HITT- Understanding agentic AI and its impact- May 6, 2025

In a recent session, the evolution of AI was explored, highlighting the emergence of agentic intelligence as a significant advancement. Sam Nelson led the discussion, emphasizing how agentic AI can autonomously reason, make decisions, and enhance customer service by resolving issues without human intervention. The session also covered the three waves of AI, with the third wave representing a shift towards more autonomous and proactive AI capabilities. As businesses begin to trust and integrate agentic AI, it promises to streamline operations and improve customer experiences. However, challenges remain, particularly for small to medium-sized businesses in accessing these advanced AI resources.

Transcript is auto-generated.

Today’s high intensity technical training is time to go. It is, as always, your comments are welcome in the chat window for live q and a both during and afterward.

AI continues to evolve. We have talked extensively about this over the last weeks. We’ve seen AI answer questions. We’ve seen content generation.

But now we’re seeing a third wave of AI, agentic intelligence, a new class of AI that can plan, reason, and execute tasks autonomously, all to drive real business outcomes. Today, Telarus VP of CX, Sam Nelson, presents today’s HIT training. She is joined by Mikey b. She is joined by Megan Tai. Sam, you always bring something new to the table. This discussion impresses me as being an enormous leap in the AI evolution.

And as always, our advisors need to understand how to help their clients understand it and use this resource. Welcome. How are you doing?

I am good. How are you doing, Doug? How are those flaming hot Cheetos?

You know, I I I I miss the flaming hot Cheetos. I had to give them up, like I say, because of the the dust. But, you know, it’s raining outside, so I could just, you know, go outside and wash them off. So if I disappear for an extended period of time, you’ll know what happened.

Excellent. I love it. Well, good stuff. Well, thank you, Doug, for the introduction.

And as everyone knows, Megan Tighe, Mikey Bee, our CX and AI solution expert solution architect slash experts are on the call with us today. You guys know my style. Right? It’s it’s very casual. It’s very dynamic.

Go ahead and pop things into the chat. Say hi to us. I do have the chat window open.

So if you wanna pop anything in there that we can go ahead and call out or answer on the fly, do let us know. I am going to apologize in advance that you sound a little stuffy, going through, like, allergy season. I think everybody is, battling I don’t know if it’s cold. I don’t know if it’s allergies, but you know what?

It’s something, So it’s all good. Hey, Kenneth. Good to see you here. Alright. So today, it’s all about AgenTek AI.

It’s unleashing this third wave of AI. And this session is a little unique. Okay? And I say it’s unique because we’re talking not just CX or customer experience technology, but we’re actually blending quite a bit into sort of the cloud swim lane.

And, I’m gonna go ahead and kinda clarify all of that for you because I think people think, oh, yeah. All this CX stuff is really cool, but they don’t really think about sort of the data, the infrastructure, all those things sort of behind, artificial intelligence.

The reality here, folks, is that AI is only as good as the data, right, that it’s got to work with. And if that data is not clean, it’s not accurate, or it’s not being handled appropriately, or if the infrastructure isn’t, there, you know, inside in a company to support the capabilities of artificial intelligence, technology, then it it’s it’s not gonna be any good. Right?

Okay. So in true Sam fashion, I wanna get to know you. Who’s on this call? I know we’ve got a lot of long time listeners, maybe first time callers, maybe hundredth time callers.

But I want to know how familiar are you with agentic AI, and it’s okay. We’re gonna go over it, but I wanna know. K?

Negative nine. Okay. That’s a new one, Mikey b.

Lies. What does agentic mean? Perfect. Excellent. This is great. Five. Okay. Alright. Y’all now y’all are just messing with my number system.

I see how it is.

And it’s okay to be completely unfamiliar. I you know, you could put zero. You could put zero. It’s totally cool.

But the goal is to make sure that this session is really tailored to what, you know, you guys want negative zero. That’s a what’s AI. I love it. Hey.

If you’re be you’re being honest. Okay? Good. Now, you’ve all passed the test, meaning that you’ve all participated.

Congratulations. That’s a win for Tuesday.

Pat yourself on the back there. I super appreciate you all participating.

Infinity and beyond. I love that one. So let’s dive right into the content. I’m gonna tell you exactly what’s coming up, what to expect, and then we will roll from there. But like I said, keep the chat rolling, folks. And the way I do this too is it’s kind of like a podcast where you participate.

