HITT- AI and Cloud Understanding and Implementation Challenges – Sep 3, 2024

In this HITT, Koby, Donovan, and AJ explored the growing interest in AI and its applications, emphasizing the need for organizations to prepare for effective implementation. They highlighted the importance of clear use cases and the role of data quality in AI outcomes, noting that bad inputs lead to poor results. Donovan introduced himself and discussed the evolution of AI technology, stressing the necessity of security and proper data management. The conversation also touched on the pressures organizations face to adopt AI and the need for clients to understand the complexities involved. Overall, the session aimed to educate clients on navigating AI technologies and aligning their goals with effective strategies.

Guys, I think we’re ready for the hit series. So let’s bring on Domin and AJ. And we got we got a a big one to to follow. Last week, we had the mayor of the northeast and and really maybe the mayor of the channel and Mike Balodran in.

And can we just also guys, just personal note, lesson learned. When you agree to do something, ask more questions. A, I didn’t realize we were falling at them again, which is great. When I when I said I would pinch it for Doug b, I didn’t realize I was gonna have to say Mikey b’s full last name.

So I hope that I did it justice and that, as much as Mikey b brings value to all of us in the channel, he did a great job last week. So we we have to we have to step up here, guys, AJ and Donovan. And I think it’s a topic that everybody is really excited to to learn more about when we talk about AI. Right?

And I know that we’re beating that drum quite a bit, and there’s a reason for it. Let’s level set. Customers are really interested in it. They’re getting pressure internally, externally, trying to figure out what they can do with it.

And I think that’s really what we’re gonna zero in on today around the other AI conversation. Right? The CX AI conversation’s been really involved for the last three, four years. And Sam and Mikey b and J Lo, I mean, you talk about deep level expertise on where you can take that conversation.

I feel like it’s it’s time for the cloud side to step up and give a little bit more insight and a little bit more explanation on what’s going on on our side of the coin with AI and LLMs and things like that.

But first and foremost, there’s always the number one thing when anytime I talk to customer and Donovan, AJ, I’ll bring you guys in in just a second to expand on this.

It’s, hey, we’re looking for a use case. We wanna we wanna use the AI, but we need to find a use case, and we need to be able to really talk through that organizationally.

So what I’d love to do is just give a little background on what’s going on with the AI pieces of it and then, the background of it, and then we will dive into some of those use cases that our TAs can take and start those conversations off. So, AJ, if you don’t mind just giving a high level what is an LLM, and then, Donovan, jump in and explain, you know, anything that, that you want on top of that.

Sure. So, hi. Just quick intro. My name is AJ Kuftek. I’m field CTO with Expedient.

An LLM is effectively a way for computers to turn words or images or audio into numbers and then do math to generate the probabilities of what should come next. That is a very, very, very simple form. It’s very technical. But effectively, what it’s doing is saying, okay. This sentence starts with see Jane.

Okay. What should come after that? Should probably be a verb. Right? So what are all the verbs that are possible to come after that?

And then based on the context of what was asked, it generates what that verb should be. So that’s really what’s happening when you type in the chat g p t and you ask it a question, it’s responding in a very, very human way, but it’s doing that based on training against publicly available data, sometimes private data. And it’s doing that in a way to try and provide a response to you. The other thing that it does and a lot what a lot of people forget about AI and when they’re typing into things like ChatGPT, ChatGPT is a front end to a number of different models.

So an LLM is there’s multiple different models behind it. They’re usually trained to either be very general purpose like GPT. That’s what the GP in GPT stands for. But you also have things like DALL E, which is an image generator, and it’s been trained on images and how to generate images.

There’s ones that can generate video. There’s ones that can generate audio.

But they’re all designed to output something that is consistent with what was put in. Right? In a lot of ways, it thinks there’s no thinking to it. There’s in a lot of ways, it acts like a person. And when you think about that, what actually makes LLMs very interesting is that the new programming language is the human language and communication.

