AI: From Basketball to Business - Part 1

Welcome to the ECI Pulse, where we break down the latest trends shaping the business world. In this special two-part series, Jeff Schmidt, CEO of ECI, and Rich Eitri, Chief Innovation Officer, dive into the fascinating world of AI and its impact on business efficiency, workforce adaptation, and the balance between automation and human expertise.

In this episode, Jeff and Rich explore the latest trends with AI and explore how AI is reshaping the industry and our world. They share real-life stories and use cases, including how AI is being used in unexpected ways in everyday life.

Join us as we uncover the tools empowering businesses to innovate without extensive IT resources and discuss the future of AI in everyday business applications. Whether you're an executive seeking a strategic edge or simply curious about where this technology is headed, this episode offers valuable insights and engaging anecdotes that will keep you hooked. Tune in to discover how AI is giving us back time, driving agility, and transforming the way we work.

Transcript

Hello, and welcome to the ECI Pulse, a podcast where we break the latest trends shaping the business world. I'm Jeff Schmidt. I'm the CEO of ECI.

We are the leading provider of cloud services, cybersecurity, and digital transformation solutions aimed at the alternative investment firms worldwide.

Today, I have with me our Chief Innovation Officer, Rich Eitri.

We're kicking off a special two part series on AI, its role in business efficiency, workforce adaption, and the critical balance between automation and the human expertise.

AI isn't about replacing jobs. It's about scaling businesses, streamlining processes, and unlocking deeper insights.

In this series, we'll dive into how AI is reshaping industries from mergers and acquisitions to investment strategies and the growing importance of governance, security, and infrastructure, and AI deployment. Plus, we'll look into the tools, like robotic process automation, low code platforms.

These are the things that are empowering businesses to innovate without extensive IT resources.

The future of AI is here, reshaping the way businesses operate and create new opportunities for growth. Whether you're an executive seeking the strategic edge or simply curious about where this technology is headed, I'd invite you to join us and uncover the answers as we go forward.

So, Rich, how are you doing today?

I'm doing good, Jeff. I'm doing good. I'm excited to talk about AI. I feel like you and I talk about AI quite a bit these days.

So crazy conversation. Right? I'm at breakfast with my son-in-law, and he starts talking about his kid's basketball team, my grandson, and how they're using AI to develop the roster and the time that kids should be spending on the basketball court. So special things that are there. And it just blew my mind that it seems like AI is starting to hit mainstream just even in the simple things that people are using today.

You know, as we open this up, I mean, as our audience and listeners start to get into this, like, what are some, like, crazy things that you're hearing about AI where you just wouldn't necessarily believe somebody might be using it or where it just kinda takes you back for a second?

You know, I think, like, personally, I I've used it, like, recently. I had to get my hot water heater replaced.

And, you know, typical, I delegated it to my wife who got, you know, probably a couple of quotes back and like four different solutions. And I'm like, none of this makes any sense. Right? So I went into the reasoning model and began to just type in.

Hey, here's the size of my house. Here's how many bathrooms faucets things that I have. Here's some recommendations I've gotten. You know, here's some of the challenges I have today, right?

With hot water and things like that. And it gave me a really good thesis around what solution I should go with. And I was able to iterate on it and said, hey, like I live here. It's here than what the seasons are like and different things.

And it came back with really a kind of full analysis of why I should go to certain directions and even estimated, the cost. And I was able to go back to my wife and say, look, I do all this research. You know, I have the answer to, like, the direction we need to go. And I did it in about ten minutes.

Right?

So, like, to me, it's beyond, like, a research tool. It really is, like, a kind of smart assistant in a number of different areas. Right? I even used it, you know, things like the NCAA tournament, things like that.

You can, you know, go in very quickly, begin to ask it, like, trends and previous winners, and help you make decisions on, you know, brackets just based on historical data. Right? And, you know, I think people don't realize the power in everyday usage. Right?

Which then I think allows you to start thinking in terms of how to use it in, in the business world.