I’ll just answer stuff. Megan, Mikey, we are all over the chat. They’re gonna stir up some stuff as we go along. Alright? So today’s agenda that’s right. Mikey b, stirring the cauldron. Even though it’s not Halloween, he’s always doing that.

Okay. So the agenda today, AgenTic AI. We’re gonna talk about why it matters, what in the world is Agenstic AI.

We’re gonna cover a little bit about what the AI brain sort of is, giving you a little bit of insight into some terms that you may or may not have heard, kind of the components of what makes up the AI brain.

We’re gonna talk about some personas to target with this stuff. Right? Because the goal here is not just to, give you insight as far as what AgenTic AI is, but also how to go out and actually sell it and have the conversations with your clients or even prospects.

Everybody’s talking about it, but it’s also kind of that thing that, you know, people wanna talk about but not too much because they might get into the weeds a little bit. And that’s where we can help, of course. We’ll go over some use cases so that we could tie everything together. It’s gonna make sense for you. And then last but certainly not least, talk about next steps. We’re gonna have a lot of time for q and a.

So like I said though, ask the questions. K? We we love we love to see it, and, we love to go off topic. Marketing loves us, for going off the menu all the time. So that said, story time. You all know that I love starting with story time.

The title of this session, is quite literally unleashing this third wave. Yes. So we’re gonna actually talk about, like, what does this mean? What do you mean by the third wave?

I think there was a book called The Third Wave. Not relevant here. It was a good book, but, we’re gonna go over sort of the first, the second, the third so you kinda understand where we’re coming from. Okay?

So we’re gonna do some lightboard action here. I know I’m gonna get the questions, so I’m just gonna say it now. Revolution lightboards. K?

You can Google it. It’s fantastic.

Alright. So, wave number one. K. We’re gonna start there.

K. What in the world was it? This is when you had a lot of sort of early AI systems. They followed really, really strict rules, and so we like to call this rules based AI. K? That’s really what happened in wave one.

And, this is where AI systems could really only respond to, like, very specific inputs, with, like, very predefined outputs. So, like, let’s say, for example turn this down a little bit. It’s giving me a headache.

Let’s say for example, when you call in and you hit one and it goes to, sales or you hit two and it goes to support. It’s very, very basic.

Think of, like, the very, very basic automated phone menus. Okay? That’s what wave one was. It was very useful, right, because it didn’t require someone to actually help you, through a menu and route you to the right person.

But it was really kind of rigid. Right? And it wasn’t able to adapt to new situations. Because let’s say, for example, you called in and you didn’t wanna talk to sales.

You didn’t wanna talk to support. You had to hit zero and then someone had to, you know, come to you and say, hey. What is it that you need? And then they had to route you anyways.

Right? So not as effective. Okay? Alright. Now we’re gonna move on to wave two. Alright.

So wave two. This is where, like, your statistical kind of machine learning based AI came in. Right? So this is ML based.

Alright. So this stuff. Alright.

Data and sort of the computing power grew quite a bit. Right? And, this is where we started to see artificial intelligence kind of advance into more of these statistical models. Okay?

Machine learning became a thing, ML. Right? And then, what these systems could then do was actually analyze significant amounts of data. Right?

Because we came to a point where people you know, we moved to the cloud. Yay. But now everyone’s got all this data. What do we do with it?

And so the systems could then analyze the data, and it could not only do that, but it could recognize patterns, and then it could actually start doing things like making predictions.

Right? And and this is actually where AI kind of took off to start sort of predicting behavior.

It started recommending content.

It started having really, really basic conversations. Think of, when you call in and the robot says something like, what can I help you with today?

Right? And you could tell it, and it would go ahead and figure it out and route you to the right person. No menu needed. No buttons need to be pushed. Just tell it what you want to do.

Another basic example of this is, say, something super relatable. If you’re looking on Hulu or you’re looking on Netflix, a very kind of basic example is it making recommendations to you based on shows that you liked. Right? This is where it start to take that data and then, you know, find and figure out kind of what you may want to watch next. Not only did it do that, but it also took your profile, looked for similar profiles across other users, and then made the same recommendation.

Right? So started to kind of learn and take all this data and make a lot of sense with it. Now let’s talk about sort of this third wave and what this means. Okay? So wave number three. This is what we are all about today.