And this is where I think a lot of people fall down when they start to use something like an LLM, whether that’s Copilot, Chat, GBT, or others, where they just ask it a very direct thing. And it goes, I I, this maybe? Because it doesn’t have any sort of answer. Right? If you asked somebody, hey. Tell me the best sandwich.

I don’t know. I happen to like an Italian sub. Maybe Kobe likes a Reuben. I don’t know what his background is.

I don’t know if this is something from his his I don’t know. This is the best sandwich that I could think of. Well, why would you say that’s the best sandwich? All of these things turn into what the in the bad inputs that go into LLMs and the bad output that comes out of them.

So it’s really based on being able to take in this input and being able to put out things based on context and based on communication, and they’re all trained on top of the the, you know, sort of language that it sees out there on the Internet. And Donovan, I think you’ve you’ve discussed also in the when we did the panel before, you kinda broke down into a little bit more on the LLM side.

Yeah. Two two quick things, Donovan. I definitely wanna make sure and do, give an introduction for me. And secondly, yes. I was actually gonna say Reuben. So that’s a little bit, scary that you’re able to to, depict that from whatever my face was saying at the time. So, Donovan, yeah, anything you wanna add into that, feel free and and jump in and expand.

Yeah. Absolutely. Thank you. I think AJ might be, an AI bot himself.

But No.

It’s a it’s AJ. It says AJ right there. It’s not AI. It’s different.

Yeah.

Nice to meet everybody. My name is Donovan Brady. I run solutions architecture at RapidScale.

We help customers develop their own AI solutions. Right? So, I think AJ just hit the nail on the head. At the end of the day, I think people get caught up in the marketure, the marketing materials around AI, and get very confused. And they’re like, yeah. I want an AI. You know, I want AI.

How many AIs, Donovan? How many do they get?

I want five AIs.

There it is.

But at the end of the day, the AI is really just another evolution of your computer. It’s another evolution of the technology that we use every day. And it’s just solving a mathematical equation, as AJ said, to come to an outcome that, fits whatever your request was. Right?

And back to Kobe’s question before about use cases, I think the use cases there are infinitely many use cases. Right? And they fall into a number of different categories based on what you’re trying to do. It could be customer experience, as Kobe was saying.

It could be, product design. It could be content generation. It could be predictive analytics. Right?

It’s heavily dependent on the industry that you’re in. It’s heavily dependent on what you are particularly doing as a customer in that industry.

And then finally, what are the customers spending their time on?

At the end of the day, AI is just a tool to help replace human time. Right? Oh, think about how you use ChatGPT right now. Right? I use ChatGPT almost every day. I’m thinking about something, and I’m like, I could go pull four hundred people, or I could just have a quick conversation with ChatGPT and ask it to generate a more thought out version of what I’m generally thinking of. Right?

Take that and apply it to any business context, and you have an AI use case. Right?

Yeah. And I think what what I was putting out of that panel was the the amount of work that goes into getting to the point of that of the production of the the technology. Right? And I think for our advisors, it’s really good for them to know that so they can start to have that conversation almost in a checklist manner with their customers and the ability to guide them through, like, hey.

If you wanna get to here, there’s a little bit of the road map, and it’s always gonna vary a bit. But, you know, AJ, what are you seeing as far as, like, getting customers prepared to use AI and make it efficient? Because, like, if come what I’m gathering, if if we just drop the technology in a bad bucket of data, it’s gonna give you bad outcomes. Right?

So what do organizations need to do? And and this is and this is important to pay attention to for the audience because this is the products that you start to sell. Right? These are where you start to make the money.

It’s it’s all the preparedness that they need to do in in alignment. So, AJ, maybe walk us through what that looks like a bit.

Yeah. So the quick background on Expedience, we’ve been a cloud service provider for twenty years now.

