107. The AI Culture Shift with Steve Mika

Host Amber Bardon sits down with Steve Mika, the commercial lead for Data & AI at Fresche Solutions, a Microsoft-focused data and AI consulting firm. Together, they break down the current and future state of AI in senior living, focusing on realistic, strategic implementation that can drive outcomes.

Steve dives deep into how senior living operators can move beyond buzzwords to build sustainable AI initiatives backed by clean data, sound governance, and a modern architecture using tools like Microsoft Fabric and Azure AI Foundry.

Contact: Steve Mika LinkedIn & Fresche Solutions

Resources Mentioned: The Guide to AI in Senior Living

Find us online:
Website
Facebook
Instagram
LinkedIn

Transcript

 Amber: [00:00:00] Welcome to Raising Tech podcast. I'm your host, Amber Bardon, and today's guest is someone that I know very well. Steve Micah is the commercial lead for data and AI solutions at Fresh Solutions, a Microsoft focused data and AI consulting firm. They help organizations modernize their data reporting and AI capabilities using tools like Microsoft Fabric and Azure AI Foundry. Steve: Thank you, Amber. It's a, it's a pleasure to be here. Amber: Steve, I was just telling somebody yesterday because I did a webinar on ai, and I was saying that somehow in the last couple months I've become a spokesperson for AI in senior living. A lot of what I've learned has come from you, so. I'm really excited to have you on this show. You know, Amanda and I always talk about how you're one of the smartest people in AI that we know right now. So I think our listeners will really benefit from this episode, because AI is the buzzword, right? Everybody is talking about it. Steve: [00:01:00] Yeah, and it's very kind of you to say that. I have also been working with data solutions and AI as it kind of merged into one thing, I don't know, three years ago with senior living operators and healthcare specifically for several years now. So I definitely have seen. The sort of hype cycle come in as this huge wave where everybody tried to jump on the wave, a lot of people couldn't surf, fell off, and now we're kind of in this almost rebuilding phase with AI where it's almost the second generation where we're starting to see all these more point solutions. often called agentic ai, where they're, instead of being just this kind of broad AI tool to catch them all, which quite frankly failed because it fails to capture a lot of the nuance of individual industries. a lot of these LLMs are trading on just all the data across the internet. But that makes things tricky 'cause when you're trying to differentiate, SNF from Sunday night football to skilled nursing facilities. That nuance in that context is often lost in these large generative AI tools. [00:02:00] So now this year you've seen a big push in this kind of rebuilding of a wave with agentic ai, and in a lot of ways we're seeing that same cycle crash. And there's a lot of reasons for that, but I think it's really important to start thinking through. A lot of the early state of your AI exploration and thinking through the mistakes you made in the past with your data architecture as a whole picture, rather than just, Hey, I need ai. Think about how you need to improve your entire AI and data culture to really achieve success with any generative AI or agent ai. I think, I mean, I think that's a great thing to point out and if anyone listening to this episode has heard me already talk about ai, you've heard me say this, we said it in our presentations together, you don't wanna go out and just get AI to have AI and say you have it. we, at Parasol, have put together an AI guidebook. Amber: we've done a white paper, and then for our clients only, we have a step by step. Guidebook that's very specific to our clients and our roadmap and our processes and how to actually execute ai. we have put together basically five pillars of ai. And I really wanna talk about one of those today with you [00:03:00] primarily. But Steve, one of the things I've learned, because since we did our original presentation back in March, I've been doing a bunch of webinars and conversations about ai. And one of the state, conferences I was at, I got 40 minutes into my presentation and no one was saying anything. And typically I engage with my audiences a lot. And then finally, someone said, we don't know anything that you're talking about. And I was trying to break it down as much as possible. But, let's just start very briefly. I know that, we've talked about this before, but just a very simple, like couple sentences. What is AI. Steve: In 2025, it's difficult to really describe what is AI in two sentences, but to me there's really two. Key pillars. It's machine learning, it's large data sets, and leveraging, artificial intelligence to basically help you with predictions, help you take what's happened and predict what's gonna happen in the future. really in senior living, I think it's using smart tools to help staff work more efficiently, improve care for residents, handling tasks like writing notes, [00:04:00] predicting future risks, answering questions, and moving data between systems seamlessly. Amber: Okay. I think that is a great way to break it down. So basically what you're saying, and I can even think about our five pillars that we've developed being in two categories based on what you said. So the first is really to help with productivity, so taking notes for you or, generating policies for you or things like that. So sort of that busy work and that