center in Las Vegas. It's 26 surgeons, four offices, two surgery centers. And in 2023, they wanted to run some campaigns because they wanted to increase what most of you probably want to increase, which is total hip, total spine, or total knee and spine. Of course, there are lots of other value cases that come out of that, but if you could just magically pick, you might pick those because they truly do change the numbers for the practice. So in late 23, as I said, they started an AI-based patient targeting campaign. The goal was to add 2 million in revenue. So sorry, 2 million in charges. We'll talk about how it actually turned out. The results were 484 appointments came from the actual individual humans that they targeted using the AI targeting system. Of those, 283 were net new patients they had never serviced. So you can say about 200 patients were reactivated and 283 were net new patients who had never gone to Desert Orthopedic Center for care previously. Most of those cases were spine, total hip, total knee, but there was probably about an 80-20 mix on that. About 20% of them spilled over into other things. The financial results were a little over 2 million in charges, about 1.2 million in actual collections, they spent $37,000 on the actual targeting campaigns that they ran. So if you're keeping track, that's a 3,200% return on collections, not on charges, which translated means every dollar they spent, they got $32 back in collections. They counted as a success, as I'm assuming you would guess, and have since turned on numerous other campaigns. But this is the one that's the most complete that I was able to highlight. And I have dozens of other campaigns, but Desert Orthopedics was great with me sharing their data here, so that's why I'm sharing their data. So one of the postcards that they involved in this thing, so we're going to talk about actually what they did here in just a minute, but I just wanted you to see the simplicity of the message. Let us help you with your knee joint pain. That's one example of a postcard sent to the right patients, or right potential patients makes the big difference. That's just one example. So how does it work? This is their team, an example of their team, looking at the Las Vegas area, having identified that they want to find people who need total knee replacement. And they're looking for people in the extreme category here. You can see that they've then clicked with the blue outlines there, a bunch of zip codes listed there at the top, where they want to get more patients. They've chosen 14,600 people by selecting the zip codes they live in. So it is showing them by zip codes. Darker the zip code, the higher the density of potential cases in each of those zip codes. So it's generally looking at and grouping them by zip codes, but the individuals who are behind those are known. So what they're really doing here is identifying a group of individuals. Then they narrowed down who they wanted to target. So if you look at the first demographic line there, it says that those are by age. And they chose to turn off the 85 and older crowd. Not that they wouldn't service them if they came in, but they're not gonna spend marketing dollars to go chase the older folks. And they also turned off the sub 40 in annual income. It's because it's probably not the most targeted. They'll take them, they'll do work for them if they can. And then finally they turned this last link where it says insurance down there, it says non-government. They chose to do this marketing campaign to individuals who are not Medicare, Medicaid patients. I think they felt like they get plenty of those. They were going after more commercially insured patients who are in the right age, the right income bracket, and could hopefully help them increase the profitability of their practice through this process. They then did two things that went out to those patients. The right type of postcard. So if it was a knee postcard, a hip postcard, a spine or back pain postcard, went out to those individuals, got delivered to their homes. And all of those individuals identified through this process, they uploaded them by an integration into meta. So then Facebook and Instagram ads could be delivered to those individuals. My little example here doesn't show a matching ad, but the ads did match to the postcard. So the person got something in their mailbox, saw something on Instagram or on Facebook, and if they were in the right mode of pain or issues with that joint or their back, they were likely to engage. Seems all very simple. The key is identifying individuals who are very likely to need care. In our data sets, as we look around the country, we typically see about 0.2% of the population in any given area is likely to be in need of a certain procedure type. Question is, which 0.2%? I'm going to talk to you about how we help, how we identify those. So then it allows you to focus your marketing spend on that small and likely needy population, which means your response rates are going to be very high because you're not talking to everybody, you're just talking to those who are likely to need care. So the return on investment is typically $1, gets you $25, which is 2,500% return. Obviously, desert orthopedics was higher than that. I think it was more like $32 for every dollar spent. So the key is AI tools that can identify the needy population. So you might not want to know this, but I'm going to give you a quick brief on how using AI to identify individuals who need certain medical procedures can be done. This is how it works. So item number one there, where we have the audience