And what we’re seeing now with AI is the same thing we’ve been seeing for the last twenty years as people want to go to cloud. Right? They think they wanna do this because it’s going to make them magically more efficient. It’s going to magically make things better. And when they don’t have a use case, they end up spending their money and their time in the wrong places.

So what we have found is we have to help them understand, like, okay. What is that actual thing you want to do? What is the outcome that you’re actually looking for? And once you start there and you say, okay. I have my internal my internal engineers and my internal staff constantly ask my subject matter experts questions.

But I have all this documentation.

How do I make my documentation more available to my engineers?

Great. That’s a great use case. Right?

Because I, as a person who writes a lot of documentation, it hurts me deeply whenever I find somebody who hasn’t read it because I’m like, I put a lot of effort into that. I wanna make sure that, like, I’ve tried to answer the question before you ask it.

But you’re because it’s not in a place that you’re normally thinking, and you just go, hey, AJ. How do we do this thing? And I can just answer it. That whole pile of documentation just kinda sits there and doesn’t do anything.

What you can do, though First, AJ, I’m gonna let you know that myself and most salespeople feel attacked right now, because the ones that this I’m not it’s not really an attack.

It’s more just a request from your friends who write documentation. Just read it first.

But what ends up happening is when you have this documentation in a place where maybe it’s not accessible, maybe it’s on an internal system, maybe it’s on a SharePoint, maybe it’s in Confluence, and you don’t necessarily just I can just pull that up very quickly versus I have Teams or I have Slack or I have Zoom, and I can just go, hey, Ajay. How do I do this thing? That access is what really is happening there and what makes that successful and why people do that. So when you can take that internal documentation, have something like a chat g p t, have access to that documentation, safely have access to that documentation, and be able to be a human like front end to be able to show that data to someone else.

I think that that is a powerful use case that’s actually really simple to deploy. The trick though is doing that safely, securely, and at a cost effective manner. So things like enterprises now expect role based access control out of the box. ChatGPT, Copilot do not have that.

They expect to be able to know that that data is secure. Everybody is scared to put their data in places where they don’t necessarily can wrap their arms around it. See Amazon Azure ten years ago where they were like, we can’t put our data in the cloud. It’ll get leaked.

Amazon and Microsoft weren’t leaking your data. It was your engineers not configuring it properly. That’s what leaked the data. So all of these sorts of challenges, you have to go solve.

And that’s what what we’ve built a lot of tooling around is to be able to provide that level of safety and security while still providing that level of access. And I did see a question here about integrations. And that’s really what this comes back to, and we’ll answer that question more full later. But in a lot of cases, it’s your data is not in one place.

I actually asked this question in the room at the pump at the partner summit. I said, how many of you are running internal file servers? And I saw nobody put their hands up. And I said, okay.

Office three sixty five, about three quarters of the room. Google Workspace, about quarter of the room. Salesforce, whole room.

All of these places have all of your data.

And so being able to pull all that data back together into one place so that you can actually get access to it and actually run things across it is really where the magic happens. And that’s additionally what we’re doing with our platform is being able to pipe that data in safely and securely so that you can then use it in a LLM in with an LLM that can read against it without necessarily putting your data at risk.

Yeah. Doug, I got a question for you. I wanna hit an analogy that I just kinda popped into my head. And what you’re we’re talking about with all the data and all the different places, I always go back to technology as like a house project.

Right? You’re always, like, remodeling. You’re doing something to your house. So the way that I look at the data is, like, those are your tools that you’re trying to get access to and the applications, and they’re all over the house.

And and what we’re looking for is a way to say, I need a screwdriver. And not only is it in one place that you know where it is, it’s in your hand at the time you need it. And that is the power of what we can deliver.

But first, we have to go figure out where all the tools are to put them all in one place and then and get them magically into our hands when we need them. You know, Don, what are you seeing from the security side of thing? And is that the biggest hesitation when organizations start to go down the road? Because there’s the desire.

