AI in Action: Use Cases, Data and Workforce Trends - Featuring a Live Panel Discussion with AI Leaders

workforce AI event panel. So we have assembled a group of subject matter experts in artificial intelligence across multiple sectors of our economy, as we learn about how artificial intelligence is transforming the world, here in the state of Wisconsin and around the globe, it is my great pleasure at this time to introduce our four

panelists. Number one is Dr Brian Kay, who serves as Chief Strategy Officer for Rogers behavioral health working closely with executive leadership to define and implement new operational strategies with extensive expertise in data analytics, process improvement and machine learning algorithms. Next to Brian is George Forge, who is the Senior Vice President of client technology and product

development at quad. George is responsible for crafting quad strategy to identify, build and deploy industry leading marketing solutions. He also oversees quads research development and implementation of artificial intelligence across quads. Offering next is Nathan Lasky. He is the chief technology officer at concurrency, a leading consulting partner for data AI security and digital operations in the Midwest. Nathan is a 14

time Microsoft MVP. He's also a three time Iron Man, Completer, as I learned this morning, and he serves on several boards. Last, but certainly not least, is Sarah grooms, and Sarah serves as the chief administrative officer of banking strategies at wind

trust. Sarah leads a wide variety of special initiatives to support strategic priorities across the company, especially those involving innovative technologies and solutions in the highly regulated world of finance, we are going to have a really lively conversation with these folks. So thank you so much for being a part of this panel as a way of introducing our panelists to the audience and learning a little bit about their organizations. And I'm going to start with Brian and

we'll move across the stage. Can you share with us a little bit about your organization, your role, Brian, and where you are on your AI journey, and then, before you're done, I also want to know what are your three go to AI apps that are loaded up on your smartphone and you can't answer chat? GPT.

in the art of therapy. What we're trying to do is, how do we move it from the art to the science? And AI is an excellent way of being able to do so. So we're mining a lot of our data sets to provide more personalized medicine. How do we build efficiency into our workloads? And as well as, how do we reduce administrative burden or clinical burden with our therapists? So I could speak about a few of those later. The three apps that I got loaded up.

Personal favorite is one called cleft, which is a transcription app that gives you markdown style notes. So when I'm driving in the car, it's fantastic. A gimme app is google assistant. I have my home all smart, automated. And then the last one is called research rabbit. So if you need to look at anything for peer reviewed publications, it's a great way to start

majority of that time. We're actually now the 14th largest marketing agency in the world, and we've been on that transformational journey. The short version of it is, print is a form of media, and as that form of media obviously had some organic declines and challenges globally, we got on that bandwagon and pretty aggressively diversified our offering. So in terms of favorite apps, my first one is

perplexity. Love it. If you find yourself like we probably all do googling stuff, I would submit to you that it is a far better Google because it will answer your question, as opposed to give. A bunch of paid search responses, it will actually answer the question directly, similar to a chat GPT, but importantly, it will index the sources of the answers. So that is a phenomenal tool that kind of takes out some of the guesswork that a lot of AI

platforms will give you. Second from a media perspective, mid journey is phenomenal from a content standpoint, there's some some challenges there in the legal space, but I'm not going to get into that. And then the last one that we've been playing

with a lot is called Cassidy. So Cassie, as we're a big company, we have 13,000 employees, and there's just many different types of employees, but if you kind of take each piece and you say, Okay, what is the content that these employees are constantly interacting with how do we make that more accessible? One of the classic examples for us is RFP responses. So we've

got about 2900 clients. We work across tons of different industries, and so how do you build an AI assistant that can help give the best answers for that particular industry that are current and relevant and need minimal, minimal people, that, frankly, takes a ton of our time as senior leaders to make sure that those RFPs are right. So it sounds trivial, but to your earlier point, it just kind of, you know, efficiency in your day to day. It's, it's a phenomenal tool, absolutely, by