workforce, assistance and things like that. And then the next level, which is the other category which I wanna talk about today, is that. actual, ability to take your data and produce meaningful information and results versus just like more of a, day-to-day workflow process. Would you say that's accurate? Steve: I think that's fair in applications in the senior living industry is you're definitely seeing that breakdown happen over time. I think part of that is due to, relatively immature levels of data maturity when it comes to a lot of operators who maybe have too many source systems, haven't centralized all that data, haven't [00:05:00] created a data warehouse. And kind of that single source of truth that they use for reporting. Oftentimes what we see with our partners in the space is a lot of kind of siloed data sources. You've got your CRM over here, you've got your EHR over here. You've got a finance platform. And rarely do those speak well to each other. So when you're trying to get that second group of ai, it's very difficult. But in the space, things like note taking. Things like policies and procedures, that's relatively low hanging fruit that can be achieved much more easily than those kind of broader use cases. And what I often talk about is instead of choosing between, oh, I, you know, if we're gonna use ai, we can only use it for these kind of simple, low hanging fruit. It's looking at it as an overall culture shift within the organization of, Hey, let's deploy some tools early on that can help drive productivity, help improve care delivery, reduce staff burnout, but also look to the true value of how we can start being more predictive. How we can start, predicting move in and move out rates, how we can start [00:06:00] predicting, patient health outcomes. Really building the building blocks with those easier use cases with an investment in your data culture in parallel to deploying AI tools that get the company comfortable, with a new form of a tool. Amber: Wow. I love everything you just said. That was really well put. So I wanna touch on a couple things you said. So the first thing is the data silos in existing senior living, and that I think everybody listening to this podcast knows exactly what you're talking about. And I think this need for data, information, data analytics, the ability to take our data and do something with it to make data-driven decisions, has been a huge need in senior living for so long. And a lot of the challenge has been, what you said is that data's all over the place. It's in spreadsheets, it's on paper. Most communities I've been to still have paper charts of some kind, you know, at least for like the physician order side, or the physician side. it's very common to see a lot of financial processes still in spreadsheets. if you're listening to this, raise your hand if you're doing your budgeting and spreadsheets, because I see that almost everywhere I [00:07:00] go. I honestly don't see that changing. there's no major enterprise system out there besides maybe Yardi that can bring all of those together in one centralized system. And I see that issue continuing. But I see the, the, the AI side and the data analytics side is the way to bridge that gap and bring the data together. So let's talk about how do you do that? what would be the first step? Steve: you know, the first step is really deciding on what your path forward is gonna be. And what I mean by that is there's this classic debate between build versus buy. We see that debate common in senior living, and it almost always goes towards the buy solution. It's difficult to build on your own. there's kind of two paths when you're deciding to build it's ramp up internally, hire new staff, hire folks with business intelligence and data analytics experience that can help build your kind of data architecture foundation or work with a partner with a lot of experience. The buy path oftentimes is an easier path to success. You'll have, preexisting integrators, but the flip side of that is you're also [00:08:00] beholden to that path. You're really making a decision early on. I think one of the nice things about the build path is with some of the latest tools that exist in the world. I'm a big proponent of Microsoft Fabric because of the open architecture. You don't necessarily have to go all in on build, you can use a partner to help you, but you can also just kind of dabble. You can put your feet in the waters fabric allows you to say, Hey, you know what? I want to take a couple of my data sources. I wanna take that spreadsheet, that work, that behemoth of a workbook with decade plus of macros built in with formulas that nobody really knows who wrote anymore. You can start to ingest that data into Microsoft. Or there's other platforms out there, snowflake and AWS. You can est that into a data platform that then helps you start to build reporting, start to own that data. Start to merge data that comes from your EHR and your ERP and your CRM into one common area so you can build more modern solutions. And even if you still wanna use Excel as a workbook, that's great, but it's much better to Connect [00:09:00] Excel to a data source that's automatically refreshed, then just be entering in data into Excel every week. We all know the pitfalls there. If a formula's bad, your data's bad. If you enter the wrong data into a certain column, you can throw weeks and months of work out of the window for that one error that happens. So it's always better to find solutions that automate a lot of these mundane tasks and refresh from the