HITT- Understanding AI's current impact and future opportunities- Feb 11, 2025

The video discusses the immediate relevance of AI technology and innovations, featuring a panel of experts from Telarus who share insights on market impact, data readiness, and security concerns. As AI conversations evolve, advisers are encouraged to guide clients through the complexities of implementation, emphasizing the importance of data cleanliness and accuracy. The rapid growth of AI solutions presents both opportunities and challenges, with a focus on understanding client needs and IT methodologies. The panel also highlights the significance of specialized AI solutions and the necessity for organizations to develop clear AI strategies. Overall, the session underscores the critical role of advisers in navigating the fast-paced AI landscape.

Transcript is auto-generated.

As always, your questions and comments are welcome and encouraged in our chat window for live q and a, which follows today’s high intensity tech training. Today, we examine AI. It is not the future. It is happening right now.

It’s transforming business operations from CX to security and cloud. Today, you will learn the real world technology, what’s here now, the innovations coming along, disruptions marketplace, and the biggest opportunities available to advisers right now.

Today’s panel of experts is comprised of State of the Union Telarus all stars, Jason Kaufman, cybersecurity solutions architect, Sam Nelson, VP of CX, Koby Phillips, VP of cloud, and all of it hosted by Telarus senior VP of sales engineering, Josh Lupresto. He may have a special guest in there as well. Hey, Josh and all. Welcome back to the Tuesday call. How are you doing today?

Hey. Good to be here.

Excited to come back. We we’ve got a lot of good things to cover here. Excited to have everybody on. You know, just trying to give everybody a glimpse of, a little bit about what’s hot, but also just the way that we have some of these conversations internally.

And you can hear how we talk about what it is that’s critical and some of the key updates that I think are gonna help you as you, you know, go into this twenty twenty five go to market strategy. So, I would love to kick it off here. I’ll give you a little bit of highlight. I’ve got four or five topics I wanna get through today.

We’ll see how far we get. We tend to get really excited, and we do back and forth. We might not make it through all of them, but I’ll give you a a glimpse into kind of what a couple of those are gonna be. So we’re gonna talk here we are a year out, from, you know, when we started a lot of these AI conversations.

We’re gonna talk about market impact and opportunity. We’re gonna go into a little more discussion around the data. We’ve talked a lot early on about the data and the classification and things like that. Koby’s gonna drop some things, about that.

And then we’re gonna go, hit a little bit on the security angle. Jason Kaufman’s got some some topics around that, the opportunity, the product, you know, some of the gems, around that. We’ll get into a little bit about tools and marketplaces. And if you hang on until the end, Jason Kaufmann and I might debate deep seek and we might get into that.

Some pros, some cons, and we’ll go from there. So hang with us. I think we got some good stuff. And as always, if you got good questions in the chat, fire those away, and we’ll get to those as soon as we can.

So first off, Sam, Koye, Jason, welcome.

Sam, I wanna I wanna kick it off with you. Here we are a year out.

We’re talking overall market impact opportunity.

We’re seeing Matthew McConaughey on a lot of the Super Bowl commercials. We’re seeing Salesforce. We’re seeing Agent Force, all this stuff. Just walk us through. What are you seeing? Where are we a year in?

Yeah. There’s a lot going on. In fact, it’s moving at a faster pace than we can keep up with. But the good news is that it’s not just us.

It’s also end customers. Right? So we’re all kinda trying to figure this out at the same time, and we’re in a a really prime position to be able to guide clients, in a lot of these AI conversations. But yeah.

Like, after watching the Super Bowl, you know, you’re seeing, ads not just with Salesforce Agentforce, but also, OpenAI, but Chat GPT, out there. They had a spot. And so, clearly, it’s more relevant than ever. Now, what’s interesting I wanna point this out.

What’s interesting with, companies like Salesforce and Microsoft, right, is now they’re actually capturing sector really, really quickly. In fact, I think the goal here is for them to take over that agent desktop. And what I mean by that is they’re actually packaging up and enabling end users to create agentic AI. In other words, like, AI agents within the platforms to handle a lot of the interactions now coming in, which we’ve never really seen before.

They teased it out a little late last year, but, it’s something that we definitely wanna keep an eye on because we’re specialists in this space. Right? And when we start to see some of these other platforms now singing the same tune, something we wanna watch out for, but also figure out how we guide customers to the right solution that’s best fit for their needs.

So as you think about that, I wanna I wanna get Koby’s perspective here. But as you as you think about that, how do we handle how do we how how do we see success in handling those conversations? Right? To your point, we’re we’re very proficient at CX.

We’ve got a lot of experts in CX, both in the partner community, on on the broader Telarus team. But how do you want the partners to be thinking about here’s where, you know, the the Salesforces and the Microsofts are now? There’s ways that we can certainly help with that. But how do you want everybody to kinda consider when the customers start asking questions around those?

Talking to me or you’re talking to Sam?

Sorry. Sam. Sam. We’ll we’ll finish the we’ll finish the the Salesforce thought.

Guys, we’re doing so well. We’re doing so well.

Hey, guys. I I just wanna pause really quick. Josh, you really need to get it together right now. There’s a lot of people on this call.

Now we can get it in the next There’s I thought it was just us.

Alright. Alright. Sam, I think the question was yours.

So Dang.

Dang it. Okay. Well, no. I’m kidding. I’m kidding.

No. I I think there’s a lot of things that we can do that actually a lot of our TA don’t realize, in that there’s this whole segment of the market where we can actually help customers with the managed services, the implementation components of these different CRMs.

And so we’re actually looking more into that. And this is where, like, Koby, where I’m sure, Josh, this is where you were going, but this is where we start to have the cloud conversation, right, about the data and how it’s being managed and the the, you know, how the data flows because, essentially, the data is not clean, then it’s that is trash in, trash out from an AI perspective.