identification, these are people who we know have had knee replacements. Let's say we start with a million of those. Since we do service hundreds of orthopedic practices, we do have a lot of this data. Now that's HIPAA controlled data. We can't use that data for marketing purposes, but we learn from it. We overlay that data with publicly purchasable data about every adult in the US. We buy 2,200 data points, multiply that times the million people there, and you got a lot of data that you need to sort through. We run a model creation process in step two that creates a machine learning developed AI model that can then identify people who look like from non-HIPAA attributes, the same individuals. So it's the same data, the big scary data set that's out there about all of us, about every detail about our financial life and all kinds of other other data points that are useful in this kind of analysis. We then test that against that control group of individuals who we know have recently had a knee replacement, and we make sure it identifies 80 to 90 percent of them. Then we know have a working model. We run that model in step four against all adults in the U.S. and look at the accuracy on that data, and then we present it based on your geographic location. So I showed earlier, I showed Las Vegas, and a little example down there on number six is showing Houston. We actually have this running in pretty much every state in the country has used it, and as more customers use it, the more people that captured through this process feedback into the system, and the system learns through the AI process how to be sharper at identifying people who are in need of care. We use, if you really want to get nerdy, we use neural network and gradient boost trees algorithms in our AI processes to continuously sharpen these tools, and the data gets updated every quarter, and the models get rerun every quarter. So as potential patients move in and out, our systems stay very tight with that population. I could go into a lot more detail, I'm going to spare you on that, and would happily do it with you at some other time. I'm just going to finish by saying many AAOE members are already using these tools, probably some on this call. If you're interested in trying them, it is actually an AAOE member benefit that we've made part of the deal as our participation in AOE that lets each of you as members use this, play with it, and analyze your own areas, and run small campaigns. So if you want to give that a try, I'd love to talk to you. Again, my name is Ty Allen, and you can see my email address there is just tyallen at socialclim.com. Thank you, Jessica, for the time. Yeah, absolutely. We had a question come in for you really quick. Do you have any data about whether the postcards worked over the digital? Because postcards seem like a waste, probably from a financial standpoint, they're more expensive than the digital in some ways, but they love the targeting. This is hard for me to say. But when my engineering team and my product guys came to me and said that they wanted to build postcards, I said, why would we do postcards? That's the past. Are we just doing digital? Are we just doing social? They said, let's try it. We tried it. Every very successful campaign of ours has postcards in it. The conversion rates from postcards are way higher than I thought. We're typically seeing customers get somewhere around 2% of all targeted people and burnt. And postcards, campaigns with postcards have almost twice as high a conversion rate as purely social or digital campaigns. Okay. One more really quick. Will this be against the ethics for age and income for health equity? And is there a risk to be flagged? So we've had that conversation with numerous groups and there's nothing in this process that says these groups do not service individuals who fall outside those specs that I highlighted. It simply says, if I'm going to spend my marketing dollars to target individuals to bring into my practice, I can pick and choose who I target. That's all we're doing here. We're not precluding anybody from getting service of any sort. We're just trying to target individual with certain attributes. So my answer to your question is, no, we've had attorneys look at it. Our customers have had attorneys look at it. There's been no one who feels like that is an issue. Awesome. All right. Thank you so much, Ty. Our next speaker that we're going to be welcoming is going to be Pat Williams from iScribe. He is the CEO and founder of iScribe Health. So welcome, Pat. Thank you, Jessica. Can you hear me okay? Yep. Very great. Thanks. Great. And here we go. Slides coming up. Yeah. Thank you very much, Jessica and everyone. Really appreciate the opportunity. Just for clarity, iScribe is a physician-facing clinical documentation tool. There's a variety of tools being discussed. We live and breathe on the physician-facing documentation side of the house. The way that it works, it's a very simple process using the smartphone of choice. Physician walks in the room, hits a record button, records the conversation with the patient, and then within about 30 seconds or less, we're fully automating the documentation summarization based on that visit, fully populating the EHR, including orders, diagnosis codes, billing codes, et cetera. Physician then reviews and signs off after the fact. So that's a high level of what's going on and how it works. This is our mission statement. This is something that we love to hone in on here and the lens that we sort of bring to the table when we think about our why, right? Why are we doing this? And