It's an interesting way of putting it right because I think of maybe when my kids were younger than them. And my question might be as, well, tell me why we should go do that. Right? Now your kids can go off and come up with a business case and hand it to you in a matter of minutes, which is interesting.

So it's interesting. Right? And as you see this, right, we've been maybe in this in this world of real GenAI, maybe for eighteen to twenty four months. Right?

Whereas as it really started to gain momentum.

What's the tipping point where we see adaption happening faster, whether that be in business or how business users are using the tool sets today? Like, what changes, like, if you think about the advancement of an iPhone.

Right? We started with Blackberry's, Trios, Palm Pilots, some very rudimentary smartphones, maybe not so smart. Right? And then you got to the iPhone, and all of a sudden, this acceleration just happened almost overnight where everything became, you know, I wanna use this in business, how we work, how we how we play, etcetera.

It became this very personal device. So what's the tipping point or what changes in AI where maybe some of the things that are happening today mimic the iPhone and how we see that starting to happen because AI isn't new. Right? I mean, it's not a new thing, but where we're sitting with the data and the information, it is.

I there there's definitely a couple of trends that I see where it steers adoption in one of two directions.

I think there is now AI being embedded into so many of the everyday business applications that people are using.

And every now and then there's that moment. Like they might ask a question of the tool or the tool surfaces.

A trend or a piece of data. That they hadn't seen before. And people like, like I see the value in this and I see the value if it understood my data better, if it had more context. And then that's when they're willing to engage in a more deeper discussion around how to use, you know, generative AI across the enterprise.

Right? I think the other piece is, you know, there's a lot of people who are reluctant. Right? Whether it be because of security, compliance, you know, data privacy concerns, or just believers that, like, look, the technology is still too new.

You know, look, I generally feel that we're in the second inning, right, of the baseball game. I don't I don't I think we've just begun this. And, you know, the excitement that I see around the tools, like, people, like, won't really fully understand the impact of this until, like, four years from now when they see the level of automation, you know, creeping into their everyday lives because of generative AI. But I think there's one of two paths that that that people either take. I mean, look, I mentioned this. I when we did our Microsoft events, you know, like, over a year and a half ago now where I have been in this business for a long time running IT and so have you, Jeff. I have never seen business leaders so engaged in technology and when it comes to AI.

You know, like, we have so many, you know, across our clients, so many of the business people interested in in discussing it, whether it be use cases or ideas that they have around automation.

And I think to me, that's where a lot of the value begins to come. When people start thinking about the art of the possible and looking at their businesses through a fresh lens. But it takes some moment for them to realize that that value.

It it's an interesting lead in to the next question. Right? Is in the examples you gave and as we see this starting to come to light, it seems like speed, right, the speed and rapidness of how fast you can work, and then that would apply itself then as time.

Right? That that somehow AI is going to give us back time. We'll find ways to fill it, I'm sure. Right? We've been pretty good about that with computing and compute power that we've had today. But when you look at that and you look at the conversations you're having, what are some things for the listeners to maybe grab onto as ideas in in our space, in the in the client base that we have today that are maybe interesting of how it's being used today, early adoption, even if they're simple examples, but where they're saving time or, driving the agility and the speed of the business that they have.

Look. I tell people all the time that if the tools that I have were taken away from me, I'd be really upset because I found, like, a number of ways to create efficiency, and I see it, you know, across our clients. Right? I think the first is, you know, there there's mixed reviews around, you know, the value of Copilot.

And I think Microsoft Copilot in in particular is extremely valuable if you're using it for the right things. Right? Especially that Microsoft three sixty five ecosystem that that we live in. You know, Jeff, you often say this.

Right? Like you like, how do you get back time?

And I think the pace of business has grown so much over the last four years. We always say, like, I don't have enough time to do this. I didn't get enough time to do this. And these tools give us some of that time back.

Right? Double booked meetings. Just to be able to record a meeting, get the simple AI notes from that meeting, and then being able to answer questions, you know, that may have been directed at you in that meeting. Like, I used to have the team record calls, and then I would listen to them on the weekend while I walked my dog.