There’s the pressure. But is that the the breaks that you see getting hit as you get into these engagements? Is it around the security of it? And what are you seeing to to really prompt them for that conversation?

Yeah. Absolutely.

I think, can I actually take a step back before I address that specific question? Absolutely. Alright. I want to just hammer home what AJ said because AJ, the first thing that AJ mentioned is the most important thing for anybody, whether it’s you as partners, the customers, us, it’s the most important thing that we need to consider when talking about AI. And that is what is the problem that we’re trying to solve. Right?

So Kobe asked, what is the work that has to happen? Well, there are really three questions that we should be asking. Plain and simple. One, what is the problem that you’re trying to solve?

Okay. Two, what is your expected outcome after solving that problem? Okay. I have a problem.

What do I want to get to by solving that problem? And three, do I have the data to support a mathematical equation to solve that problem? Right? Those are the three things to ask.

Back to your question just now, Kobe, about security.

Yes. So RapidScale specializes in highly regulated, hyper compliant, industries. Right?

So we are internally audited for PCI, HIPAA, HITRUST, SOC one, like, you base you name it, we’ve got the certification. So that means that the customers that we’re specifically seeing are all hyper aware and conscious of security.

But it’s not just because these customers and their industries. I think it’s also the nature of AI and this data that we’re talking about. So if we’re talking about ChatGPT or one of these LLMs that’s open source, it’s open to the world, it’s pulling data from a number of different sources, that means that the data that it’s getting access to could be malicious data. Right?

It could be bad intent. It could be bad data in general. Right? So I think, back to AJ’s point about RBAC, role based, access controls, you need to consider specifically where your data is coming from and who has access to it.

If you have third party bad actors that are putting bad things into your equations, into your code, into your data sources, there is potential that you’re putting viruses into your own applications, into your customer applications. Right?

That’s just from a security perspective. There’s also the deeper applications of just the the output being erroneous. Right? But from a specifically security based perspective, I think that’s probably the biggest concern that I see.

And I if you don’t mind, I just wanna I wanna highlight what you just said even more so for the TAs because you just turned that’s a talk track, in my opinion. As people are going down and customers are still trying to figure this out themselves. Right? This is a lot of new technology and new and new acronyms and new things like that.

So if you really wanna look at it from what you just said, Donovan and AJ, what you guys have put together, that that talk tracker on, hey. You know, you can utilize these public data pools, But in doing so, here’s the risk. And a lot of times that might seem as though it is very much, you know, a dot that would connect on its own. Highlighting it could put you in a different conversational position with your own clients from an advisor position and give them more insight than that.

And assuming that they don’t that they have it is probably incorrect. So that’s why I wanted to hit on that so, so much is that is a really relevant talk track because it becomes a warning sign of you might be going down this road, but you probably need some help because you could be doing it wrong and immensely, are immediately putting value on what you’re able to bring to the table.

So Sorry.

Off. I just really wanted to highlight that.

Yeah. No problem. Were you gonna jump in for a second, AJ?

Yeah. I I mean, when we’re talking about data and where data could possibly go, eighteen people just, their Otter dot AI summaries just dropped into the chat, which means all the stuff that we’ve been talking about here has been summarized and put into a number of different, various Otter dot AI summaries. So in case you’re wondering where what AI can do, it’s join a meeting for you and take notes for you. But it also means that the things that are set on that meeting are going out to this third party service.

So another thing to kinda keep in mind there that when you actually have a when you’re thinking about what you can do with AI, I do think it really has to come back to that nailing what is the actual use case. And, Donovan, you actually nailed it. It’s not just what do you want to do. It’s what do you expect to do.

Because it’s those are two different things in some cases. We expect to be able to have all of our support people have access to all of our documentation. What we expect is there’s a chatbot.