Right? And so when people think, oh, I don’t have access to selling things like Salesforce or Microsoft because they sell direct to consumer, well, the reality is those platforms do outsource that professional service work out to implementation partners, right, who we actually can work with.

Yeah. And I think that’s where I get excited when I hear when I see all of that, all the commercials, all the stuff that Sam just laid out with Microsoft, Salesforce coming to the market, it creates the same opportunity that’s already been there, which is just alright. Let’s take a step back. Sam, how big of a deal when AWS jumped in the in the to the contact center space with is AWS Connect.

Right? Everyone was like, this is gonna be a game changer, this, this, this, and this. And then as customers got into it, it is a game changer if you have four to five people that can build on top of the platform, manage it. They gave you it’s like they gave you a LEGO kit, and you have to put it together, where a lot of customers don’t want a LEGO kit.

They want the finished product, and they want it managed, and they want the resources. And so what we’re seeing now is a crossover between a lot of the CX conversations into the cloud infrastructure, and then as Jason will tie it together in a second, all of it has to be protected. So if I’m an adviser, I’m looking at this as the opportunity to go, hey. What’s my customer feeling?

What’s the what’s the pressure? You imagine a CIO who’s already feeling pressure, trying to get away from it at work, watching the Super Bowl. Now you got Matthew McConaughey kinda mocking you. I’m like, hey.

You don’t what’s your AI strategy?

Like, that is real. That’s human. That is human. That is what’s happening to a lot of IT leaders. And so the the advisers have more opportunity now to step in and go, hey.

There’s a process. There’s a road map to this. We got you. Yes.

There’s these really great products. Do you have the team that can do it? And the cool thing about the products is they do make them simpler. But to Sam’s point earlier, if you don’t have everything set up, it’s like trying to put something together from IKEA.

Right? And you’re and you skip three pages, and you don’t have the right parts, and it’s gonna fall apart, or it’s not gonna go as well as you think. It might it might stand up, but it’s not gonna be sturdy, whatever it is that you’re building. And so you gotta get your data aligned.

You gotta have data readiness. You gotta have the security piece. You gotta have, like it’s almost like when companies early on would transform into cloud. It wasn’t just a technology change over.

It was really a managerial style change and a cultural change for the IT department all the way through the organization. As the AI comes in and impacts it, I don’t wanna overwhelm anyone, but they are overwhelmed. And that’s the opportunity to come in as an adviser and go, hey. We have a playbook here.

We have a lot of resources, and we’ve done a lot of that. I mean, Jason, I know firsthand you’ve been involved in a handful of, like, AI, centric deals on the infrastructure side, like, LM, data cleansing, things like that. To go along with Sam, you guys have been selling AI in your space for this is old. Right?

Predictive and conversational AI has been around for a minute. So I think I think we’re in really good shape to capitalize on it, but it is a short window in my opinion. People are gonna figure this out. There’s gonna be more and more prompt engineers.

There’s gonna be different AI data scientists, and there’s gonna be all these new terms and new roles, and companies are gonna hire them in, or there’s gonna be specialized consulting. So I would say strike while the iron’s hot, while there is a lot of, like, confusion and pressure equals opportunity in in looking for guidance, and you guys have that in space. So I I just see it as a really, exciting opportunity to expand and and grow the conversation for our advisers.

Yeah. I think so. Kaufman, you wanna you wanna weigh in on that? I mean, you’ve you’ve seen a lot of these.

You’ve seen some of these discoveries. We’ve been talking about data readiness. We talked about it at partner summit. What’s what’s the customer sentiment?

I think they’re heavily surprised when they get on a call expecting to get all these different ideas on how they can implement AI. And the first question is, hold on. You know, let’s look at the data and make sure that we’re classifying it, make sure you know where it’s really at to make sure that any of these models that we’re gonna be using is accurate and also the data is protected. In fact, I was just looking at, I I follow that TLDR thing that gives you all those summary news stuff, and they said that point zero zero zero one percent, if the data if the data is incorrect by that much, you’re increasing exponentially the hallucinations that are coming out of an LLM. So if you’re looking at a very small amount of data that is incorrect could affect exponentially more what results you’re getting out of an AI implementation.

So it’s very it’s very important to do that measure thrice cut once type thing to where you’re getting in to see what data are we getting it, how are we getting it in, is it already digitized, and then how are you making sure that people that have access to it are the ones that need to have access to specific data? Because you can’t use the same models and the same data for internal and external LLM provisioning because you have to have stuff that is confidential, stuff that is public facing. And there’s a lot of thought and process that goes into all this stuff. So when you’re telling somebody that they’re really excited about something and they wanna implement it right now, right now, right now, if you’re saying, hold on, you know, let’s look at what we’re gonna be putting into this first, and then we’ll implement it and get you get you up and running.

It it adds that truth factor to them that, oh, this person kinda knows what they’re talking about. You know, they’re they’re not all about, hey.

Let’s just throw all these different skews at them and implement AI before we’re ready for it.

Yeah. Yeah. There there might be, a training coming out, if you guys are able to to join the CloudSense.

And we actually break down the eight steps to preparing your data for AI because it is, an eight step process for every organization. We’ll go into the details of that, and then we will have that as collateral for you guys to be able to take and then walk through and and do all that stuff with. But, Jason, do you mind, sorry, to to just break down what an LLM really quick at a high level, so everybody kinda falls along. And I always attribute this is a weird analogy, but imagine a bunch of apples going into a barrel, and that’s your data.

Right? And You’re collecting it. If you have bad data or you didn’t, like, look to make sure the apples are good and there’s a couple of worms in there, not only does it affect that apple, but it’ll go through and start, messing up everything that you have in that barrel. So that’s one way to think about, like, cleansing data and why it’s so important.