So I spend most of my time helping organizations to look at the mission of their business and think about how it translates into leveraging AI to be able to make that real in the mission, help them to be able to gain new revenue or operational savings or better customer results as as a with AI as an asset. So it's really, really interesting right now, because most organizations don't know where to go with AI. Their executive teams know it's a thing. They know they're

interested in it. They know it has to be an asset for them in the future, but they're not sure

what to do with it. And we have the great blessing and being able to help those organizations take advantage of AI in a responsible but also assertive way, one of my favorite apps, boy, I'd say copilot, is probably the one I use the most often, and in particular, the game changing component of that for me has been meeting transcription, just not having to take notes in meetings, and being able to ask it aspects of the meeting has been just amazing for me, like I wasn't a

note taker in the first place. And there was always someone who had to take notes in meetings. There's always that like job that kept them from being able to be engaged. And for me now, that discipline of turning it on and then enabling that and even sending it to people afterward has been has been awesome. The second thing is training peaks. So if anyone is an athlete and they want to be very data driven in how they compare themselves against how they can get better

at their particular sport. For me, I had just done a bunch of triathlons and training peaks was an asset for me to be able to say where, where do I score against my competitors, and how can I get better at different aspects of my training regimen, like, when should I work out more? When should I work out less? And then the third, and this has been fun. This has been one that's been around for a while. Is this plant identifier

on my iPhone. I love it because I've got all these, these different plants in the backyard, and I don't know which are good plants and which are bad plants. My kids know, but I don't. So I'll go up to it and scan it, and it'll tell me what I should be taking out with my my tools, and what I shouldn't be taking out. Love

company. So if it's if it's if it's financial services, we probably have something to do with it, either here or in Canada. And the the state of AI at Wintrust is really interesting, because we have to take a really regulated and measured approach, especially from a data perspective, privacy perspective, and so on and so forth. And so we'll get into that quite a bit. My favorite apps. So I'm kind of a geek when it comes to images. I love photography and art and all of

that kind of stuff. So one of my favorites is actually lensa, and it lets me wipe out things that are in the background. So if I, you know, take a picture and it's like, Finally, both kids are looking and smiling, and there's some random over here, you know, they can just wipe. A person out, and I've done that quite a bit. I like copilot too, and that feels like stealing.

But the reason I like it is actually not at the work front version, but actually, if you go in the app, very similarly, if you're trying to say, Oh, I don't know, design a tattoo, you can kind of give it prompts and start seeing images and start watching it change and add things to it and everything else. And so if you're, you know, I'm a, I have all these vivid imaginations, but I'm a

terrible artist myself. And so I can start telling it what I want something to look like, and then give that, you know, to say a tattoo artist, or to marketing, or to, you know, whatever it's like here, this is what I want

this to look like. And then similar on the transcription services, I use a, I use one called plod, P, L, A, u, d, and it actually will take the it listens to a meeting or does whatever or phone conversation or whatever you're trying to be like, No, I really want to capture what we're talking about here so that we can then bullet back and then send it back to people. And so depending on the if it's not a work thing and copilot is not available, plod is another really good

topic. I'm going to begin this series with Brian, once again, Dr, Brian Kay, have you used artificial intelligence to solve a problem that you would not have been able to solve without artificial intelligence? And also would love to hear if you've got some of those distant, early warning signs that one of your AI projects is going off, off the rails and needing to start over. Yeah,

cross country. One of the most important things is, how do we create treatment plans that are very personalized and set up so they have the best possible

treatment response? A few years ago, we set it up first as a research study, so going through our institutional review board, since we are such a regulated industry of saying, if we could apply artificial intelligence and especially deep learning algorithms on day number one, can we predict if an individual is going to respond to treatment and if they're not going to

respond to treatment? What are the different levers that we need to pull as clinicians in the treatment team to make sure that they have the best possible outcomes by setting up those algorithms and leveraging different data from patient reported outcome measures, we had a positive predictive value of around 92% so what that means is basically 92% accuracy. We were able to predict if that individual was going to respond to treatment on their first day.