source, refresh from your Yardi environment into fabric. And in fact, just had a conversation with Yardi. They are going all in on fabric as well. They are definitely pushing a lot of their clients to say, Hey, you know, we have all these platforms, we have all these preexisting solutions that help me to need, but we still know that folks wanna own their own data. And so they're making it quite easy to bring data into a fabric framework, build some kind of base reporting, but then you can take that foundation and grow from there. Amber: Okay, that makes sense. So basically I think the first thing is to get the systems in place that they wanna use if you're about to make a major system change like your EMR do that [00:10:00] first. Yes. And then you're saying to build this, data warehouse, So you're saying to build that, using a vendor like what your company does or others out there that has experience doing this. Steve: I think both paths have a lot of value. I know a lot of organizations that have decided to hire up and say, Hey, you know what? We're gonna look for folks with experience building these solutions in healthcare. Bring them on board, and truly own the solution internally. And there's never gonna be any pushback from a vendor like us to say, that's not a great idea. I think ownership is always important. That said, using a vendor, using a partner with a lot of experience in the industry and with the technology can really help you lay that foundation and sort of train you as we're building that foundation. And so you're starting to learn the intricacies, the reality of something like fabric, it's two years old. You go online, there's not a lot of guides that really help you with best practices 'cause those best practices are constantly evolving. So when you have a Microsoft partner that works closely with the fabric development team, [00:11:00] you really get insight. You can really get that kind of inside track. The reality is we have resource constraints in the industry. We are constantly struggling with both staff struggles and financial struggles to really map out what the best path is. I just want to emphasize that I've seen a lot of operators go down this kind of. Let's just use a vendor for everything path, and then you just end up with all these disparate systems that don't talk to each other. And so it's really important when you're going through that onboarding process to be very in depth, really ask pointed questions about with your vendors of how they integrate, how they work with your current source systems. Where does your data live today? How do those systems speak to each other today? And what do you want that architecture to look like in five years? If your vendor doesn't really have an answer for that, you should take a big pause right there. Amber: So just to kind of break it down a little bit from what you said, for some of the listeners who, are starting at more of a basic level, You need to identify a partner or, hire people who have experience. I have seen it same as you, where we've seen some senior living communities [00:12:00] try to hire internally and they don't have the right resources and end up spending a lot of money and time going down a path that doesn't end up being very successful. And then. Once we get your data in a data warehouse, what can we do with it? I wanna also touch on, you said outta the box solution. So I wanna just differentiate before we get to the question I just asked, the difference between going with like an out of the box static dashboard type solution, of which there's several, we've had quite a few of them on this podcast. Versus building your own private AI data solution. Can you just talk about the differences between those two? Steve: So, and again, I don't wanna sit here and say, even on the hiring topic, like I'm never gonna tell somebody that's a bad idea. I do think you've brought up a good point that if you have leaders that are making hires for technology, they don't quite understand. There are definitely pitfalls to that. You're making a big investment in one or two individuals. That's a lot of work for one or two individual. So there is [00:13:00] benefits, when you work with a partner that we have a team, we have data engineers, power BI specialists, AI engineers. You get to basically have bits and pieces of different employees rather than making a big investment in one. But again, I also think there's value in ownership and really understanding the technology. Deeply in how it impacts you. Now when it comes to like where to start. So you've made the decision. You're either using a vendor that puts these static dashboards together, or you're going the path where you're using a partner that really wants to kind of create these solutions customized for you. There's a middle path. Now again, I mentioned Yardi has this solution now where. They can kind of give you a starter pack of reports and then you build from there. But once you have your data landed in a data warehouse, once you've gone through that process of, integrating different data sources, cleaning that data up, getting that data to a place where you have confidence, it's been validated. The financial numbers tie out, all the, the population health data is in there. That's really where you've unlocked so many potential. use cases. And [00:14:00] so the first and most obvious one is gonna be reporting, within a framework like Microsoft Fabric, power BI is included. It connects directly to that data warehouse seamlessly. There's no future costs in terms of like moving the