And you’re trying to pull that information out to be intelligent, and it’s now corrupted. Right? So in LLM, Jason, if you don’t mind, just say now you kinda hit on it a little bit. Just wanna make sure you can, break it down simply in a non over technical way for everybody on the call.

Okay. So how would I explain this to Koby? Yeah. So think about it. To help.

So think of it as like a digital brain to where you’re taking all these mathematical algorithms and you’re giving it a ton of data and you’re telling a machine, hey. Based on all these all these data, I need you to learn how to understand it assuming somebody’s gonna be talking to you in human language. And then how are you gonna process it, and then how are you gonna get a response? So an LLM is basically the human factor that says, hey.

This data I’m going to to receive, I know I’m understanding it based on all this data I was provided. Now I need to give something back, in response to it that is educated. So then the LLM is the brain portion, but then there’s also models that say, okay. Based on based on something, how do I go get this research so I have knowledge to give that response back to to the human agent?

So, for instance, you know, take the the original piece of it, which is natural language understanding and processing. I’m gonna put something into it, so it’s gonna take it and break it down into individual tokens. So that’s basically saying, I’m gonna take a sentence. I’m gonna understand it basically words.

So what comes next in the, you know, language? You know, you have adjectives, verbs, nouns, and all that stuff. I’m gonna take all of that. I’m gonna break it down so I understand it.

I’m going to put it to a, a knowledge base that says, hey. Based on based on my understanding, I’m gonna get an educated response here, And then it sends it back to the large language model, the LLM, that formulates a response that somebody is going to then understand it on the receiving end. So the LLM is the brain function that takes everything and consolidates it till somebody can understand what is being said and what is being understood.

And so that’s driving so much, goodness. Right, Josh? Like, in across the board and then how it dips into to the CX space and made that bigger impact. Like, let’s set on top of the conversational predictive and just give it even more intelligence to react in, appropriately.

It’s it’s this is the perfect tie in to to expand all your technology conversations. But if you guys have noticed the number one thing we keep talking about, it starts with data in in all of that stuff. So if you kinda follow the data around, you’re gonna be able to land in some really great conversations. Thanks, Jenna, Jason.

Yeah. No. And and just to give an example, because I know, like, a lot of folks ask for use cases around this. So, like, one prime example that was a global example here is that, like, last year, tennis tournament, Wimbledon, they actually used an AI tool to, provide all of the app users for that tournament with information about the players.

Well, little did they realize that they did not protect or look through all of the data that they were using, right, in that LLM, and it started feeding from outside sources and was actually giving false stats for all of the players on the in the tournament. And it was really bad. Really, really bad. They’re still recovering from that.

So, just going back to data cleanliness, I mean, it’s a big deal.

And when companies say, oh, I want this AI feature and that, it’s like, well, again, trash in, trash out. Right? So you have to make sure the data is clean to actually have AI effectively work for the company.

Yeah.

Yeah. And and, like, the other big thing to be aware of, as you guys are having these conversations, your target audience, guys, on this organizationally, the personas you wanna attack, CIO, CTO, CEO, CFO, CXO, any c level is gonna be open to this type of conversation and probably will hand walk you into the the right members of the team to, to expand on it. VP level, you get into an IT director stuff, you can you can start to get a lot of information back by saying, hey. What’s the what are they tasking you with?

What’s the road map and things like that? Here’s what we’re able to do if you can get us to help climb the ladder. But I just wanna make sure you guys are understanding who to target with these conversations, because these are high higher level business conversations that will get you into a different seat if you need to inside your own accounts. And it’s really gonna be that consultative approach of, like, let me understand what your vision is, where you’re going, and here’s what we can do for you.

And it and it seems, Koby, it seems to me like there’s sprawl. I mean, like, we talked about initially when cloud came out. Everybody was excited. Everybody wanted to move.

First, it was, it’s just for dev. It’s not for production. Then everybody started testing it. Everybody started playing with it.

You had people jumping into this hyperscaler, that hyperscaler. It seems to me that this is kind of proliferating in the same way. And to your point earlier, it’s moving so fast, but I think we we our role in this is to give guidance and say, I don’t care whether you’re working on a CX project, a security project, a cloud project. If we don’t back up and look at these components, I think we’re finding sometimes in these opportunities that people aren’t even aware at what departments are looking at what.

There’s so much product being thrown at these c levels down to the IT managers.

To your point, we have to bring it back to do they are they even aware of what they’re looking at or what their other folks are looking at. Right?

So, Josh, if you don’t mind me adding a hidden gem right there, thanks for the for the coin term earlier.

So I actually had a really quick close, last month, to where somebody came asking, hey. Do we have access to ChatGPT or Google Gemini?

And, you know, because we have users that are already using it, but we wanna make sure that we can protect that data as quick as possible. And they didn’t have the they they didn’t have the internal expertise on their internal data to kinda know, okay. We wanna go build out a Copilot instance, or we wanna tune or create our own LOM, or either even scared about using a fast paced application in the CX world. So what we did was we actually went to, Expediant, if I can talk about suppliers on this one.

And what they do is they front end any of those foundational models. So if you go to Google and you type in large language model or AI, you’re gonna get those Google Gemini, Anthropic, ChatGPT, OpenAI. And what they do is they front end those exact models and they take out that that PII, PHI, any of that confidential or identifiable information and use that that actual model itself. So the users are pretty much using the same exact thing.

They just have a different chat window that they’re using, but the company data is already protected inherently. And then they’ll add functionality to it after the fact. You kinda see what the users are using it for. So they’re getting the discovery information in house by just letting users use it day in and day out.

This was actually a three week time frame from first call to close just because it was so effective to get them up and running for the enablement part of it.

Well, fun fact, next week, the Expedience AI team will be on to break that down in even more detail, and and the other ones that they’re seeing. So a little plug for next week, and special guest host Dan Foster on that one, hopefully. So, the, the what I hear out of that, though, is, like, so exciting, because you’re getting multiple ways to go in and monetize this technology.