the way we like to think about it at iScribe is what type of physician experience would we want for either our parents or our kids, right? I think we can all agree. Clearly, we want for our parents or our kids or ourselves, for that matter, a physician who is present, who's focused, as opposed to one who's rushed and running an hour and a half behind schedule and is burned out, right? And I think that's the experience that we're giving back to our customers is that ability to show up, bring their best selves to their work. How are we able to do that? Number one, it all starts with time. We're giving the physician their time back, which is their most valuable asset. We know the orthopedic landscape very well. It's where the majority of our customers are to date. This is a very busy, very high volume, very high demand specialty that demands time. It demands the best utilization of that time now more than ever with increasing patient population, you know, going through the roof and provider shortages. We've got to be as efficient as possible with our time, and that's what we're able to give back to our customers. We've got a lot of documented evidence of that that we're going to go into detail with about today. What I like to call you a reduction of sort of the cognitive load, right? There is no room for any distraction in the exam room for physicians to poke around in a complicated EHR while trying to see 50 to 70 patients a day. It just doesn't work. I think we can all agree upon that. That's what we're seeing is one of the real benefits of AI and incorporating it in this way in the exam room is, again, all about that focus on the patient, which is, you know, nothing's more important. There's now no such, there's no such thing as turnaround time anymore, right? When we think about legacy transcription services, legacy virtual scribe services that take 24 to 48 hours, those days are gone. The ability to generate a note immediately, regardless of the time spent in the room is quite frankly amazing. It's still something that it's hard to believe we're able to accomplish with this technology. And then the revenue optimization, right? I mean, again, in a high volume, high speed, high demand world to be able to optimize each and every encounter at the highest E&M level based on the LLM's ability to analyze each note against 2024 E&M guidelines is something that we're doing with each and every visit and optimizing revenue compliance like never before. Again, what are the real benefits of AI in the exam room experience? The short answer, it's time and money, right? I mean, with iScribe AI, there's no such thing as after hours documentation time anymore. All the things that have sort of plagued the industry for so long now can be gone in the utilization of these types of tools. And there's also, I think the real takeaway here that we understand is there's no one size fits all. So I think the ability for solutions in AI to meet physicians where they are, you know, a lot of physicians still have that traditional transcription process preference, the ability to match to that versus someone who's more comfortable with a point and click. AI can really, based on the way it's deployed within the iScribe AI workflow, meet that provider, again, on the spectrum of what they're used to. If they're comfortable in the exam room doing everything in the room or more of a hybrid, which is what we tend to see, a lot of providers like having that HPI story capture conversation in the room. But then when they step out, they like to go into more of a directive mode. AI can handle those two things really, really well, those two different scenarios. And I think that's something that we've operationalized very well that our customers are evaluating, again, with regularity. What are some specific outcomes that we're seeing? We've surveyed, we've taken an average of all of our customers that are using our AI tool today. We're reducing documentation time by as much as 75% across the board because, again, it's all being done, it's a true two-bird with one stone. Provider goes in the room, they have the conversation, the chart's getting populated with everything that's required for billing, for documentation, for pre-certification, prior authorization, without having to go back and repeat that, right? They're only having to document the encounter once as opposed to two or three times historically, and not to mention not having to remember, okay, what did I talk about at 8 a.m. with a patient when it's now 7, 8, 9, 10 p.m. at night? The documentation quality is going to go through the roof when we're getting this sort of immediacy and turnaround time. Patient satisfaction scores are improving across our customer base. Why? Patient wait times are going away because physicians and staff are staying on task, they're staying on time, again, because there's not this lag period of going back to the workstation after each and every visit, which, if you have to do that, you're going to get this inevitable backlog of wait times that are going to accumulate. We're avoiding that, which, again, that's going to drive patient satisfaction through the roof like none other. You want to frustrate a patient, I think, as we all know, make them wait an hour and a half and fill out a bunch of clipboard forms, that's a recipe for disaster. And then lastly, the ability to automatically incorporate everything that payers require. Payers are playing this game with AI now, by the way, I think everyone realizes that. We've got to equip our physicians to fight the same fight. We've got to give our providers the ability