data or transforming the data. And then you can build reports to your heart's delight. It's best to start with kind of a framework foundation, but there's so much customization that can be done. Typically, the path that we like to see is, analysts or even just citizen developers on individual business units, they can kind of help. You can truly make informed data decisions and hey, if you wanna use an Excel, you can just connect to that data set and just see it in Excel. But then that data refreshes automatically. And again, you move away from that kind of, what we call fat fingering, where you might have columns or data that's entered incorrectly. What that does really unlock for the long term is AI use cases. And so once you have all your data centralized, you trust that data. The sky is the limit because then you can bring in all your policies and PDFs, throughout the company. You can bring in your emails, those paper documents that you mentioned [00:15:00] before, you can scan those in and use documents intelligence that exists within the framework to scan those and read those and integrate that within your larger data. So then you can start standing up. AI agents that can help process claims or help with customer service. You can stand up generative AI tools for internal knowledge management. You know, maybe you have an HR bot, maybe you have a policies and procedures bot. You might have, communities that are in seven different states. Each state might have different regulations. Maybe you have staff that spreads across different states. You can instantly go in and differentiate, Hey, what are the policies in Illinois versus the policies in Oregon? And really dive into that easily. And that's honestly, as we've seen this whole AI kind of wave of crash and wave and crash cycle where we're consistently seeing value is what's existed for a decade plus in machine learning, where you take all this data that you've got at your fingertips, you've spent the time and effort to clean that data up. Now you can start being predictive and that's where you can easily see that kind of path. The ROI that's a little bit less clear when you're talking about agent AI and generative [00:16:00] ai. Hey, I can start predicting my move in move outs for the following five years based on trending data from the past 10. You can really start to see the power that could be unlocked there. And then there is one more part that I think is less talked about until this year that I know you and I are both in, really interested in RPA robotic process automation. It's taking new forms, it's getting a little muddy in this AI space, but there's a ton of value in automating a lot of these mundane processes. And it all starts with that first bit. It's where is your data living? Is that data clean? Are you confident in that data? If you can't say yes and you don't know to all three of those questions, all that AI downstream potential, it's really all for Naugh. You need to start with that foundation of where does my data live? Do I trust my data? Is it at my fingertips? Amber: Clients who are starting to go down this enterprise data strategy path, you know, they should think about, do they want a static dashboard that will just pull information and display it to them, or do they want that plus the ability to do this [00:17:00] machine learning, this generative information, this querying. And you know, personally I've been advising clients to go that second route because that's the direction that they need to go. But that being said, I do wanna mention that I think a lot of those static dashboard companies out there are starting to build this into their products. So I think it's good to look at both options, but also understand like, where is the future of technology going? And you don't wanna be stuck with a product that can't eventually get you there Steve: Yeah, and that's where when you're reviewing vendors, it's really drilling into what your future plans are. 'cause I feel like a lot of the times I've worked with operators when making these kind of, different siloed source system decisions is it's addressing a need or it's addressing a problem. But very often it's not looking forward to what does this look like for my organization in five years if I'm going down this path. I really need to think about the financial and the data implications of what that looks like years from now. And again, I'm never gonna sit here and tell somebody not to [00:18:00] use a preexisting solution because I'll be candid, that is almost always going to be more affordable than building your own solution or using a vendor to collaboratively build your own. Enterprise data, architecture ecosystem, it's, costly, especially on the front end. It's more costly, but those downstream costs, that's where I think that's more debatable on what's truly a larger cost. When you own everything, at least all your, you know, you're not obviously owning your own EHR, you're not gonna build that. But when you own all the data that comes from your source systems, your systems of truth, and then you can really trust in all that data, merged together, creating, reporting, creating Excel workbooks like. That has a ton of value. It's that knock on effect of, okay, now what ai ml, process automation, it's those tools that you really become free to hire up, you know, hire a vendor, whatever, to be able to build those out, custom fit, bespoke for your needs. And this is the kind of the thing that I've seen in the past with static dashboards. [00:19:00] Everybody gets some value outta