If they weren't, we were identifying 40% of patients who fall into the not in response category of the 40% by identifying what levers we were able to pull an additional 80% responded to treatment. So it would not been possible if we didn't employ the use of AI, because there's so many different variables and so many different correlations underlying in the data that a human wouldn't be able to do so.

And when you think of the amount of data that's collected in the medical record on someone's treatment stay, it is astronomical. So the pure human putting it all together is just not possible. But having AI as a tool to assist our clinicians have been a phenomenal aspect to do so. In terms of warning signs, how we set up all of our AI governance, we have a board that mirrors the same composition of an institutional review board. So we have

different disciplines. We have actually a lay person on that board who has no idea what AI is, and they could provide some different ethical perspectives. And then we review our projects based on our evaluation metrics at a very consistent standpoint. So we want to understand how is the positive predictive value? What is some of the qualitative data coming from our clinicians? So we're always looking at it and making sure that we're on top of it, especially

1520, years at times, right? So, you know, how do you tap into the expertise of someone who may have fixed that piece of equipment super efficiently, but they were a 40 year employee, and they're since retired. So you know, that was a really great opportunity for us to tap into that data. Yeah. And it's

called maintenance GPT. I think it is, is and, you know, it's just amazing that you think about a community that, frankly, you know, are maybe not as welcoming to the technology, but they're finding that, hey, if I can just go here, and I can ask a question, and I can get the knowledge of the dozens of people who have fixed this piece of equipment in the past, and then I can do my job more efficiently. So it's, it's super helpful from a training

perspective. You know, at the end of the day, we measure our equipment on uptime. And so there's, you know, there's incentives built into to use the technology as well. We subsequently did a very similar thing that's not quite ready for prime time yet, but we're testing with our safety data as well. So similarly, decades of safety data that's highly regional. It's highly subjective on weather conditions, just other patterns of geopolitical things that may happen in that

region. So how do we bring all that information together and get that information to our leaders as efficiently as possible? And then what's really cool is, once you identify the things that are most likely to occur, creating preventative measures on the fly is so easy. And then on top of that, we have, you know, dozens of different languages that are spoken across our global platform. So Polish Spanish, there's many different dialects

of Spanish, right? So our leader of Europe, for example, is from Provence, right? So his Provencal French is very different from other areas. You know, it's very different from Canadian French. So like, how do you pick that up dynamically and build these things just super

efficiently? So that that's been a huge win looking more, you know, forward, looking towards, you know, where quad is going versus where we've been a little more on the marketing services side, our primary differentiator is that we're an integrated media company, and that's not very common in the in the media industry. What I mean by that is there's usually different agencies that take just a

portion of your media spend. So there might be one company that just does social media, another that just does search another that just does out of home media, another that just does print media, we do it all, and that's a great opportunity for us to use data to say, Okay, how do we look at a marketer's goals and how do we understand what media mix is gonna be most effective, either at a regional level or at an individual level? We've got tons of data to support that, which we think is

a big differentiator. It helps us create better content that's more strategic, that's faster, and frankly, it's cheaper as well. Awesome

about themselves. An example that would be, I'll stick in manufacturing, but this is this fits in professional services and other spaces too, which is understanding the needs of your business from the market, and then being able to translate that into whether it's inventory or staffing or skills, and being able to translate that into the loads of that you have in order to produce your goods. So for example, we worked with a

company in Chicago. It's about a billion dollar company, and they were able to optimize their inventory holdings by about $40 million year over year, simply by knowing more about what the demand was from their customers, by building AI models around it. But what I really think is interesting about this is that once you know more about yourselves, you start to know more about your customers and your customers and your

customer's customer. So what companies like that can then do is they can translate that understanding into a conversation with their customer about their strategic needs. So in this case, it's about I know more about your demand needs than you know about your own demand needs, and I can form a partnership with you to help serve you better. Some organizations are taking it a step further. They start to understand the picture of their

customers. So working with a global restaurant, provider of food services, and one of the things they start to figure out about their customers is, what makes a successful restaurant, why does a restaurant stay in business, and why doesn't a restaurant stay in business? Well, who knows more than the food providers? Who knows more than who's delivering that every single day as to why those restaurants are staying in