Wave number three. This is where agentic comes in, agentic AI, or something that we like to refer to as the reasoning phase. Alright? So reasoning.

This is a big, big deal, folks. We are already here, and we’ve come such a long way, in such a short amount of time, which is crazy. So we’re now in this agentic era. I’ll tell you right now.

What you’re gonna see is this word, agentic, used in everything. Just like how over the last couple of years, we saw AI become that buzzword. K? The new buzzword is agentic.

Mark my words now. This call is being recorded. Agentic is the new word. It’s the word of twenty twenty five folks.

It has arrived. So this is where the wave, essentially, AI understands and it also predicts.

But in addition to wave one and wave two, it starts making decisions. Okay? And this is where enter autonomy.

K? So think, based on the conversation it had with you, it’s gonna start making its own decisions, to assist, but also to help you achieve the outcomes that that you have.

So let’s say, for example, instead of just like a product upgrade. Right? Something like AI can actually analyze the customer’s data. It personalizes the offer.

It actually initiates outreach to you, and it even schedules, like, a follow-up session with you. Right? It’s starting to make decisions based on, excuse me, based on the data it it has actually collected. Right?

So the big word here is autonomy, autonomous agents, agentic, reasoning, all that good stuff. So in true Sam fashion, I do have to erase this. So start lighting up the chat. Mikey b has put a lot of stuff in there, folks.

Take a quick look, ask some questions.

I’m hoping that nobody here has a PhD in AI.

Definitely one of my worst nightmares as you’ve wondered this recently. Someone was correcting me the whole time, but it’s all good.

Let’s see. How does it deal with bias and ethics?

Oh, I love this one.

Bias and ethics. How do you police that? Alright.

So first thing I’m gonna say is that AI is something that’s constantly learning, and you do have to teach it.

It’s like this machine that’s just constantly taking do you do do you do Windows?

Wes, I’ll get to that after this question. So you gotta train it. That’s correct. So Zachary is right.

You have to train the model, folks. And this is typically what we call a nesting period. So it’s anywhere between, you know, sixty to ninety days where you park, whatever AI tool it is. Right?

And you’re constantly making adjustments. You are teaching it how to do its job.

Yeah. Windows as a service, no spiff. We can go deeper into that question, John. It is gonna, differ based on supplier and functionality and capability and all that good stuff.

But there’s absolutely a learning period. And here’s the thing, as technology advisers, you can be really sticky in these opportunities naturally because you have to paint the picture that the customer has to, in fact, continue to train the model, revisit the solution, make sure it’s continuing to work for them to help them reach their goals. Right? So it’s a super sticky product, but it’s something that, helps you stick around as well.

Alright?

So let’s keep alright. So why does it matter? Why do we care about this stuff? It’s cool.

Right? And the waves are pretty neat, and the stuff I wrote on the board is pretty cool. But let’s talk about why it really matters for your clients and for prospective clients. So first and foremost, the wow factor is huge.

It is, in fact, a new competitive edge.

In fact, k, but Gartner threw out this stat, eighty percent of customer service issues are predicted to be resolved autonomously by twenty twenty nine. That’s eighty percent, which could potentially lead to a thirty percent cost reduction in operational costs. Okay? And not necessarily the people. Okay?

Rather, it’s just making them more effective, potentially, you know, scheduling them to be, in more effective sort of time time slots here and there. But also, if you have agentic AI working the system here, you’ve got things like twenty four seven support, right, where, AI AI doesn’t need to sleep. Right? And so, this is a new sort of competitive edge for customers.

And that said, let’s just dive into, like, what exactly is it. If you had to, give it a particular definition. Alright? Yeah.

It’s it’s a it’s autonomy. It’s it’s all good. But at the end of the day, we call it this sort of emerging class of AI systems, and it interacts with users and systems with autonomy, deep reasoning, and this ability to act independently toward achieving complex goals. So in three things, it sets goals, it learns, and then it adapts.

Alright? Those are the three things when it comes to agentic AI. And that’s what makes it so unique compared to the AI that we’ve been seeing, emerge over the last couple of years. K?

So next up, let’s talk about the AI brain. Oh, this is fun. Okay. So predictive.

Alright. That’s the first one. So it plans ahead. K. Let’s dive into this a little bit.