those static dashboards, but every single operator I've ever talked to looks at their organizational data slightly different in a way that static dashboards very often don't properly represent. So that is where you start to have that internal debate of does it make more sense just to build these ourselves? Amber: and I think, people, when I talk to 'em about this, they're thinking about the cost side and how much more expensive this is. But the other things you can do with your data when you have that ownership, people are using chat GPT for everything right now, maybe people know this, but I did a webinar yesterday where a lot of people didn't know. They didn't know that all the information is in the public domain. And so you can build your own internal, chat, GPT, like tools that are private to you, that would add a lot of benefit in time savings. To your organization. So, I think there's other applications outside of just the data piece. Steve: And that's a really key point. So one of the benefits, and I'm just gonna take this scenario 'cause this is something that we've done with a customer recently. So we did the dirty work, we brought their source systems, we brought the CRM, their EHR, their [00:20:00] finance platforms, their manual workbooks. We brought all this data into fabric. We cleansed that data and now what we built reporting, that's great. That's kind of standard operating procedure, but it was that next layer of AI use cases that was the most valuable because within that Microsoft framework, there's a product or a platform called Azure AI Foundry, and essentially that's taking all these l lms, so that's taking, four oh from chat GPT that's taking even deep seek, but disconnecting it from the Chinese backend and it's giving you these models. So that then you can take them and put your data into those systems and run with it. And so what that really allows you to do is basically put reins and put a box around something like chat GVT, so you're not training the internet at large. When you're deploying these internal solutions, you're parameterizing them. So they're only available. Within your source systems, they're only basically they're taking that training but then putting it in your box and it's only able to go and look at your data that's available to it. So there's a couple really cool things that are a knock on effect of that. First of all, it's the compliance and the regulation [00:21:00] side. You can have confidence that, whatever Joe from HR is typing The co-pilot isn't just being used to train, the internet isn't just being used to train the LLM more broadly, but it's also secure you still own that data. You're being compliant with whatever regulation pops up in the future 'cause it's still a patchwork. But it also helps you control the outputs. So when you're an AI foundry, you can control temperature, which essentially means how creative responses are gonna be. If you just want something, policies and procedures, you don't want the LLM getting really creative and spitting out something that you have very little confidence you wanna say, no, I wanna see what's in paragraph three of page 19. I want you to cite that. And I just want to be able to access that data quickly. And then I wanna be able to cross reference that against, patient 19, to be able to really get to the core of the confidence level with ai. And to really be able to build your own tool internally. And it's not as, you know, this sounds like a super expensive process, but we've had employees where once they have their fabric environment stood up, they just loaded a bunch of policy procedures into AI Foundry and some of their structured data that kind of [00:22:00] like modeled the EHR and finance data. And all of a sudden they were able to internally just use this kind of internal process where they had a few core individuals, about 10 people working on, the feedback and making sure it was right. And then voila, they had a tool. With very little outside investment, just because they made sure that that data foundation was sound, the intuitiveness of these tools that are available to us now allow us to do things just like upload that unstructured PDF document and be able to query against it. Amber: Yeah, it's exciting. I mean, I think organizations who aren't considering or thinking about this are gonna, and it does take time to build and you have, there's a lot of pre-work you have to do, but I think they're gonna get behind other organizations that are ready to move forward with this. But I wanna switch topics a little bit, and I wanna go back to something you said maybe like 10, 15 minutes ago. So we're started talking about what you can do with the data and the directions to go. But the other thing you mentioned is the culture change. And I wanna. Talk about that. You know, you mentioned that we have, and I just had this conversation this morning with a client that is starting down this path and I was talking about everything. We're talking about what can be [00:23:00] done, and they said, well, we're just worried about, you know, people not understanding how to use these tools or how to go into a chat bot and ask it a question. They just wanna see the data. And then of course there's always, you know, is it gonna take my job? That fear. So let's talk a little bit about how can you get your organization ready to embrace, or start down this path of culture change with technology? Steve: Yeah, and generative ai, I kind of think of as the gift and the curse for the IT budget because you have got a lot of organizations