business? Well, they start to get to a position where they're not only delivering that low margin service of the individual product they're delivering, but they become a knowledge broker to their customer, and they can provide now layered on premium best practice services to say, This is what the top quartile is doing in your market. This is what makes them special against your competitors, and that translates into almost any

industry. So understanding more about yourself allows you to translate that into understanding your customers and what your customers customer wants from that organization. That you're serving helping them to be more successful. Ultimately, that is what our business is trying to do, is make our customers more successful at serving their

customers. So AI really enables the mission of every organization to be able to be more successful, if we look at it through the lens of what makes every organization unique and special, and allow that to translate into leveraging technology to make that true. And that's what's really made this an exciting time.

like that 100% Yeah. We will be back to this week's episode in just a moment, but before we continue, I have to tell you where I will be in early December, a CTE career tech vision happens this December, 4 through the seventh at the Henry B Gonzalez Convention Center located in the heart of San Antonio, Texas, along the banks

of the beautiful river. Walk. Vision is the largest annual conference in the nation for career and technical education professionals, and offers you opportunities to gather with your peers and grow professionally, expand your professional development and ignite your inspiration. There is still time to register visit career tech vision.com, to learn more about this premier event. And now back to this week's

episode. All right, so Sarah, I know win trust is kind of at the early stages of your your AI journey. And one of the reasons that I love having you up here is that that's where most of us are right. Most of us are are just starting to figure this out, just getting going. And so I think you have a really, really unique perspective for our audience. So as you're starting that AI implementation journey at wind, trust, are you just like jumping right into projects?

you're using, right? So you know you you might not have anything going on internally, like in a lab or in a, you know, fun little innovation group, or anything else. But think about all of the folks that you are

leveraging, right? So if you're, if you've got a CRM, you know, a sales force, so, you know, one of those, and things like that, there's just AI everywhere in that, and it's leveraging your data, and it's potentially giving your sales people recommendations of how to approach their customers or their prospects, or what have you. It's all built into all of the Microsoft apps. Now, if you don't automatically shut it off, it's odd. And so it's there,

right? And so what we needed to do was not only take a look at, you know, where is that from a vendor perspective, and our partners and everything, and how they're leveraging it, but also then, are we sort of inadvertently using it and not even realizing it? Does that affect our customers? Is it

affecting our employees? And so I think that, you know, one of the really, you know, good points to that a couple people have sort of referenced, but I'll really be explicit about it, is, where are those opportunities for your employees to start leveraging it and, you know, in a way that's not going

to scare them. And so that's another way that we're looking at it as well, like, but we're not going to just jump in and suddenly have like, you know, some sort of AI algorithm, Robo advising our entire client base on their financial investments like that is not where we start,

right where we would. We need to start first of all from regulated perspective, as policies, procedures, standing up potentially a center of excellence, similar to the review board that Brian was talking about, where you have people looking at all the different aspects and things like that, of how that might impact you, and particularly, how are those, and I forget which app it was that you said that actually gives you the references perplexity. Yeah,

that's huge. So to me, anything that we're going to do in the AI space, once we really actually start turning things on and leveraging that, is you have to be able to go back and say, Where did this information come from? The transparency of it is going to be crucial, I think, in every industry and every application, but in particularly in regulated industries, because we're going to have to answer for, okay, how did you make that

recommendation? Right? So at some point, if we're going to say here, you know, for me, I, you know, I chose to put a new branch in this town, and I chose that corner, right? Well, how, if I use any AI, and there's all sorts of applications we're evaluating, I'm looking at different vendors right now that have all this different data, right? It's cell phone pings, and it's, you know, everything, every little piece that's going to drive these algorithms and then be prescriptive, be

predictive. About like this is where we think you should be next. If I'm going to actually leverage something like that, and as the human then look at the look at the recommendations, and choose one that's recommended by something that's completely, you know, artificial, based on data from outside and inside, I'm going to then have to, at some point say, Okay, here's why, and here's all of the information that created that recommendation. So the transparency is going. To be