You we looked last year and we’re like, hey, here’s what we see on the horizon. And then to see a lot of suppliers execute on that vision and productize it and lay it out there, you see a lot more companies that are wanting to turn over their entire database. Right? And you’ve got companies like Entirety, Expedia.

You see, like, we’ve talked about Rackspace, rapid skills, AI components, and there’s more and more ECI, etcetera. Where we used to, we’re, like, hoping we could add throw out two or three about twelve months ago. Now we can throw out ten, twelve that are doing this. And I think that’s also kind of the head spinning part for the advisors of, like, where do I take this and where do I go?

Josh, I think you guys are doing an excellent job of staying on top of it and figuring out how to pull it through in every component. How many I’m gonna flip and ask you a question.

Taking over as host for a second here.

What do you see, you know, with your team? How many deals are getting opened up that start in one space in technology and AI pulls it through to three or four others?

Oh, any of those conversations. Because I think what we’re finding is people that we thought know exactly where they wanted to go at on the customer side really aren’t sure where to take this, and they aren’t they they really weren’t aware of how many different things that they can already do that are in this ecosystem. Meaning, if they came in and said, hey. I just I wanted to go down this one road for, you know, just an LLM build. They didn’t know that there’s help out there to do consulting around the data readiness, the data preparedness software around that implementation support around that, and really how it all ties together. And so I think more than anything, you know, back in the day when it was CX, we’d ask a discovery question on a CX deal to go, where’s your database? Where oh, it’s on this Microsoft SQL on prem.

Well, that’s not really gonna integrate super well with a cloud based SaaS API driven, you know, UCCC provider. Have we thought about moving that to, you know, just some cloud based VM and hosted it there and put it behind a private subnet and things like that? Well, okay. Well, what’s the long term security strategy?

What else how are your rest of your endpoints secured? And so I don’t to your point, this is this is different though. This has got a this has got a rate of proliferation on it that I just think is is wild because everybody I don’t know. It just seems exponentially more driven that people are being tasked to figure out to do this, do it now, do it faster, outpace our competitors.

How do we how do we how do we build a moat around all of these things? And so it’s it’s every conversation, because it has to be because it touches everything. The data touches everything.

Well, it makes sense why everybody’s so excited. Right? Like, whenever cloud, you know, hit the scene, there were arguments. We we we spent more money and we didn’t realize the business outcome.

We did this. And, you know, you could point to them and said, well, you did it wrong or, you know, whatever it is. But their viewpoint was still their viewpoint. At this, at the rate that AI is going, you’re hearing and you’re seeing immediately business return.

And so that adds that pressure because if it falls into, you know, CX, like, there’s all of these amazing things that come back through of, like, better customer experience ratings, higher customer retention, all this stuff. And I, you know, I don’t wanna step on Sam’s toes on, like, what the impact is there. And then what Jason’s just laid out, and not only is it fun and exciting, and it’s, you know, boundless as far as, like, opportunities, it’s really, like, okay. How do I add to the bottom line?

Again, that’s the opportunity because everyone’s so excited. They’re gonna go out and they’re gonna do it wrong, and it’s gonna be the same thing in two or three years. Yeah. We tried AI, but it led to this, this, and this.

And so, like, the more that you can get ahead of it and say, here’s the playbook. Here’s how you work through this. Here’s here’s how you stack it and pull it through.

The better our advisers are gonna be setting for not just one or two deals, but multiple opportunities within the same accounts.

Yeah. I love that. Sam, you wanna go I wanna I wanna come back to yeah. I wanna hear your perspective, Sam.

Yeah. So I think the the main theme in this call that I think we’re all seeing is the extensive crossover between these swim lanes. If you haven’t taken it away yet, it’s that. Right? Because here’s what’s happening.

Now more than ever, we’re seeing all kinds of decision makers in a deal. Everything from a chief revenue officer to a chief marketing officer, which is usually now where a lot of the CX, exploratory sort of discovery calls are happening because they’re seeing all of these cutting edge tools that are revenue generating, but then they have to go to procurement. They have to go to IT, and then those folks get involved and say, hey. How secure is this?

We have to check the boxes. And then you go to operations or, you know, the folks, looking at the data, and they’re like, well, our data isn’t ready for this. Right? And so it’s all driven kind of by that data conversation at the end of the day, but reality is it’s all interconnected.

And so we’re seeing more and more deals where they may start with a particular persona. Like, Koby was giving you some some, personas to target an organization, but it very well may include several others, and it could lead down, a totally different path than initially intended. So I think what’s important to realize is that you have to have to have to incorporate this sort of born in AI approach. Whether or not you think you have, you’ve been born in in the AI era, right, where it’s now more relevant than ever.

You have to be able to sell to that, and you have to be able to take this sort of agnostic approach because, again, you have to be able to guide those customers to those conversations. And and there was a question in the chat about, hey. Can Solaris AI experts, join discovery calls? Absolutely.

Hundred percent.

Bring us on, and we’re more than happy to have those conversations.

Yeah. I think that’s the key part in this is you’re we’re throwing a lot at you. You’re never alone in this. I I I even if you don’t even know where to take this, but you’ve got customers and prospects that you’ve started to drum up this interest, you’re you’re leveraging the things that Sam talks about in the ascends, you’re leveraging the things that Koby’s gonna talk about. You’re you’re not alone in this. Koby, I know you and I, we talk about a lot about how we do some of the process, some of the architecture.

Would love your thoughts on how we help partners in that. I think you got a cool angle on it.

Yeah. I would say process wise, it’s trying to figure out where they’re at with their IT methodology, which is gonna open up a ton more, a ton more. And I saw a a question from Hans or more of a statement. It’d be cool to hear some case studies. We got you. We’re gonna throw some out here as we conclude.