Right? Those are those are two somewhat different things. This is what you actually want it to be. And what we have found is that a lot of organizations wanna get down into the into AI, but really they’re just putting ChatGPT in front of people, which is a general purpose LLM, which, like, it can make an email go from mad to happy, but that’s not necessarily solving huge business problems.

Yeah. And I think this is where, from both of our standpoints, we’re able to get towards a cleaner outcome. We’re able to help customers do things and actually put their arms around AI in a way a lot of what we’re doing is providing access to the different models and allowing a front end to access them, whether that’s a chat function or whether you a client wants to write an application to use that to get to the LLMs, but it provides all of that. And, you know, Donovan’s in the same boat as I am is in terms of helping customers figure out what those use cases are and actually driving towards that.

And this is where you can leverage us as suppliers because we’ve seen this from our own sake, but we’re also building it. So we’re able to sit down with them and say, okay. Kinda walk them through. Where do you wanna go?

How do you wanna get there? I don’t expect everybody, all the attendees of this room to be AI experts at the end of this. Absolutely not.

This is a big, big challenge for a lot of organizations, and they’re getting a lot of pressure from not the technical side of the house. They are getting pressured from the business side of the house.

Can I jump in for a second there?

So Absolutely.

Great segue.

It’s everybody’s job on this call, including Kobe, AJ, myself, and every single one of UTAs on the call, to dis dispel the FUD. Right? Dispel the fear, uncertainty, and doubt around AI because I think it is such a new topic, and, also, entertainment hasn’t helped us very much. You think about iRobot.

I was just looking at, you know, figure two is coming out, and there’s a new Unibot or something like that. All these robots in the house, and people are getting scared. They’re like, oh my god. It’s it’s Terminator.

You know? So there is definitely some fear there, but I think just the newness of it makes people very confused where it’s actually not that different from anything else that we’ve done in the past. Right? So let’s go back to the security question for a second, Kobe.

It’s relatable from everything that we have done in the past. I’m sure that everybody here used to have LimeWire or Napster or some something like that. Right?

Like The AI is listening. The AI talks to the feds, Donovan.

Don’t give me this.

Right? But these things were things that were in the world. Right? Torrenting.

If you download a file from a malicious source that you don’t know where it’s from, you might have downloaded something that you didn’t expect. You tried to get a song, but you actually got these other viruses and all this stuff that just broke your computer. That’s the exact same thing that the AI is doing. It’s pulling from those same sources.

It’s pulling from those torrents. It’s pulling from the limewire. It’s pulling from anything it has access to on the Internet. Right?

So, that’s why the access and and permissions is so important. And if you break it down to the customers like that, like, hey. It’s just accessing anything it can. It it’s talking to the feds.

It’s writing down what we’re hearing on our conversation here, and then it stores it in a database. Right? That’s why we need to set up the proper parameters and design the solution, so that it’s so that it’s working properly. Right?

Alright, guys. So as we kinda wrap this up, there’s some really, good questions. I’m gonna take a, stab at a couple of things, and then I, I would like you guys come over the top, and then we’ll bring Adam on to conclude the call.

And, on the marketing team, I know I got a couple other announcements to make before I bring Adam back, back, so I’ll get to those too. I just don’t want everybody in the backroom to to freak out that I’m, jumping over some stuff. But on what we have, it’s there’s a couple of things. I think sometimes for organizations that you the conversation I always start with is, hey.

What kind of pressure are you getting on AI? Like, where is it coming from? Is it other executives? Is it sales?

Because now you’re you’re you’re gonna lead into a technology conversation, but you’re having just a human, like, pressure conversation. And it’ll and say, what use cases are have you had? Now the the number one thing that we’re gonna start putting together and make sure you guys have access to is really just a lot of different use cases that you can almost suggest and go, hey. There’s some here are some other things I’ve seen other organizations do that really help them along.

And if you’re getting pressure and you’re looking for something, this might just give them a little bit of insight on what they could go down the road and maybe it relates. So it’s a little cart before the horse, but this is one of those times where that actually works a bit. So you have that. Every other situation as they’re going through it, to have a goal without a strategy is really just like false hope.