that for the first time in a decade, leadership has bought in. They're like, we need ai. We don't really know what that means, but we need it and it, here's some money. Go find us AI tools. And it is sitting there like you have neglected us. We have been in a dark corner in your organization for the last 10 years. You've said no to every big budget expenditure we've requested, and now all of a sudden we're supposed to roll out these advanced AI tools. When we can't even trust our data when we're still operating our entire business off Excel. Speaker 3: This is so true. I've had a same conversation so many [00:24:00] times in the last couple months. Steve: And, but it's still a gift to me because at least the conversation is now happening and hey, you know, if it took AI to really give leadership a view into what the potential is and why the investment in your data infrastructure is a valid one, I'll take it. And so that's the first culture change We want ai, it's not a switch. You don't just flick that switch after years of not messing with your wiring and expect your entire light system to work properly. That is a rebuild. You have to tear down the walls a little bit. You've gotta go in and say, Hey, how do we rebuild this foundation? And part of that is a culture shift. Part of that is understanding that, hey, we may be a little bit technologically immature, but we are going to change that. We are making decisions today. With five years in mind, 10 years down the line in mind. And that all starts with data. And you'll hear this from basically anybody in the industry. And if you don't hear that, please run away from that person. Everything that you want in terms of an AI output all [00:25:00] starts with trusting the data, Basically downstream from dirty data is gonna be dirty. You're never gonna be able to trust it. And if you're talking about building confidence and using new technologies, the first way to destroy that confidence is to have one of your expert, nurse practitioners go into a copilot, ask a question they know about and get the wrong answer. They will never go back again. And so. Instead of just rolling out tools willy-nilly, be very deliberate about that process, about testing with a small group first, and I mentioned that previously. It's always important when you're thinking of these new tools, be narrow first with ai. Really pick a narrow use case that drives value for you, so you make sure that that initial rollout is productive. You bring in some super users, some power users, some longtime experienced staff members to really give you that initial feedback that then helps the iterative rollout when you roll it out to the entire organization. Amber: Yeah, I mean, it's just like any other system, right? You don't wanna go out and just get it because you know, a guy, you met somebody at a conference, you heard about it, you know, you need to do your business. Case analysis, you understand the use case. You need to appropriately plan, you [00:26:00] have to have project management, understand your deliverables, you know, training, communication. you know, just like any other new technology, you need to go through that process to be successful at the end of the day. Steve: Yeah. And it's very difficult. And I do sympathize a lot with vendors because there's so much noise. Every process automation or workflow automation vendor now is just calling themselves an AI company. We've seen this big shift where everybody's marketing department is like, I don't care if we actually have anything that's ai. We need AI enabled features on our marketing material. We need to tell everybody when they come to the booth that we're all about AI because no one's gonna buy our, system if we don't start talking about that. But that's where it takes a lot of diligence to really work through what the difference is. It is that workflow automation, which is kind of the most value. So take the scenario. You've developed a call center AI agent, and this is what I really see is like a lot of the things that we've offshore over the last five years. A lot of AI can really help with that low hanging fruit, so AI automation of that early customer service interaction. Oftentimes, We're frustrated if [00:27:00] it feels like we're talking with some computer. AI has improved the experience to at least get you to the point of if it needs to go to a human, it's been well vetted, but there's also the automation off these workflows. Hey, you know, we've got this claim number, they've got this insurance. We can start to automate some of the processes that occur in the backend with a agentic. AI flows. There's so much potential there, but in the same vein. They've had robotics process automation for a decade plus now they've had workflow automation for years. Now, a lot of these workflow automation vendors are just saying, Hey, this is ai. It really requires a lot of interaction. Understand, well, okay, what reasoning is occurring on the backend that truly makes this an AI system? Versus this is just basically a tree that it's following. If this happens, then that happens. for operators in the space, it's so difficult to differentiate between the fakes. And the real. a lot of that is why I kind of lean towards build because at least you can see what's being built on the backend. You can control things like, the creativity of my AI tool. And you truly, [00:28:00] the ownership. A more cultural buy-in. There's more dedication internally 'cause the staff feels like they're partially responsible for building this tool. But you