Right? But I would be home, and I'd still be listening to meetings because there was so many I missed. Now at the end of the day, I just look at the ones I missed, and I look at the key content from there. And, yeah, maybe it takes me a half hour to reply to all of them, but that might be four meetings that I never would have been able to attend that I now understand what's going on and could provide some help guidance, around those.

Right? And I think, again, like, that's a couple hours in the day that that I get back. Right? Email, chats, you know, we're all traveling.

We're on the move.

Being able you come back into an email thread, and it's like there's sixty emails, you know, on one particular thread and realize, like, in there, there's probably, like, four bullet points that are really important. And being able to pull those four bullet points out and then mark the other sixty items as read feels so good, I could say. And I think people don't, like, realize those small use cases that are extremely valuable and being able to access that tool from your phone. Right? So not even and, again, in, like, that secure kind of compliant way. Right? And I'm sure you see some value in those, Jeff.

I think it's really interesting, Rich, on the on that statement, right, is the simple release probably a month ago, follow a meeting.

Right? And it just recaps and you get everything on it. You don't have to sit on a meeting anymore. The transcript comes in. You can you can then generate, like, are there any action items for me on the back end of this?

That's hours back in your day of sitting on a sitting on a meeting that before you're like, I wanna listen to it. I have to listen to an hour of meeting that now you can recap in five minutes.

For the listeners out there, it's available in the new Outlook capabilities within the calendar section.

So you have the accept the meeting, decline a meeting, and you can now follow a meeting, to the right of the to the right of the, the what is it? The maybe side, which I probably did with all those meetings in the first place. I made them all I made them all maybes, and I didn't show up are tentatives.

The other thing I think is really funny, right, is I remember at one point, you just said something that was really interesting. Right? It's we were attached to devices.

Right? I love my Mac, but don't take my Mac away from me. Right? And in fact, I have an incident where I wanted to show somebody how passionate I was about client success, customer experience.

And I walked in, and I took two guys' Macs away from it. I saw I'm gonna replace them with PCs. I hope nobody from Microsoft's listening to this at this point in time.

And, like, literally, these guys were like, like, if you take that away from me, I'll quit. Right? And I think we're moving away from device reliance to application reliance. Right? Is this is gonna change how we think about how we work is Copilot and like, take Copilot away from me in my daily job and work. That's much more important than the device I'm actually working on.

You know, example for me is I left the meeting the other night, and I'm going on from the dinner.

And I just wanted to really just do a quick recap of what was happening. And so while I'm walking down the street, I just do a quick, hey. I met with so and so. Here's the five things we talked about. Here's the three follow ups I have.

Create one email for me for the client as a follow-up with a nice thank you, and I enjoyed the dinner and the time and the company, and here's the three things we're gonna follow-up on. And then can you also write another one to these people in my company that I'm gonna need support to be able to go do this and a follow-up, and I need dates and times with them. So that would have been back in my hotel, you know, pull out my laptop or grab my phone and start typing it out. By the time I get back, I already have that curated for me.

It's ready to go. And I just do a quick review on it, and then I can send that out. So there is a certain amount of getting time back, right, as you find different ways to fill it. But, hopefully, you're filling it with more meaningful ways where you're using the critical thinking of your brain to apply it to the areas that really require human interaction in it.

And I think those are probably good examples of, like, mundane tasks. Like, nobody's getting a benefit out of writing, like, the follow-up.

But what you have to do in the follow-up is really important.

Well, the natural language piece has gotten, like, tremendously better over the last, like, year and a half. So I do, like, similar thing where our I'll walk around the house or the office, and people think I'm on a phone call, but I'm really just dictating notes into because then it I can codify it with other content I have and, you know, produce, content. So even, like, presentations now, I will dictate them into OneNote or Word and then save it down and have PowerPoint put a first draft around it into the corporate template. And then I look. If even if it's fifty percent there, which most of the time, it's a little more than fifty, it saves me so much time. Right? And I, you know, I do work on some client strategies.