Anybody that I go back ten plus years with from Integra, they get what they want on these things. So, Hans, hang tight. We got a couple good ones for you. The, the the the process or the methodology that I love to look at is figuring out and you guys will hear this from me all year long.

And it it’s because it it works really well. You’re gonna figure out, alright, are we selling products or am I selling services to this client when it comes to the cyber and cloud space? Right? And it it’s an easy conversation app.

Are you guys insource, co source, or outsource as far as your managerial style? If they’re all in source, we’ll go through some trainings and some stuff to maybe move them to co source. If they’re already co source and outsource, you’re gonna be set up for services conversation.

If they’re like, hey, we do everything internal, cool. You’re gonna sell them some products, and we’ll have those conversations and those talk tracks locked and loaded. But what it also does is it opens up a really good opportunity to start changing the perception and account. If you sold network, you sold UC just by simply getting to understand their business more and then giving them little nuggets to where you can bring in resources to make an impact.

As far as doing discovery calls and things like that, I personally really like to to jump on as many as I can depending on schedule.

The one that I was on that I and you guys have heard this before, if you’ve heard me talk, so you’re buckle up. You’re gonna hear it again.

It’s it was leveraging AI and AI engine to help drive in the call. So first, you know, Sam, I know you go over this in-depth. But during the call, of course, it was an AI, a call to explore what they could do. We got into their data. Their data was actually in pretty good shape. They know that they had to do a little bit of cleanup.

This was a, an organization that’s, you know, fifty plus years old. And as the second thing that everybody always looks for is, well, I don’t really have a great use case on how I where I wanna take AI. So I just threw in the name of the company into an AI engine and said, give me top five use cases for AI that they could utilize this for, and read it off to them. Let them know where I got it because that the wind is not originating the idea.

It’s just getting them to jump and and go into one. And so they picked number two. We’re, like, a month into their ProServe engagement to cleanse their data, to start implementing their deep archive data, to then analyze it, and then put it back into play. So it was a simple, simple request when it came down to it, but it was just something to kick start them.

And utilizing AI give me like, here’s five ideas. What do you guys think? And the guy’s like, I like number two. It was like he was picking off a menu, and off we went.

Because they already set budget aside to go through this, and they just needed to see it, see it carved out. Jason, I know you’re you’re in a lot more deals than than I am. You probably have a couple other really good use cases or, case studies that you’ve seen hit with AI. Right?

Yeah. I’ve honestly probably probably have two a day now it’s on it’s on averaging.

I think if you would’ve just stone faced it no, and then just left no. I don’t have any. But, yeah, like, which ones, like, which, what are a couple that are top of mind?

So, one actually last month, it was a a company in the oil and gas industry.

Coincidentally, I’d I just bought some of their stock right before that because it came up on one of those, you know, recommendation sheets. So it is a sizable company and recognized name, and they wanna leverage AI to see where they should look at drilling next and kinda see where they should look at land grabbing and all that stuff.

So when it came down to it, like, they they the the partner brought me in because he said, they’ve already looked at some some AI models that are already out there. They’ve tried, you know, building this on their own, but they can’t you know, they they’ve been struggling a lot. Is are there ways that we can help them? So came coming into the conversation, I just knew that they already had some previous experience.

It wasn’t successful. So but I had no other data on that. And after digging into it, kind of what what they wanted to do was they had all these different databases out of all these different use cases where they were success successful previously. They have accounts in all these geography platforms that kinda show where, like, hot hot areas are for potential drilling.

They they have all these different models to where we even have to go out and do some screen grabbing from a public website in order to pull that in and create our own database, and then feed that into a data pipeline, which basically says, we’re gonna grab data from all these different areas, and then we’re gonna correlate it in one single single area, and then give them some visibility on it. And then we’re gonna use that data to train an LLM, that large language model. Somebody can say, hey. Where should we do this next?

And then that gives you that kind of that future data. So the data pipeline will allow you to look at, you know, data that the historical type stuff, and then the LLM will allow you to look at future type, you know, road mapping type, information. So we have this large type of, engagement where it’s periodic milestones to where we’re taking the data, consolidating it, and then we’re building them a custom engine because their problem was they’re trying to use those foundational models like CHAT, GPT, and Gemini, and it wasn’t as accurate as they were looking for. Again, you know, all this is about hallucinations and data accuracy.

Some of these companies are so specialized in their vertical, they have to custom develop it. And we have the the the partners that can physically do that and host it within their own infrastructure.

So don’t be scared. If something sounds like it’s way out of left field, that’s where we specialize. Hey. I played left field in high school.

Do you mind, I think I I think we’ve talked about this before in the, the partner had sold them. Was it network? And just was asking kinda like, hey. What’s on your roadmap for AI and got into the conversation and grabbed you? Is that the is that the origination of this from the partner side as well?

Yes. Yeah. Started in connectivity network stuff and then yeah. There was like, what’s your a AI rep you know, what’s your AI plans or goals? And they just started.

And you probably joined, what, three to five calls, kinda getting them prepped and then getting them over to the right supplier. Is that about the process that you saw on this one too?

It was it was, two calls, customer facing their agnostic, and then calls with with a couple different suppliers to give them a couple different viewpoints on what the solution would look like. So prepping them, and then, yeah, the the agnostic call after the fact to get, you know, kind of feedback besides the, you know, supplier customer calls. Yeah. Long answer.

So super complex thing that was just opened up by one simple question of what’s going on with your AI road map.

And just just to let you guys know from the from the partner community, training your own large language model kinda starts at the forty k realm.

That’s just from a resources perspective, and then you have, like, the management, the pro services on top of that. So this isn’t like a small endeavor. This is this is investment that these companies are still gonna recognize multiple x ROI on it, so they see the value in it, but it’s very lucrative from from an adviser perspective. And all they did was just bring in a Telarus expert, and we we ran with it from there.