also can see the process. You can see how the sausage is made. So if there's something nasty in the sausage, you at least know what happened on the floor. Amber: I really like your point about how vendors now have to shout out AI to get attention, and we've even experienced that. we have been told sometimes we're not innovative enough because we're not shouting about AI the loudest, even though I kind of think we are because we talk about it all the time. They perceived that someone else was shouting about it louder and that that's innovation. But we can't forget about the basic needs. We still need these systems to do. I think AI should be part of the conversation, but I just helped a client pick a new document management system and at the end of the day we needed to do, x, y, z basic functions. And we did ask about ai. So we asked like, how is AI incorporated in this? And is it, is it learning or is it just using like process automation? you can't like just get overwhelmed [00:29:00] by this innovation side and forget about those pieces as well. Steve: Yeah, and that's such a poignant point in the space because oftentimes what I've seen is the folks yelling loudest about AI are more often to be closer to that fake side of the spectrum rather than the boring and mundane vendors that are like talking about data cleanliness first. the reality of successful AI is it's a lot more boring and less, I don't mean to be offensive, but in the senior living space, I very seldom see operators that are really far to the maturity side on that kind of technology curve. And so really what you wanna be talking about is the boring stuff early. And if you're not having those conversations, there's a problem. So I appreciate less fizzle and more sauce. Like really, what is the substance of what we're trying to do? Get to AI needs to be a part of your roadmap, but it needs to be a pragmatic roadmap that achieves success instead of just throwing money at this and then abandoning. Because every year we're seeing a Gartner article say. 40% of last year it was generative AI use [00:30:00] cases, not making it to production. Now they just released one that subs like 38% of agen AI use cases. Don't make it into production. You can just think about the amount of investment that's been made there that has just wasted and really go back and say, Hey, what if we had done this differently? What if we had just achieved some basic foundational needs first and to really trusted our data? Because there's so much that comes from just trusting all that data in your report. And so many operators I've worked with, nobody really knows what the single source of truth is. Nobody really trusts the data, and that just creates so much cultural disconnect. So yeah, you're gonna throw AI on top of that, and then people just don't trust the AI that's sitting on top of your ugly data systems. Amber: Yeah, I mean, I've seen everything you're talking about over and over again and at the end of the day, it's just like any other new shiny system in senior living. You know, I could talk about resident engagement or falls management and we could really apply the same thing that someone met a vendor, they did a great sales presentation, they went out and bought the product, and they put in their resident engagement app, but they never identified who's [00:31:00] gonna manage and update the data that is in that app. And they stop using it or they buy a false prevention tool, they don't really understand what it's supposed to do, or they don't have the wifi infrastructure to support it and they don't use it. You know, I think this is just a, a, endemic, is that a word? Endemic? Speaker 3: That's a cover. Amber: Yeah. This is a problem we see in senior living in general where they don't have the processes, the people, maybe project managers or a steering committee or something to make these organizational wide smart decisions versus siloed department decisions and AI's no different. It might be a bigger mistake than others in some cases. this is something that I think is more of a bigger picture that, we need to think about from every system perspective. Steve: And I've been in those booths at ASHA and all the senior living conferences across the country, so I know like A, we're very convincing and B, it's hard not to be tantalized by the easier path by the, Hey, this is, a [00:32:00] couple thousand dollars a month. I don't have to make a deep initial investment in this. It's starting to understand the impact of all of those decisions that happen over years and not months, and then starting to put all those decisions together. That's where the cost differentiation really comes. I kind of liken this often to like leasing a car versus financing a car, versus buying a car. Like I think everybody in the financial space will tell you. The best thing to do from a financial perspective is to pay cash for your car. Don't have a payment, don't pay interest. The middle ground is get a low finance rate on a used car generally. And then the third path is a lease car. I see oftentimes that senior living operators go with the lease path because it's the path of lease resistance on the front end. Oh, my monthly payment is only X, but what are you paying in the backend? To me, that path on the right side where you're truly taking ownership, you're making a deep investment, hey, it stings when you have to take all that cash outta your bank account and put it into one thing and you're really just like, Hey, I wanted to buy all this other stuff, but you own it. [00:33:00] There's so much to be said for the value of ownership over the long term, and that's really how I like to look at this from a senior living operator perspective is shifting that perspective of, it's just easier to buy all these point solutions. 