So predictive AI, is essentially, it it analyzes the data to forecast what is likely going to happen. Right? It’s this piece of the brain, the AI brain, not my brain. I wish I could, but I have a crystal ball and those are all guesses.

But the predictive AI component of the brain is is one that plans ahead. Right? And its impact is to actually help clients proactively. Right? And it meets the needs instead of reacting late. So it’s proactive versus reactive. Okay?

So let let’s, give an example here. Let’s say a provider, right, uses something like predictive AI to flag accounts that are most likely to churn.

Right? And what agentic AI would do in this situation, is it could actually notify the account manager.

It could schedule a check-in between the account manager and the customer, and it could actually proactively offer personalized discounts or maybe even, like, a training session or a check-in session with that customer. So what you’ve done, essentially, as a business owner is you’ve automatically turned a risk into a retention. See what it did there? So that’s the power of something like predictive AI and that brain and how agentic, is related to that. Now the next one that we are all very in tune with these days, because everybody does it now, is conversational. Right? So the listening, the speaking, that’s a part of the AI brain.

It’s understanding. It’s responding to actual humans.

And so the impact here is that, ultimately, this part of the brain reduces the friction. It increases accessibility.

It’s gonna speed up a lot of the decision making.

So how does a Genesys AI kind of work in conversational? Let’s talk about that a little bit.

So think of, agentic AI being embedded in, like, a customer portal. Okay? And when it’s chatting with users about, you know, product issues, what it can do on that con in that conversation, is that it can identify the root cause of the issue. It can automatically escalate a ticket or create a ticket and escalate it if needed.

Or it could even solve the issue on the spot saying, hey. Here are some links to help you. Here’s some documentation.

Or guess what? If it even, like, identifies what the issue is, again, power of cloud, power of data, on the back end, it can actually go in and, well, fix the issue itself. Right? Doesn’t require a human to do it.

It can go in the back end and fix it. So what this means for a business? Well, it’s faster resolution, happier customers. Right?

People are slowly moving to that model of, like, trusting the AI and training it to do these kinds of things so that they can dedicate actual human resources to more kind of complex initiatives or, you know, just other ones.

The next one is generative. Okay. We’re all familiar with this.

Put a one in the chat. Put a number in the chat. If you have used something like chat g p t or Gemini or something.

Yeah. That’s what I thought. Good. We have a lot of honest people here. I love it.

Yep. You have used some sort of Gen AI, within the last probably twenty four hours. Right? You pop something into ChatGPT or whatever.

Doesn’t have to be work related. It could be like, hey. Like, give me a workout for thirty minutes while I’m traveling, and it’s something that I can do just in my hotel room. Right? Not like I’ve done that at all.

Okay. So generative AI. Literally, the brain that creates a piece of the brain that creates.

It does things like create new content, whether that’s text, for programmers. It’s code for, designers. It’s graphics.

It can even do things like workflows. It could create, gosh, so many things, mind maps, like, you name it. Right? And the number one reason why that’s so effective is because it’s gonna save you lots of hours of manual work, and it kind of enables this rapid iteration.

And so we’ve all experienced it. When we need quick resource, we just pop that thing into chat GPT and out it comes. Or we want to make an email sound different. Right?

Change the tone of this email or make it sound more professional. Pop it into chat GPT. Of course, leave out the sensitive information. Okay?

Don’t forget that.

But they can do things like that really, really quickly for you. So let me give you an example of this. If, like, a marketing team wants a new product landing page, the agent, the agentic AI, can not only write the copy, generate the visuals, but it can also optimize things like, search engine optimization, SEO. Right?

It can optimize that metadata and then maybe even push it live, via kind of marketing system integrations. Right? And so as a result, something like a marketing team, can actually, like, reduce launch times. Right?

Because they don’t have to do that much more. They could actually reduce that from weeks to hours potentially.

And it gives them more time to really focus on sort of refining and enhancing materials rather than the launch itself. Right? So just a couple more use cases for you. But that is sort of the three components of what we reference as the AI brain.

Too funny. I’m seeing some stuff in the chat. This is great.

Alright.

Let’s talk about personas. Who cares the most? Now I’m not saying that these are the only personas that you should target. Okay? But what I am saying is that these are some of the most common, that we are encountering in a lot of these conversations we’re having around things like agentic AI. Alright?