So as you go through this and you say, okay. What’s the goal of the outcome that you’re looking for of this project? And what other things do you need to do to get it ready? That’s where you can really gauge where the customer is at.

If they go, well, we don’t really know. And you can the number one thing you can come back to when it’s anything around, like, a chat g p t style of AI is it starts with data readiness and then the security of that data. So you can go, hey. In you know, some insight here is a lot of organizations struggle with what you’re struggling through.

Here, we need to get your datasets ready. So we need to attack your data data sorry, data readiness, and see where you’re at on that. And I can bring in the experts to really dive into that. You have two great ones on this call, and AJ and Donovan, the organizations and Expedient and Rapid Scale that they represent and that have been doing this for clients now for quite some time.

Right? And and the knowledge they use of this is just cloud and the same variations that happened before, the same the same fears, the same doubts, or the same things that open up the opportunities for all of you. And here’s what that opportunity is. Clients have a need in a technologies space that they wanna go and explore.

They don’t know how to do it, and they don’t have the resources and skill set to do it. However, you have the access to that and to be able to put them together. And, again, that is the magic of what everybody here has the ability to do, and we’re continuing to expand what that reach is. It’s also overwhelming for all of you, I would imagine, to try to keep up with all of this to know all the different nuances, and that’s why we’ll continue to have these type of conversations and give you guys as much ammo and talk tracks as possible to kick start it and then to create that world class sales experience that you guys do for your clients and getting them through their projects.

I didn’t see I think that addressed most of the conversations in a little bit of a high manner. AJ, Don, do you guys see anything in particular that, and the questions that you want to address?

Yeah. I saw there’s, like, a set of questions here that that came in very recently. Basically, where do I start?

And I know that you were going there, but there’s, you know, asking more of a very specific, like, what are the set of steps? And the first step is have the actual conversation around what, like Kobe said, around the business pressure and why. Like, why is the business pressuring? Because you’re talking to a CIO, you’re talking to a director.

They’re probably not hearing it from their engineers. Their engineers are using AI for sure because they’re using ChatGPT. They might be using one of the code based LLMs.

But the CIO pressure of we need to make something, we need to use AI in our product, it’s usually just a squawking hand next to their head. It’s a lot of pressure either from outside, from something they’ve seen, from something that that is driving that desire. Get an understanding of what that is so that you can actually take advantage of it. There was an org you know, say an org uses three sixty five.

Microsoft three sixty five has Copilot. And the one thing I will say is that Microsoft has the benefit of being the king because they can put it in the OS. They can put it in the Office. They can put it into Teams.

They can put it in all these places, which is great. It doesn’t take into account every other piece of data.

That’s it. Right? So once you get outside of the Microsoft ecosphere, Copilot doesn’t necessarily have the ability to do that. Understanding those use cases and understanding their vertical, to Donovan’s point, when you’re talking about regulated industries, they don’t like to put their data anywhere.

So if they want to take advantage of AI for things like analytics or even something as simple as we wanna make sure that our folks have access to this tooling, but we don’t want them to be leaking credit card numbers or be leaking account numbers or Social Security numbers, that’s a big deal. So understanding, do they have a specific, you know, compliance need? Do they have a specific security need? And, really, this is a business conversation. This isn’t a technology conversation. When it becomes a technology conversation, when it goes from what and why to how, that’s when you bring in Kobe, Telarus, you bring in Expedium and RapidScale to actually talk through the how. But understanding the what and the why first is where USTAs come in, where you can provide some of that additional sort of, you know, warm and fuzzy feelings to them on actually getting them to where they wanna be without necessarily having to sit down and put it all together yourself.

Donovan, as we wrap up and, leave some time for q and a for Adam, any, any anything that you wanna hit on?