'cause it's not really, when you think about the long term impact of that. Yeah, and there's a, I also don't wanna disparage 'cause there are so many great point solutions and there are, like, I've seen some of the fall detection software with AI integrations. Amber: I'm not. Well, right? But you don't wanna just go, by the first person you meet who's talking about it either. So either way, I mean, just you have to be thoughtful about your selection process, your requirements, and also in selection process. Steve, every time I talk to you, I learn, I learn something new and it's always such a great conversation. Is there anything we haven't talked about yet that you wanna share as any last words? Steve: I think we typically cover everything that, really to me is driving the conversations now. I will also say. I've said some things that might be disparaging about senior living operators. we're thinking about things the wrong way. We're stuck in the past, [00:34:00] technologically immature. I also say I've seen a significant culture shift over the last two years, and that's where I say the gift in the curse of ai. I think that gift side is this is forcing operators to have those conversations, to really start thinking about. Build versus buy. Do I own this? Do I use a vendor for it? there's multiple valid points and valid paths to take, but I wanna give some kudos to leaders that are really starting to think forward. I think we all know with this kind of generation that's about to move into retirement, that there is a lot of potential to do our jobs better, to deliver better care, to have staff feel less burnout, to have better tools to basically have a better, user experience, patient experience. Sorry, the. The, resident experience, but also families like how you can interact. There's so much potential to leverage technology, to improve the reputation of the industry. And I'm seeing those decisions being made now across the board. I'm seeing so many leaders in the space pop up with really important decisions being[00:35:00] Speaker 3: you just lost your audio. Steve: Sorry, my mic just fell down. Amber: Yeah. Steve: Should I say that again? Amber: Amanda, do you think we can just, Steve: I we can either cut it or you can say it again. Amber: I was, I was gonna add, so I mean we could cut it after the part where you say like, this is the opportunity. 'cause I was gonna add something about that really quick. Okay. So. I think, I mean, I think we have these like watershed moments, which have really changed, the technology culture in senior living. So COVID was a big one. So, you know, I've heard people say that technology was the silver lining to COVID because people, both staff and residents were, forced to make changes with the way they think about technology. And our workforce practices have changed for sure since then. the cybersecurity industry requiring cybersecurity measures to be in place to get policies was another huge one. We saw a huge increase in people taking that seriously. And I think AI is another big one. it's, forcing people to think about technology in a different way. Steve: Yeah, and that's a great [00:36:00] point. I think what we've seen in those two, three years since the pandemic tailed off is a collective sigh from the industry. It was like, okay. We're back on level, ground now or on like even keeled. Let's start being proactive. Let's make our technology decisions now future forward. Let's really start thinking about the downstream impact to our employees, to our care delivery staff, to our investments, and start making the best choice today. Rather than, Hey, we just need to fix this problem. Or, oh, like we're kind of reacting to a crisis. Let's start planning the foundation to making those decisions that will be productive for us for years down the line, and again, I've been a little disparaging in the industry in terms of the technology maturity curve, quite frankly, a lot of operators are quite a bit behind. But I also will give kudos. I have seen a lot of leaders pop up with a lot of great things to say and a lot of guidance to achieve our outcomes in a much better way and address the incoming retirement boom, that's gonna happen. We all know there's a ton of [00:37:00] potential. Let's lay the foundation today to make sure that we better care for our aging population, that we make better investments, that are more productive for us internally, and that we're not spending all this money down the line fixing the problems that we've made technology wise. Amber: I love it. Let's do it together. Steve: Yeah. Amber: Well, Steve, where can people find you? Steve: so you can find me on LinkedIn. I love to interact with folks on there. I'm always impressed with the social media community. you can also find me on email fresh solutions.com or omni data.com. We were acquired recently, so you can still find us at the old site. I would actually recommend going to Omni Data to like kind of get a true understanding. Or reach out to you guys 'cause we love to partner with Parasol Alliance, fantastic relationship. And we have worked in collaboration on multiple customers and had great success. So if you ever wanna just reach out to Amber and, and let's schedule a collaborative call, we love working with these folks. Amber: Great to see you. Steve: Thanks.