Yeah. Jason, when we’re talking about pricing, you bring up a really good point. There’s a question up in the chat earlier, about small businesses, SMB taking advantage of AI tools. And I think right now, we have to realize is that a lot of the AI components are built into solutions out there in our portfolio that do cater to SMBs today.

Oftentimes, we just miss it, right, because there are so many of them. But the reality is SMBs can absolutely still consume AI tools and use them, through the existing providers we have. So if you need guidance there for some of your SMB clients, feel free to engage us, and we can kind of review what’s available for them. Because a lot of them just don’t don’t realize it. And, again, so many suppliers to keep track of.

Sam, it’s really made an impact in your space. Right? You see a lot of your suppliers that added a lot of AI backed, tech to their seat licenses that enable a lot more than they used to for that SMB client, and then that can that can certainly, trickle into what we’re talking about here. And, guys, to let you know, a lot of those SMB clients wanna do and analyze some of this stuff to go to market.

And there are, there are companies in the portfolio, many of which we mentioned, that will take those on. Jason kinda gave an idea what that expedient, product looks like. That’s not an overwhelmingly expensive product that an SMB can jump into and then, start to tie into some of these other, components that, like, the suppliers on the CX side of the house, like a Zoom, Nextiva, Vonage, etcetera, and you start to build out a very automated, very intelligent, baseline of technology for them, That’s all hosted and managed. That’s the that’s the great part about SMB.

Let us go do it all for you, and more and more there’s there’s the willingness and the desire to do it. Imagine a company that starts their business today. They’re not ordering pots lines anymore. Right?

They the way that people, take in voice technology is by Zoom, RingCentral, all the other ones that we’ve already mentioned. And so, this really just cultivates more ability to go, hey. We got a company that’ll come in and do all of this for you. And we have companies that specialize, especially in the Microsoft stack on that SMB side as well.

So there’s not gonna be a company that’s too small to do an ex to explore. It’ll just kinda come down to, like, hey. What exactly do you feel comfortable with budget wise?

Yeah. Do we, do we wrap it on deep seek? Should we talk about deep seek, Koby?

Yeah. So, I have one job. I’m just gonna reveal the curtain here. I have one job to do, and Josh is like, make sure you bring up DeepSeek for me and Jason to talk about it.

So, guys, I’d love to get your opinion on how DeepSeek has hold on. Let me double double check it. How DeepSeek has affected the marketplace and potential ripple effects of what’s going on there. Maybe give a little background exactly what it is for anybody that that might not be up to date.

Yeah. I’ll kick it. And then, Kaufman, we we can go back and forth here because I wanna I wanna take your angle. I think DeepSeek was interesting.

We keep hearing a lot about these models. We’ve heard a lot about JAT GPT. We’ve heard a lot about Llama and, you know, Meta’s model. There’s a couple of these big models.

Right? Big, big models.

And they’ve cost a lot to train, you know, upwards of a hundred million dollars to train to train GPT four. They’re talking about a billion plus to to train GPT five. And then lo and behold, out of nowhere comes this model, DeepSeq, r one. And so this is a model coming out of China, created by a guy that manages a big hedge fund there.

He’s used it for kind of some of his quant trading algorithms and some really pretty complex things. He’s piled up lots and lots and lots of GPUs, and and then all of a sudden this model comes out. And the the thing that hit the wire that I think blew the stock market apart that day was they did this allegedly for six million dollars. So I’ve I’ve I’ve read some interesting takes, hot takes, true, not true, but, wanna set that as the foundation.

Kaufman, your perspective, there’s security components here.

We can go into some of that privacy policy. What’s what’s your take on DeepSeq as you’ve seen it kinda compare to the others?

I mean, honestly, I’d I’d I’m still kinda undecided on the the overall impact. I mean, from an efficiency perspective, I think it’s revolutionary. They used open source models. They used synthetic data to train, and they’ve come in a lot cheaper than it is for any of those other foundational players that that are out there to where it’s it I think we’re gonna see a battle for power power from cape compute capability.

So who can who can hit AGI first? You know, that’s the biggest thing out there. Who can actually clone pretty much a digital human to do everything that we can do. And then the other one is efficiency because we’re running out of power and resources in order to to power this stuff.

So I think DeepSeq is actually a leader in the space on the on the efficiency side. And actually, there’s actually, what is it, University of Berkeley. They just released that they train the lowest DeepSeq model that’s, like, six hundred and eighty million, or billion parameters that can actually fit on a small local endpoint. They’ve actually, trained it for the core functions of it for thirty dollars, which is which is amazing.

Because now anybody can be a player in this market space. And I think it’s gonna change how people change, how they how they train these models because the idea on using something that that is open source and synthetic data is gonna be kind of the the future path to it rather than trying to consolidate all this all this, human made data and creating stuff completely from scratch. I think this is probably gonna be kind of the path forward that people are gonna see.

Yeah. It’s gonna have a ripple effect on a few things if it hits.

Right now, any any AI company with a good business plan is getting a ton of funding. Right?

That’s also trickling into the data center space and how we’re seeing that explosion. So there could be potential ripple effects here if this is a viable, scalable model as Jason just kinda laid out.

Do you mind if I ask a question? I know this wasn’t this isn’t part of the program, but I’m gonna ask.

Josh, Jason, what do you guys see from, like is there bigger security risk in the way that you just laid that out for back end?

So my my x feed lit up that day. You know, the stock market was red. I the the the first thing that I went and did was, alright. Let’s where did this come where did this come from?

I mean, DeepSeek has been an open source model, so there’s been a lot of improve and expand upon that. Yeah. They had some old stuff v three that was trained the same way. And then reasoning the reasoning model comes out.

Right? We this is where we’re stepping from pre training on a bunch of data that they’ve pulled from all sorts of places, all the models have. And now we’ve moved up to GPT has o one reasoning. That’s great.