And within, you know, our LS solution, I use the reasoning models. And I take all of my historical strategies that I've built for similar type clients. I load them in there. I take all of my notes. I load them in there, and I give it some ideas around how I wanna steer the strategy. And it'll produce a very good set of content for me to now, like, give back to the clients around all the things that, like, I brought into the, you know, the context.

And, again, it's might be sixty to sixty five percent there, but that process would take me normally, like, eight to twelve hours. I can get them done now in, like, two to three hours. Right? And that's a huge difference, right, in terms of time. Right?

And those are basic scenarios where you don't need, like, special tools. Right? Like Copilot could do something very similar.

I use Ella because, you know, we have the power of the reasoning models embedded in there. But, you know, you could still leverage regular Copilot to do something similar.

So you did you did a good job of talking about how you're inputting data and using it.

Going back for a second, when you talk about a reasoning model versus just normal Copilot, and how you ask questions to it with prompts.

And then the difference between, like, a large language application and a large language model.

Can you maybe help our listeners a little bit more with what that like, how you use a reasoning model compared to prompts, and what those are, and then maybe the difference between the LLA and the LLM?

Sure. Look. Reasoning models are extremely powerful, but not needed in in every use case. Right?

So the standard large language models like you see from OpenAI, like chat g p t, the three five to, like, four o people might hear about, or some of the llama ones, you know, they they've been around for a while. They're really good at, you know, reading large unstructured content, summarizing it, you know, finding little pieces of data and giving it back to you, doing searches across content. That's what they're really built for. Right?

They had, you know, I forgot how many billions of parameters that, four has, Chad GPT four, but the similar models with, you know, Google and Anthropic.

But what the reasoning models do is they're actually doing more kind of analysis of, like you use it for more analysis. Right? So you could actually ask questions of it because what it does is it goes back and it tries to build the strongest strength of the answer. Right?

So with a lot of the kind of GPT models, which is why sometimes people get frustrated when they use them in the wrong use case, it's kind of, like, you're taking a shot, you ask it a question, and it'll find, like, the best match, but it may not always be right. And it may be a little deterministic. It may fill in the blanks, right, if it has it. Reasoning models are designed not to do that.

The reasoning models will go through to give you the lineage of how it actually came to the answer for the question that you asked. And it takes time because it goes back. It verifies. It rechecks itself.

It finds gaps. It'll try and re answer the question. So they're inherently slower as well because it's doing more compute, more work. But generally speaking, the answers that come out of it when you're doing analysis and working out things like strategy and, you know, other types of complex things where maybe you wanna put a revenue model in there.

And you wanna be, like, play with things like CAGR or, you know, IRR, things like that. Those are what the reasoning models are designed for. They're really good at caching that information that you give it, but they expect you to iterate on it. So I think that's the other kind of piece that people need to understand about reasoning models.

It's storing everything that you're asking it in one context.

And it's gonna learn from, you know, questions that that you ask it, but also when you correct it as well. Because, again, like, they're very much about, like, trying to figure out an answer, but it's not always right. And if you say, hey, what about the revenue for twenty five? You have the information from twenty four.

Why didn't you calculate that? You'll be like, oh, you're right. It goes back and it'll calculate it. Right?

Because it realizes, like, oh, wait. I did do that, but some reason I skipped over twenty five. So they're much more, you know, thorough. Right?

But they're also slower. Right? But I think you're you get better quality answers around.

Interesting. So that's helpful. So if I'm maybe, like, a use case in this scenario for the listeners would be, I'm looking at a SIM and evaluating a business.

And I've got maybe the perfect company as one of the examples, and that's my thesis is if I can find more companies that look like this confidential, memorandum that I have, that this will show me if I can match this, it's probably a good fit for the thesis that we have. And then it isn't just running it once like I would with Copilot. It's give me the give me the give me the output of the meeting. Right?

And it it'll give you, like, here's your updates and what's here is I can now become conversational with it Hey. Go check the gross margin, the growth prop the gross profit. Look at it quarter over quarter for our twenty four. And I can keep doing this, and it's gonna contain it.