automatically to make sure everything that's required from every commercial payer, Medicare, or whatever it is, that every I is dotted and T is crossed. And if it's not, it's going to get denied. And that revenue is going to get delayed for 6, 9, 12 months. And the ability, again, to arm the physician with the same tools that the payers are using to deny claims. This is not a nice to have. This is stuff that we have to start to adopt, and we've got to do it quickly. Dr. Brett Rayner, he's a huge supporter of what we do. He's spoken on our behalf historically on AAOE sponsored events. He couldn't be here today. So he wrote a phenomenal post in terms of exactly how iScribeAI is impacting his practice. And I think he's echoing the things that I've said here already. It's again, the instantaneous turnaround time. This Friday clinics are the stuff of legend. He sees upwards of 70 patients in a busy Friday clinic. And when he walks out the door at five o'clock, he is done. He has nothing left to do. He can go home to his two kids. He can go to his kids' events free and clear without having to catch up on notes over the weekend. We're optimizing his E&M, average E&M coding levels have increased substantially since adopting our solution. And he would be obviously a great person to interact with after the fact if you so choose. So again, I think sort of the combination of things that makes AI work, there's a lot of different solutions out there. I think number one, that what the separators for us specifically, the versatility AI is a component, but there's a lot that goes into the iScribe platform that's important. The ability to also leverage speech to text and some other tools are critical to our workflow and the success that we've had to date. Integration is everything. If these tools don't really deeply plug in and dynamically work with the underlying EHR in question, they're just not going to be adopted at scale. I think we've already seen that. There's a low tolerance for copy paste, which only again adds to the workflow for the provider at the end of the day. And everything we do, we're providing customers with the ability to try before they buy. So if there's a trial period that we're always able to spin up for any interested customer that wants to work with us, and many of those customers are listed here. We're operationalizing AI at scale. We're now processing close to 8,000 AI encounters per day. I think we're doing this at scale as much as anybody in the industry right now. Customer roster is growing. We're extremely proud of our reputation in the market and our collaboration with AAOE. And lastly, rather than listen to me the entire time, we'll listen to the voice of the customer here before we turn it over to the next speaker. Dr. Kyle Kessler, another one of our great customers in Chattanooga at Center for Sports Medicine, I think sums it up pretty well in less than a minute. The reality of this is, this is not just me blowing smoke, but this is going to be the difference between me dreading clinic and being okay and finding clinic. This is going to be the difference with me being burned out and not wanting to do that because my burden, my clerical burden from clinic these days is nonexistent. I mean, I look through my notes still just to make sure that nothing has gone awry, but it's almost just a ceremonial thing I do at this point, more than actually finding anything to change. It's made a big difference just in my overlook or my outlook on clinic and how I feel about that. And now rather than, I mean, I'm still tired at the end of clinic, but I'm tired because every three weeks I tell them, add three more people to my template. I can do three more, add three more to that because I'm faster now, add three more to that. And I'm just ramping up every couple of weeks. I'm adding more to my template because the system is more and more and more polished. All right. Yeah. And if anybody is interested in learning more about what we do, we've got a little QR code here. Feel free to scan that. Again, there's a lot of information available to you. Feel free to reach out to us. I would be happy to chat after the fact, but Jessica, thanks. Thanks so much again for the time and the opportunity. Thank you so much, Pat. All right. Our final presenter for our use case panel today is Matt Seafield from MedEvolve, who is the chief commercial officer. So Matt, welcome. Thank you so much. And I'm excited to hear what you have to say. All right. Let's see if I can get the screen share correct. It's always the first test. Is that working? Looks good. Good. Awesome. I love hearing about all of these solutions out there. I've spent 25 years in this crazy industry, primarily focused in the revenue cycle, consulting at first, software development arena. And over my 25 years, I've seen a lot change when it comes to margin pressure. And I love all the innovations in health care, all the new tests that we have, all the new drugs that we have, all the new expensive things that we have. And what it comes down to is, are you actually getting paid what you're supposed to be paid, hopefully on time? And how many people and how many touches does this actually take you on the administrative side to get paid? Billions and billions of dollars a year is wasted on touches that generate no outcome whatsoever by the rev cycle staff, front, middle, and back. I threw a few slides in here just to frame up the problem. Your margin is being impacted. What you charge, it doesn't matter. What you get reimbursed is a reflection of the quality of the service and the quality of the administrative staff to