Yeah. I’d like to leave you with a practical example. So I’m gonna use the same example that we talked about during the panel.

I’m just gonna give him just preface.

This is a great one, and I’ve actually stolen it.

I’m sorry. Really? Okay.

It’s it’s literally the it’s it’s everything we’re talking about in a nutshell. So, this actually was a a TA came during the summit and said, hey. Got this customer. They are a lobster fishing company.

They don’t have much technology. They just have this ERP solution, but, you know, that’s about it. They don’t know how to identify the proper market price for their lobster. Right?

Right now, they literally phone a friend. They just call around. They’re like, hey, guys. What do you think the market price price should be?

They come up with something. Right? That’s a very bad practice, and they properly so said, AI could probably do this for us. Right?

There are a lot of variables into determining market price. There’s shipping costs. There’s gas costs. There’s all that stuff.

Right? So we’re talking to them. They have the business problem. We don’t know how to identify market price.

They have an expected outcome. We wanna have the ideal market price based on a number of variables that we’re talking about here. Okay. So then the third question comes, which is, where’s your data?

Their answer was, most of our data is in three ring binders that we’ve been writing and storing our market price data over the last number of years, for, like, thirty years in three ring binders on paper. Right? So great problem to solve, but very, very little aptitude or ability to to execute on it. Right? So we would need to extract this data from these three ring binders, put it in a database so that we could create an AI solution to then query the database, get the data, and then answer the question and come up with a market price. Right? So where do we start?

Back to those three questions. What is the problem that you’re trying to solve? Do you have a problem? If not, let’s do a work a use case workshop to identify some potential use cases where we might be able to help you.

Right? Second, okay, if you have a use case, what is your expected outcome? And third, do you have the data to support it? Both RapidScale and Expedient have solutions to help you identify help customers identify any part of this process.

And somebody also asked in the chat, about use cases per industry and things like that. Rapid scale, actually, we have, like, a a sixty slate sixty slide presentation of a ton of use cases that we’ve built for customers per industry, all those things. So definitely give us a shout, reach out, and we’re we’re happy to talk to you about all of that.

And and we’ll put that into Littlest University as well. We’ll make sure and get it uploaded so you guys have access to that in one centralized place.

Hey. Real quick, Donovan. I I love the story because I never thought that you’d be solving, like, a lobster market price Yeah. Like, that just kinda gives an example of how, wide wide ranging this this can go. Well, how large would that comp how large was that company, like, market size wise?

That’s a great question. Honestly, I don’t wanna take a guess.

I don’t recall the size of the company or revenue, but, I’d say pretty good.

Not not fortune five hundred. Right? Like Yeah.

They they were not fortune five hundred. They were big enough, but small enough that they were they didn’t have much technology. You know? Yeah.

And I think at that point, it’s not an AI conversation. That’s a this is effectively digital transformation. Yeah. Like, in the actual sense of real life, honest to goodness, digital transformation is taking things from three ring binders on paper to an actual digital function.

And I think that’s really the other the other part of this is if your data is not in the right place, you don’t get to talk about AI yet. You have to make sure your data is actually in a place to have AI even look at it. Because as cool as all of this AI stuff is, it is still not able to read pen and paper yet. So being able to get all of that data in is where the actual process starts.

And, again, that comes back to talking through. Okay. You wanna use AI to look at your market prices? Well, where is all your data stored?

What’s in three hundred binders?

Well, you don’t have an AI problem yet, bud. So let’s actually talk through that.

I think there’s a ton of different a ton of different capabilities that you have with AI. I see some questions in there around services and and what we’re each doing we’re each doing things differently there.

What I think I’m realizing is I I think that there’s a lot of great questions coming in.

Yeah.

I’m gonna grab, both yourself and Donovan and do a recorded we’re gonna grab all the questions. Yep. We’ll go through a record, answer them, and then add on some additional stuff, and then we’ll get that out to the market for everybody as we conclude this call.