And then as I as I get to the output that I'm looking for, that then is gonna produce a much more robust outcome similar to what you said when you were doing relating it back to the beginning. You said reasoning model when you were doing the water heater. Right? It has give me an output that I can actually use, which is now you have a definitive answer for something or at least close to what you think it should be.

You're applying it to your house, all the different variables that you gave it, but you were able to build that over time. So it makes sense that it would take longer because you're creating a longer set of code, if you will, I guess, in that scenario. And then you mentioned Ella. Right?

Which Ella if I understand Ella correctly, as you explain it to people, Ella has that ability to be able to save a lot of those routines or those formulas that you're building and be able to use them repetitively over and over again where you don't have to keep your reasoning actually gets better over time, and you can continue to use those reasoning models, from a business perspective and repeat and use them versus having to start from scratch.

Yes. So, Ella, we we've built a RAG engine. Right? Retrieval augmented grounding where it pulls in and creates a data domain.

So as I mentioned, I uploaded previous strategies that that were similar, you know, other content around industry trends. You begin to load all of that data in, and now you've created a very, very specific database, right, of information around, you know, content. And now you can actually begin to ask questions, right, of that all that data that you've kind of loaded in there. And that's really important.

Right? I think that's where the power is. Like, even on the SIM example, you know, you could upload previous SIMs of similar companies and say, you know, first you ask the LM to analyze each one of those SIMS and they say, hey, can you now bring this together? Show me in a table, how gross profit lined up across all of these companies.

Show me how gross margin was different. Show me how many salespeople each have. Right? And you can begin to do that, very easily, right, when you have that that data.

And, again, it doesn't require what you know, we used to do these projects all the time when I was on the other side of the fence.

We'd have portfolio managers and, you know, analysts come and say, hey. I have all this data. Can you guys, like, read through it, produce some, like, re reports for me? And we would have someone work on it.

It would take a day or two, maybe a week, depending on how complex it was. Now you could do that yourself in minutes. Right? And that to me allows you to keep up going back to, like, the speed of business.

Right? And when you think about our typical ICP, you know, even if it's, you know, our PE clients who have to get bids out, right, value these companies, really understand, like, where the data is and get bids out quickly. Well, you can peel a couple of hours off a process each day. That that's meaningful, right, in in in the process.

Interesting. So and Ella is ECI's large language application. Correct?

It is. It runs in a in your own environment.

So you we you know, I'm a strong believer in, you know, the technology should be brought to the tool to the data, not you bring the data to the tool. So I am not a fan. Right? And maybe it's because I've had to spend years securing data to let it leave my four walls or even SaaS platforms that I've spent time securing, right, and doing due diligence on to now pump it to another platform that, you know, has my most confidential data.

I think in today's day and age, you could bring, you know, the technology to the data, and that's what we do with Ella. And again, like, you bring your information, you know, to it. It runs in your environment. And that's really, to me, the first starting point, especially with AI because when you talk with a lot of clients, especially in highly regulated industries, they get very weary.

Right? Like, you know, because the bang for the buck is using the most confidential data. Right? So now you're gonna take it.

You spend all this time securing it, and you're gonna toss it up to a platform that maybe you've done some due diligence on. You're not really sure how they're learning the models. You're not really technical enough to probably dive into it. You know, it's risk.

And, you know, that so that was really the premise behind why we built Ella was, you know, not to, like, reinvent the wheel, but really give the clients the ability to bring the technology to their data and then be able to create custom use cases around it. Right? So we could do you know, we've created stuff for SIMS, financial statement analysis, use cases that, you know, require special models or code around it that we know are gonna be valuable to clients. Like, we built, you know, some of those agents in into Ella.

Notably, you touched on a critical point in your answer there, and that's security.

So as we wrap up this first part of our discussion on AI, it's evident that security controls, governance, and regulatory measures are still evolving, and the regulators are actively working to figure out the best approaches to manage these challenges with AI.

Now we're gonna continue this discussion in our next episode, so stay tuned for part two of

we're gonna dive deeper into the intricacies of security and AI. Again, thank you for listening to ECI Pulse. Have a great week.