get paid. And then what does it cost? Well, we know what the supply costs, we know what the clinical labor cost is, we know what the OR costs, we know what the facility costs are, but what about the administrative costs, right? And all of that's impacting margin. So what do we control? We control the ability to look at touches. And where AI comes in is it starts to look at those human touches and it starts to make those recommendation sets on where you can reduce waste. Some of the things that you look at a revenue cycle from the time a appointment is scheduled to the time a claim is at zero balance. Look at all of the different areas in rev cycle that people make mistakes. And by the way, even the AI makes mistakes. The RPAs make mistakes. The bots make mistakes. So the question is, how quickly can we determine where those mistakes are occurring, and what do we do about it? Metrics that really matter is health care providers can't answer these questions. And it's scary for me, being the guy that has to always live on the revenue cycle side, is that you don't know how many people, you don't know how many touches. You don't know what the true cost per touch on the administrative side is. We don't know how many bad touches are occurring. And ultimately, what it's impacting is your profit margin. You're collecting less than your 98.5% NCR, and it's costing you more to collect less. And a lot of the money is coming from the consumer now, who may or may not be able to pay their bills, certainly not up front, even with the orthopedic groups. I mean, I had a big procedure done in January. I've got a nice payment plan going on. They will get paid eventually. It's not like it was 20 years ago, where 5% of provider income came from the consumer. High-deductible plans have made a huge impact. I live in Southern California. Unfortunately, the orthopedic guys and girls down here, they just stop seeing types of patients. And that's not helping community either. If you have a great commercial plan, I'll see you. If you have a high net worth, I'll see you. But if you have this, this, or this, I just can't see you. And it's not the way we as a society want to operate. So what am I talking about human-generated data? Human-generated data is not something that's in these practice management systems. Practice management systems were never designed to capture structured data around truly what people are doing across the revenue cycle, right? So looking at things like how you get your work, looking at what the status is, looking at what the action taken based on the status, and then starting to look at resolution rates. These are real numbers. I looked at just a single orthopedic group, one of our clients, and look at the actions taken in the bottom left there, right? 95% resolution rate, but the AR representatives, 30% of their touches were on paid claims that just needed an adjustment taken. That's crazy, right? What could they have done with those touches, right? Allow more time for processing, right? Where AI starts to come in is it starts to move things to the right person at the right time to reduce touches and waste. It also starts to automate processes, which we'll get to here in a minute. So start thinking about, it's not that I worked these many claims, it's what actually happened to the claim, okay? When we start to look at context and variation among individuals, this is when you really start to see change, right? How many touches? Why is it taking one rep 1.6 touches to resolve a claim and somebody else 1.3, right? Why is somebody's resolve balance 70% and somebody's sitting at 33%, right? I'm losing money on some I need to recognize and reward others, okay? PM systems can't get you this data and we've seen this firsthand. What if I were reduced touches? This was a real example. You go to a small orthopedic group and you say, if you could free up 25% of your touches by reducing the waste, what would you do with another $5,000 a month in capacity? Let alone 50%. If you all are large organizations, just multiply that by five, 10, 20, right? It's real. But in order to determine where the waste is occurring, one has to understand where the touches are occurring. And AI won't solve that piece, okay? So when you think of the future of Gen AI, okay? You have your system generated data. That is what every AI company out there is gonna tap into. It's your basic data set. It's not novel. It's your clearinghouse data. It's your scheduling data. It's your claims data. It's C835, 837. It's your 277, 276, all of that data. What it's missing, right, is it's the top there. That's what your people who have boots on the ground who are trying to get you paid for the services that are really expensive to deliver now. That data, when you combine it with the commodity data and you run it through a Gen AI model, that's when you start to look at a lot of more unique ways around machine learning RPA. So exhausted claims, right? We've exhausted our efforts. We're not collecting it. Stop looking at it. Auto-write it off. Bank reconciliation. Why do I keep following up on paid claims, right? You start to think of the ability to take machine learning and prescriptive analytics and start building smarter work use, right? Based on outcomes, I have a higher probability of success when so-and-so does it this way. Therefore, create the queue and assign it to so-and-so, okay? We're not that far off from a future here, but the biggest gap in every medical specialty, not just orthopedics, is it's the top there. We don't have enough human-generated data to really start to profile what's actually happening