crucial. And so that's the those are the other things that we're going to be looking for as we evaluate, like, where and how are we going to use this? And then we're going to be able to, I, my personal opinion is that you're going to have to start internal. So how are you going to leverage things to make the employees lives easier, right? Because a happy employee is going to create happy customers. And then how are you going to do it, maybe in a way that, like,

yes, it's customer facing. So think of like a call center, right, or something like that. How many people have ever called a call center and got an answer you didn't like and then hung up and called back and got a different answer, right? And maybe that answer was even worse, thank you for the hand and so right? Maybe that was worse, but maybe it was better, and usually worse and better, or just depending on what your problem is, and did they solve

your problem, right? And if you feel treated fairly, and if you feel like you gotta, you got a good response? Well, we can go from, you know, maybe just using it to help employees onboard easier. And, you know, auto enroll in a 401, K or something, something personal, something internal, to then helping those employees when they are servicing a customer, maybe it's already working in the

background. To, you know, scan all the policies and procedures and everything that these people are talking about, and say, here is the right answer. And now you have consistency, and you have actual correct answers going out to people. And then maybe you get into, you know, the robo stuff and everything that's like, Okay, this is that next best solution for you? Or this is that next best answer? Actually, I think that you don't, you aren't going to want to, you know, a 30 year

mortgage. You might want a five year ARM right now or something, because something because of x, y and z. And here's why we think that. And I think that the empowerment that can come out of that to help people win with money, which is my personal ethos, is, you know, democratization of data and winning with money for everybody. I think that that's where we're going to see the most lift from a customer perspective. So those are just

Sarah answers right now. It starts with the data and the data and the governance and all the rest of it, but I think it'll all come together and coalesce at the right pace for our company so

instance, were air gapped. All the data stayed inside, and now everything's connected. Everything, you know, everything is censored. We're pulling all this data. Do you worry a little bit about IP and how we keep that safe? And how is quad thinking about

you know. So as we came into January of 2023, you know, kind of post holiday, and us all kind of see the impact of chat GPT and what was to come, we knew that we needed to create a steering committee. Sounds like very similar to my peers here, you know, we were very thoughtful about making sure that we had representation from every area of the business. That was key. So that was the

starting point. We also have a top level steering committee that we have, AI business unit champions, that's representative of every area. And then we kind of broke down and said, like, what type of data are we dealing with? We have employee data, we have customer data, and we have quad confidential data, right? So we have to kind of go from there, and then we have to kind of look at that intersection of

roles. So which roles are intersecting with what type of data, and how do you put logical securities around those things? So we deal with financial services, we deal with healthcare companies. And the answer there is number one, it needs to be completely segregated, and it needs to be completely containerized, and it cannot get into an AI model in

any way, shape or form. Then we also deal with brands who are trying to sell T shirts and hats, and they're like, how do we get our data into a generative model so that I can create 50 different iterations to reach my market? And I think that over time, the more regulated industries will be

leaning into that. And I'm not being critical, it's absolutely what needs to be done, but I think it's a huge challenge for quad, because we have to really be thinking about what's the role of the individual, what's the type of customer that we're serving, what is the type of product that we're serving for that customer? So it's, by far, not a one size fits all. I'm really proud of what we built.

It's not perfect, but I think the it started with creating, you know, kind of a, you know, a North Star that goes all the way up to our senior leadership, and then it can't and then, really, I think where the rubber meets the road is having business leaders that have the knowledge of how AI may affect their space and making sure that we're all I think you said it very well, like there's so many commonalities even within our

company. And then you look out in the world, you're like, Okay, this is how we need to be thinking about data protection and security, and it'll morph the leading edge marketers that have a little bit higher risk tolerance as that starts to harden for the world. And for the legal community, then some of the more regulated spaces will, you know, I think, start to lean in even further than they have lots