Now we’ve got r one from DeepSeq. That’s a reasoning model. So it’s it’s trained to to Kaufman’s point. It’s been training differently.

So I’m like, alright. Well, let let’s go explore this. Let’s go check out DeepSeek. And there’s this little button on the DeepSeek login page says privacy policy.

Who what what who besides the nerds reads the privacy policy. Right?

The first thing that I see in that is it says, Did you read it?

I I did.

Okay. Cool. I just double double checking that.

Thank you.

So it says, where we store your personal information. I’ll just read this. Just one sentence, two sentences. The personal information we collect may be stored on the server located outside the country where you live.

We store the information we collect in secure servers located in the People’s Republic of China. So when it comes to this idea of did you know, I think people are looking at this going, oh my gosh, this model’s amazing. I can do it for a sixth, a tenth, a twelfth of the cost. It’s faster or it’s different, but there’s privacy concerns.

So in my mind, everything that we’ve talked about so far, we’ve got sprawl all over the place. And if the businesses are fine with their data going to China, then great. But I my guess is that a majority of them aren’t even paying attention to that it is. So we have to have that as a conversation of, did you know?

It does this matter to you?

And to make matters even worse, that database that they’re storing all this information was unprotected completely until Wiz figured it out when they did a vulnerability scan of all the infrastructure over there. They found out that the database where they’re storing all the all the inputs that users were creating plus confidential information was publicly accessible.

So DeepSeek themselves, even hopefully, you know, whoever tries to do do this next protects it a little bit better. They had everything out in the open, until somebody figured it out. And luckily, they had it passed within the hour. But still, how long was that all that information out there for the public to see? I mean, that was a huge miss from a cybersecurity perspective.

And also, I know you’re big on stats. Sorry. Go, Koby.

Oh, no. No. Go ahead, Jason.

For, you know, talking about China LOM, fun fact is China requires any large language model, company that is, you know, trying to be an AI start up, they’re required to submit five to ten thousand questions that the LLM would not respond to that would be, like, answering, in opposition to the the People’s Republic of China. So the there is limiting on what what the capabilities are within within those models too.

The the thing that I see here is they showed somebody else another way to do it. To to borrow a sports analogy, like the NFL is a copycat league, somebody started doing a wildcat offense, and then everybody had a variation of it. I see this as getting duplicated pretty quickly by a couple other companies that might not have those privacy issues, and it’s still gonna be a disruptor, based on the way that it’s conducted.

And that that’s gonna be the big, like, the big play. And that’s gonna be the disruptor. And to see what it comes out as, it’s it’s exciting again. It’s another platform that people are gonna have to be able to build off of.

This isn’t unique to technology. Somebody comes out with something really good. Somebody comes up with a different cheaper variation that has their own pluses and minuses. Right?

Josh, am I seeing that?

Are you seeing I think so.

Yeah. I’m I mean, I I wake up just excited to see what’s out there next. This is the funnest time ever because it just seems like the innovations are happening so fast. And the things that you can’t even fathom you could do as a business, you’re probably gonna be able to do tomorrow or in another week or in another week. And I think that’s really our job here is to help our customers proactively be guided back to that earlier conversation is what is your AI strategy. Right? This is why we’ve come out with the the partner downloadable AI readiness checklist to really help people walk through.

Have you thought about this? Do you know, you know, what direction are you going in? And so while we’re here trying to be really good at the basics, the foundation, the principles, the process, we’re also equally, you know, waking up early, pay attention, reading the TLDRs, reading this x article, reading whatever the article is, and understanding what that product is. And and and really trying to help people think about how how do I go to market with this? How do I talk to my customers about this? How does this revolution or improve revolutionize or really improve what it is that that they are trying to do as a business and help them differentiate or or become more efficient or whatever it is? I think that’s really the exciting part that we get to play in this every day.

Yeah. I think it’s, it’s a fantastic seat to be in and then to be able to help try to give some better, insights to our advisers.

I think the opportunity has continues to grow. I think anytime you can come on and go, oh, this is the best time to ever do our jobs, which I think we’ve said now for ten plus years. Right? Because we get more and more access to to more and more cool stuff. Like, think of what the use cases Jason laid out. Three week close on redefining a company’s, like, go to market strategy.

These things are, like, really impactful and very exciting to be a part of. So, Doug, I think we we try to hit some of the, questions in the chat, but then we might have missed some.

So It was a really good review of the, questions that were in there and phenomenal information overall.

I’ve I’ve tried to sum up a few of the others into categories, but we’ve talked a little bit about the regulation and compliance issues that the manufacturers of these products have to deal with. What should end users be concerned about about their own legal and compliance responsibilities as they use these tools?

The end users more on the cybersecurity stuff.

Because the end users need to know, hey. You whatever you put into that model is gonna somebody else could be referencing it within a question. And that was one of the biggest concerns when they were first implemented and why cybersecurity and IT teams need to get in front of it. Because users were were putting in financial information, personal information. They were they were putting in stuff in order to get a a valid response from the generative AI that they could send to somebody, and be more efficient. But then, unknowingly, they were actually training that model with proprietary information. So the end users themselves need to know their impact to the business by using these tools.

Shadow Shadow AI is a is a real thing. Right? You tell an organization you can’t use something, Doug, they’re going outside of it. They’re not it’s just like when you change when you went from a prem based, organization, all your all your gears on prem to you’re moving to the cloud.

It changes your the way that you procure IT services. And so and this will be a big change in the way that you deploy and manage a lot of this technology as well. And so as you go in with a run book or an idea of how to do this, these readiness guides or these checklists and all of these things, and I think that, we’ll answer where to find those here in just a moment, for everybody, because I know there’s been a couple of questions on those. It’s really giving them the the road map to protecting themselves.