in this battle of healthcare. Atlas Healthcare Partners is one of our clients, and I would say they're probably a storyboard of success because they've adopted the principles of accountability from the time that surgery is scheduled until the time the claim is zero balance. So they have deployed a workflow automation system on top of their PMEMR that allows them to measure every single touch by every single person. And so as a result from that, they're able to see not only increases in productivity, but the effectiveness of work, right? They had 25% of their people chose not to be on the team because they didn't want to be held accountable, okay? Since then, she's further increased the capacity of her teams, even in areas that we weren't even focused on, okay? You know, net collection rate, getting close to 98%, reducing labor by 31%, accelerating cashflow. They had a bit of an AR problem, right? As everyone does, right? So accelerate that cashflow, get it in, clean up the balance sheet, reducing denials, right? And you start to think about the different areas in which she's been able to improve, Heather's been able to improve her labor capacity. Billing was a no-brainer, right? 52% improvement to capacity. They're buying and building lots of ASCs. They don't have to hire people. And one of the first, I think it was Cathy, one of the first presenters today said that, right? You know, you've grown a lot, you haven't had to hire people, right? Same concept. But then the other areas start to clean up too. Coding, 36% improvement to capacity. They're actually installing our AI coding solution now because she further wants to improve capacity, reduce denials, right? Stop paying so much for coders and let the machine do the work. Patient access was another area, 22%. They're actually installing our prior auth automation solution that'll actually go out and initiate the auth and status it, right? Remove the human from the equation. So unfortunately AI is going to remove jobs. There's no doubt about it. But what we're seeing is it's the right jobs that are redundant that are being removed, okay? No one's going to a humanless revenue cycle. There's a vendor floating around out there that likes to say that, and it's not happening. And as long as you have misalignment between the insurance company, the consumer and the provider services, we're always gonna have people involved in revenue cycle. What we're trying to do with our software is we're trying to increase capacity in the teams by bringing awareness to where the waste is occurring. And when you know the waste and where it's occurring, then you can diagnose the why. Is it a technology gap? Maybe. Is it a process issue? Likely. Is it an alignment issue? People process technology most of the time it is. So when you start to think of the benchmarks, some of the areas that Atlas has seen improvement on, their front end financial clearance process has been amazing. They put our workflow automation system in place. Every scheduled case goes through 12 different checks prior to service. They're nine days out now, which is amazing, right? They're nine days out. And even with their add-ons and their same day add-ons, they're able to clear those patients within 24 hours. Pre-service collections, right? 97.5% upfront. Bad debt, very minimal for that organization. Cancellation rates, again, holding people accountable, making sure the customer knows we've done all the checks, they're likely to show up for their case. Right, the middle, right? Looking at the coding accuracy, right? Better feedback to staff. Now with the AI coding solution, they're gonna have a less demand on people and they'll even have further quality on coding. Billing lags, getting claims out the door, getting them paid quicker. Clean claim pass rates, right? We measure something called a zero touch rate. That's called no human actually had to get involved in the revenue cycle after the service was rendered. And they're operating right into the 70th percentile. A lot of the orthopedic groups I work with start somewhere in the 45%. Our best is about 81%, right? And it's a metric that we have because we do measure every touch. We know where people are getting involved. And then on the backend, obviously you're looking at net collection rate, looking at bad debt write-offs, looking at some of the higher value write-offs around no off. So I think, you know, to me in summary is technology is a must. Software continues to evolve. AI is here to stay. I challenge the industry to think about the data points that are going to be driving the AI and the science. And I believe that the lack of human data, the human generated data is the biggest risk point for organizations making heavy investments in AI because you're going to be told things that you already know. It's amazing to me when somebody actually picks up the phone, gets on with the insurance company and it determines the claim status that they got back from the bot that came through the clearing house, doesn't mean anything. This is actually why it's not paid. And then the fun begins because now you start to see the bounce effect, right? Hey, Jessica, I need help with this. Jessica's like, it's not my fault, it's Pat's fault. Pat's like, no Ty, this is you. Four touches, $5.27 a touch. How do I reduce touches? Diagnose the why. And that's the position MetaBall's taken. I love working with AOE. I present pretty much at every show there. So looking forward to seeing some of you again in the spring. If you have questions for me or you can obviously put them in the chat now, trying to speed through the end here or