try to get ransomware. So we spent a tremendous amount of time making sure that we migrated our data to the cloud. We're using best breed, Microsoft, Azure, to ensure we've got all the security protections in place. The other piece with data privacy and healthcare is if you're applying artificial intelligence algorithms ensuring you don't have bias in your underlying data sets. Rogers, we see a specific population that comes

into our doors. We have chosen deliberately not to make the move of monetizing algorithms or doing anything along those lines. A, because the algorithms have learned off of our patient population, and would it be able to have the same amount of accuracy in different places we don't know, and we don't want to take that risk in B, in health care, every piece of data that you collect is someone's story. It's someone's personal information. It's someone being vulnerable, it's someone coming

into care. We treat that with the utmost respect that we want to use it to make sure that it's helping them on their journey of recovery, rather than doing things of mining information or maximizing revenue or things along those lines. So we just take it on a small little bend from that personal level level, as well as just the privacy, security side, absolutely and as

customer data. So, so what policies are you developing at wind trust and what fun foundational elements are you considering as you move forward?

are AI experts. They do understand, in our case, financial services, privacy, security, all of the all of the aspects that we need to to look at, so that as business cases come up, it can not only help us implement but also kind of help us potentially, hopefully, with, you know, ROI and stack ranking, what the opportunities are. Because here's the thing, the opportunities are going to be endless. Anything to do to me, with technology is only limited by our own human imagination,

right? And so, like, it's kind of like a shoot for the stars land on the moon perspective that I always have. It's like, what else could we do with this, right, if we weren't scared, if we weren't limited, and that sort of thing. And then where do we start? Right? Because I'm a big kind of, I always say, like, I'm gonna throw a lot of juxtapositions at you and cognitive dissonance, but I'm both a, you know, begin with the end in mind and road map it out

and dream big. And also, don't let the perfect be the enemy of the good. Don't let, don't let something be the enemy of nothing, right? And like, just get going do something, start somewhere. And so as we look at standing up both this group to help implement all of the governance and all of the policies around, like I said, transparency and accuracy, and, you know, just being fully everything, fully documented.

And I mean, I mean fully documented, you can't really understand model risk management until you've been through it. But, um, once you have all of that in place, then, to me, the the best opportunities are going to come from the front line. And so somebody that is frustrated with the way that they have something to do every single day, right? So whether it is those, you know, we don't want to chat bot, we still want

humans doing it. But how do we make sure the humans are fully, you know, fully up to date on every single thing that they might get asked every single day, you know? How do we layer it in there, right? How do we layer it in around, you know, lending and and maybe recovery if something's going wrong. I Brian, I'd love to talk about what you guys have done in terms of improving that accuracy and the outcomes and, like, how you

plan that? Because I all I could think was, I wonder if there's a way we could take our data, look at every single loan that's ever gone bad, and figure out, like, which aspects really would have mattered, so that we could predict when a loan seems like it's going bad. Do I have actually have a chance of recovering this if I restructure it? Or should I just take my first loss? Because my first loss is always my best loss.

Right? Like, what rules that we've always operated by are just gonna get completely upended, because now we can actually feed data into something and get a real answer. The possibilities are endless. And so that's why I'm one of those. Like, I've always been a I always joke that I'm kind of the weirdo. I'm like, the change management junkie. I'm like, what else can we change? When can we upend this? How do we transition? How do we get people

excited about change? And I'm I say I'm the weirdo, because most people are terrified of change. You know, you might hate the thing you're doing, but you're the best at doing it. People want change, but they're scared

of the transition. And so I think that by using the right policies, putting the right people in place, having the right leaders, both on the technology side, but more not to say, more importantly Nate, but just as importantly on the business side, and having those business champions like George talking about to surface those ideas so that we can kind of figure out, okay, where does this fit? And then roadmap are the way to the future.

important to you. Tell us a little bit from the IT angle, where you spend a lot of time, what are two or three things organizations should be doing to prepare their data for the age of AI?