And so, like, you go from, like, the old school way of, like, procuring IT hardware. You had to go put it in a request, somebody had to approve it, you had to buy it, you had to burn it in, and go from there. And when the cloud came along, now it’s keystrokes. Right?

And think of through what, like, AI now enables, where you used to have to go through and do the research, Google, pull down stuff, maybe use some other apps to help rewrite some things. You can just toss in a prompt, and off you go. And it’s a it’s a quick and easy button, but it also exposes a lot of risk as Jason just laid out. So it’s giving your customers, like, that framework of guidance of here’s the right way to do it, here’s the wrong way, and here are some trapdoors to avoid that have making a lot of impact in these opportunities from the adviser side.

Koby, you took this exactly where I wanted to go because Joanne Jensen asked a question earlier. We’d been talking a little bit about data cleanliness, data hygiene, the effect of hallucinations, and so forth. And she asked a great question. What about understood processes? And I think where she was going with this was, you wanna use AI as a tool to facilitate one thing or another, but maybe you don’t fully understand the process yourself in your own organization.

You feed in the information you think you have. You train your model. You train your product on what you think you want it to do, and then you learn later on that maybe you missed something because you didn’t understand your own process very well.

What is the impact of that on the future operations of the company, and what can organizations do to make sure that before they engage AI, they have fully vetted their processes and the information they need to include.

Yeah.

I’ll I’ll take a stab at it, and then everybody else feel free to jump in.

I would say it’s setting down starting with the data. Right? So here’s a here’s something that was eye opening to me when I started to to look into this. Most organizations have multiple streams of data that aren’t connected in their back ends.

Like, they all have accounting coming in, dropping into this bucket, sales and marketing, etcetera, etcetera, etcetera. But they’re not necessarily dropping into a data lake or a data warehouse where they’re sharing and it it’s, where they can get the goodness from all the data together. So cleaning up your data, we’ve hit on that so much. That’s gonna be step one.

Then figuring out what’s what’s the use case and where people are wanting to utilize it for, and then walking it backwards from there. Here’s my end result. How do I build it backwards? And how do I get from from a to z?

As far especially with the security mindset. You wanna have the security people in the room. A lot of companies have moved when they like, organizations have built out AI committees internally that are exploring this. Right?

And they’ll have their legal team set in on it. So for regulatory and compliance and just making sure they’ll they’re not investing in a technology that’s not gonna be available potentially because of a regulatory, change or something like that. And then walking them through all of the security. So you have your security team, your legal team, and your your infrastructure team usually involved in these committees.

Those are the people that you wanna have in the room as you’re exploring these conversations as much as you possibly can.

It’s it’s risk too at the end of the day. There’s a lot of power in these tools, if built and implemented correctly. A lot of people wanna move fast. A lot of people see the exponential power.

We get excited about the exponential power. But, again, I I know we’re we probably sound like a broken record here, but it goes back to the data because the this is a risk just like anything else is. I I don’t care what model you’re using. I don’t care what you wanna use it for.

I don’t care what your goals are, but it’s a risk to your question, Doug, that we need to make sure that the IT leaders of these businesses are understanding what what people are using and where it’s going and really what the sprawl with the shadow AI. You know what I mean? That’s a that is probably the most drilled down way I think we can quantify it.

It seems like people are getting very excited about AI overall.

We referenced the Super Bowl commercials at the beginning. Heard a statistic earlier today that there were only two car commercials during the entire Super Bowl, but there were many ads that were either featuring or created by, to some extent, AI. Overall, this seems to be much more palatable now to businesses and consumers alike. Does that make it easier for our our our advisors to have these conversations with their clients?

I think so.

It it it seems to me that once once mainstream keeps talking about this, people can’t ignore it. They can’t sit under a rock and not do anything. I think they they then get asked from their leadership. And, Sam, I’m sure you probably see some of this. What are we doing with this?

What’s our play?

Yeah. Yeah. Exactly. I would say to answer your question, Doug, they’re it’s making it easier, but it’s also making it extremely hard at the same time.

It’s easier because, people are seeing it mainstream. Right? But it’s hard because everyone is muddying the waters around the capabilities around AI, and people think, oh, I could just go to a store and buy a box that says AI on it. Right?

And that is absolutely not the case. It’s gonna be different every single use case, and you have to look at the outcomes and how the capability is actually catered to those outcomes. Right? It’s no longer a capability game.

It’s more now an outcome game, and then you have to untie the spaghetti of, okay, data and security and all that intertwined.

So true.

I need another half an hour for this call. Chandler, we need another half an hour for this call. Thanks, guys. This has just been phenomenal information.

Still some more questions in the chat. We’ll try to get some of those answers back to you.

Josh, you wanna try to sum it up for us today? We’ve gotta finish up.

Yeah. Look. As you can see, kind of the some of the points, there’s the things that we’ve driven home is there’s exponential power out there with AI. We’re seeing anecdotally from all the customer conversations that people need help in this and guidance in this as much as as they have in anything ever before, and don’t take any assumptions that they have it all figured out.

Start the conversation with what’s your AI readiness plan, what’s your road map, where are you with this, and do you have any version of where you think AI can help you and improve upon? And I think what you see here is that this doesn’t start in one silo and stop there. It goes into all these different areas. It crosses over from cloud to CX to security and everything in between.

So, have that conversation, start it, and you’re not alone. Right? We’re here to help you walk through that. We’re here to help whether it it goes deep into security, deep into cloud, and we wanna get nerdy into different LLMs, we’re we we’ve got the resources here as as you’ve seen from everybody on this call to help.

Thanks, Josh. Thanks to all of our experts today for our fantastic presentation. It is such an amazing subject. Appreciate each one of you. If you like, more information like this, be sure to check out the Next Level BizTech podcast, which Josh Lupresto hosts.

You wanna make sure that you’re subscribing to and listening to this.

It is absolutely phenomenal.