reach out to us. We've got a lot of resources on our website too. If you wanna get more thought leadership and content around what I'm talking about. Thanks, Jessica. Thank you so much, Matt. And thank you so much to all of our use case presenters. Before we end today, my colleague, Kathy Lada, who's our Chief Marketing and Membership Officer is gonna share a little bit of a wrap up and I believe a little bit of insight into what AOE is doing as well from an AI perspective. So Kathy, I'm gonna hand it right off to you. Yeah, thanks Jessica. And thank you to all of our great speakers. This has been an amazing day. I've participated in all of the sessions and I've just been blown away by the depth of information that's been shared. And we've shared an afternoon of all things AI from the basics of what it is with Dr. Muller to current uses and future uses with all the other speakers and even legal cases. So as you think about though, what you learned today, I don't want you to be overwhelmed. I want you to think about doing a couple of things that will try to make your life a little bit easier. So one is to share the recordings and presentations with your team and get started on your own AI journey. Some of you already have, and that's great. You can take it further, but many of you have not started it just because you're probably not sure where to start. And so for those of you who are wondering where to begin on your AI journeys, I wanted to just recap some of the next steps for those of you who want to get started. And I actually took these from the AI adoption challenges in the status quo session, and I just added one of my own. So keep an open mind when you're looking at AI. There are some definite pros, some definite cons. Overall, it's kind of a unbalanced more pro side, but there are some very serious considerations that you need to weigh legally, ethically, and just from a practicality standpoint, but just keep an open mind. Conduct research and educate yourself. Dr. Muller gave us a great presentation on like what machine learning is and what LLMs are and what generative AI is and all that good stuff. The more you know, the better prepared you will be when vendors who are all incorporating AI into their solutions come at you with, oh yeah, we have LLMs and we do NLP and blah, blah, blah. You'll know what the heck they're talking about. So just educate yourself on that, and we'll come out with a resource shortly that will help you demystify some of those key terms and things. Chaining executive sponsorship is important, but in order to do that, you need to identify key business challenges in the associated ROI. So whether it's clinical documentation, data analytics, prior auth, coding and billing, patient and staff communications, we didn't really talk as much about today, patient scheduling and workflow, staff shortages, whatever those business challenges are, pick one and focus on that. And then look at the ROI that's associated with that so that you can build your case to your executive team to get the funding that you need to try it. And also just the willingness to experiment a little. It is a little bit scary, and there are some things that you can do to manage your risk, certainly. And by developing AI policies and guardrails, you can protect against bias, pay attention to privacy and other ethical issues. You can make sure that the AI is not hallucinating or coming up with crazy stuff that makes absolutely no sense, which it does. You always need to keep a human in the loop, period. That's just best practice. Fully vet your vendor partners. Understanding the lingo is part of it, but also look at their experience, their reputation, their track record. And is it really AI? Some companies are touting solutions that have really nothing to do with AI, but they're saying AI because it's the new buzzword. So just be careful about who you look at. In about three months, we're gonna be starting a new AI webinar series focused on both extending the conversations that we started today and diving into the practicalities of using AI in your own practices. In February, we wrote and adopted our own AI policy at AOE to help govern the safe, ethical, and transparent use of artificial intelligence. We primarily use AI to analyze survey results with us, help create blog posts for webinar content, and enhance our marketing. But we'll be sharing that policy along with you that you should be able to adapt to your own practice on the business side of things, not the clinical side, but we also share a draft policy for clinical use of AI in the coming weeks. So stay tuned for that. And I just wanted to add my thanks to Jessica. Thank you all for participating in the summit. I hope you found a lot of benefit in the sessions that we curated for you. Thank you so much, Kathy. So with that, that concludes today's AI summit. Just thank you again for everyone joining us. I hope you found it to be informative. And I just wanna do one more thank you to our sponsors who supported this event, who made it possible for us to host something this extensive. So our session sponsors, again, were AI Health, Gale AI, iScribe, Social Climb, and Whiteford. Our keynote sponsor was Infinix, and our event sponsor was NextGen. So thank you all so much. The recordings will be available in the AOE Learning Center, and you'll be notified when those are ready to go. And thank you so much for joining us. We will see you the next time around.

AI summit

orthopedic practices

artificial intelligence

efficiency

revenue cycle management

patient care

documentation

productivity

healthcare operations