right? You're not going to make progress until you actually try to make progress. So understanding what you're truly trying to achieve starts with the business, starts with the mission of your organization, starts with what you're trying to do to help the business win. And many organizations don't start there. They start with it as a problem or it as a data holder that's not really sure

how to move itself forward. So the first thing is to understand that prioritize understanding how your strategy translates into leveraging data to make that real in the context of artificial intelligence or other types of platforms. That's true for both these very narrowly focused AI use cases that many of us have been talking about, as well as very broad commodity use cases, such as, like, what a co pilot or something along

those lines would apply. You know, we sometimes forget that, like, one of the biggest impacts that's going to happen in our organizations is every employee leveraging AI to be more every employee being able to take their individual work tasks, the data that they use on a daily basis, whether it's in a database or Excel or Word or PowerPoint or some document from 10 years ago, and leverage that information to be able to perform their activities in a more in a way that allows them

to be more effective at their daily work. So what I'm seeing organizations do absolutely organizing Centers of Excellence or strategy, organization, however they define that cross functional team, but it's really trying to understand what is the ways we're going to win as an organization over the next two to three years. How does that relate to our current state? It may be completely different. They may be disrupting their organization to be able to approach the market in a

completely different way. So they think about that strategy of the business, then understand what data is going to enable that to be true. Don't think about data in terms of just

straight data driven. Think about in terms of an objective that you're trying to achieve, what needs to be true from a data perspective, from a staffing from a skilling, from a project and portfolio standpoint, for me to achieve those goals, that then allows me to be able to build a data state that serves that objective as well as the skilling around it.

Final point I'll give around the data state, and this is probably an undervalued aspect of AI, is how much AI can enable organizations to find non intuitive insights from their data. I was working with an organization that was leveraging manufacturing data, and there's a part of their manufacturing process that they didn't understand, like they're half the time, it would produce a level of efficiency of the production process that was wildly different than the other

half of the process. And they couldn't figure out why that was, but they were able to leverage that, that sea of data that I think you mentioned, you know, so much of this information on our manufacturing the OT environment is now available to us, or the healthcare environment, or other environments that are being able to start being available for us, to be able to get sensing and information and experimentation.

Well, they were able to leverage AI to be able to figure out, what is that missing piece like? What's the missing element that keeps us from being more efficient on a consistent basis across our production process? That's simply by finding non intuitive insights from experimentation with AI, not just by auditing a process. Sometimes we kind of quickly go to like automating a process. Is what AI is. Yes, that is, that

is a component of AI. The other part of it is using AI to discover something new in partnership with a person's unique skills. Capabilities. So data has much to unlock here, and we're just right at the tip of the iceberg. At

And I think just the world needs more examples. It's one of my greatest pet peeves, and I will confidently say it didn't occur today, where you go to an event and people just say the word over and over and over and over, and they don't really say what the heck they're doing with it. And I just, I'm very critical of

that. I'm very dubious of that, and I'm proud that, you know, quad is taking the time to say these are some specific things, whether you're on the manufacturing floor or you're an executive, or you're a marketing media planner, like, here's some real stuff that can make your job easier and more effective. So I think just getting the word out and getting real, tangible examples out is key. Thank

medical record. You think when you apply some of these different AI techniques, it's amazing the amount of burden that it could reduce while increasing accuracy and meeting all the documentation expectations in a regulated industry. What I tell people, first off, AI is here to augment your job. It's helping there to reach your full potential or work at the top of your license. B, everybody associates AI, I think, in the space today, with chat GPT, AI is actually very

old. I mean, you could actually accredit it back to Alan Turing in 1940s what has actually happened the last five years is this increases computing power, cloud computing, and it says kind of reached a little bit of a plethora that it hasn't been before. But it's not as scary as people may seem, and many of the methods are already there. What scares me the most about AI is people who apply it without understanding what is truly going on in their underlying

data. I think everybody takes a shiny object, they apply it, they want to get some sort of nugget or insight out of it, but you truly have to understand your data, and you have to understand your domain expertise to do it right. And I think especially in the healthcare industry, also having outputs that are actually actionable. I think for healthcare, we should not be doing algorithms just to

do algorithms. They should be there that has actually some sort of output that a clinician can actually use to improve, help out.