So the first one up into the right is what we all like to see. Anyone in the revenue organization.

Anything in the revenue organization because and and a lot of you know this as well. Right? We’ve got a lot of sales folks here. As tech advisers, we’re constantly selling.

But if you’re not spending time selling and you’re doing something else, you’re essentially taking dollars away from your day. Right? You wanna be focused on those revenue generating activities.

And so anything that’s gonna get you more of that time or make that easier for you, is a big bonus. So, of course, anyone in the sales org is really, really, really into Agenstic AI, as well as marketing. Right? Sales and marketing come together.

But also, guess what? They got the money to spend. Okay? They have the money to spend.

Now the beauty of this is you apply technology and it’s auditable, it’s trackable.

Finance is gonna love that because typically, finance wonders what ROI is on sales and marketing activities. Well, guess what? You pop an AI agent in there that’s tracking everything. It’s tracking ROI.

It’s tracking lead conversions. It’s try you name it. Right? It’s gonna provide that full analysis, full visibility, to make further investments in things like technology that we all want.

K? The next one, CX leaders.

If you guys have heard of c x p a dot org, check it out. There’s a whole community of CX leaders.

If I were you, I’d go check it out, start stocking it, go on LinkedIn, find people who are members of CXPA or who have that certification.

Companies are hiring these folks to come in and optimize workflows, processes, technology stack, you name it. But CX leaders are looking for ways to embed agentic AI into the technology stack all the time, because that’s gonna like it says here, automate insights to action.

Last but certainly not least, and I say this, like, in no particular order. Okay? So, like, don’t come mad at me, cloud folks. But cloud architects.

Alright. So this is really neat. And this is not my swim lane. This is definitely a Kobe Phillips swim lane.

Right?

But you have to understand that this is insanely important.

The cloud conversation, the data, the architecture, the structure, all of that good stuff, is really, really important. And the way that cloud architects are actually using things like Agenstic AI today is they’re, you know, they’re enabling sort of this no touch provisioning, scaling, and monitoring. So in other words, these things, these activities no longer really require a human, in a reactive situation. Rather, if something happens literally within, like, seconds or milliseconds, something like agentic AI can come in, provision, scale, monitor, boom.

It’s done. Right? And we’re all, like, all lights are green again. No need to worry.

All the car emergency lights are off. We are good to roll. Right? So these are just some personas who are gonna care the most.

And, again, no particular order, but it and what it does, what AgenTic AI does in it is that it really enables you to have more conversations beyond just customer experience or beyond just cloud.

At the end of the day, something like AgenTic AI is kind of bridging the gap, between a lot of these different swim lanes, meaning CX, cloud, cybersecurity, mobility IoT.

Right? And so just know there are so many different applications, but it’s now this sort of tether between all of the different technologies, to have bigger conversations. And we are here to help you with that, folks. I know this is a lot of data coming your way.

The last thing I want is for you guys to sit in a dark room, you know, on the floor in the corner, hugging your knees to your chest, rocking back and forth. We’re not looking for that. Okay? Bring us in.

We’re happy to help you.

Alright. Let’s roll through a few different use cases because I wanna leave a lot of time for questions. I wanna bring on Mikey b, Megan. We can chat through some of these things that are coming up.

Doug can come up as well. But use case number one. Okay. So we kinda chatted through this, sales enablement.

So let’s say, for example, a, the simplest form. A lead interacts with a website. Okay? And without any human intervention whatsoever, something like an agentic AI system, can create a custom email sequence based on what that lead did. So say the lead went to, check out gray shirts, then they went to blue shirts, then they went to blue pants. So what the custom email sequence might be is, hey. Noticed you checked out all of these different items in these colors.

Let’s schedule a meeting for you to meet with, a personal shopper.

And by the way, here are some looks that go great with all the selections you made.

Right? Like, how powerful is that? And it and it happened all without an actual human. Right? So, just a lot more hand holding, white glove service. It’s like going into a Nordstrom, but, like, virtually.

Right? Very cool. Use case number two. We talked about proactive versus reactive support. So, this is really neat.

This is more of like a cloud thing, but very applicable. So let’s say a client, is working with a a system, they hit a bug. Right? And it’s a show stopper, and they need immediate support.

So what something like Agentic AI can do is it’s not only going to detect the issue ahead of time or or as it happens essentially, but it can go ahead and generate an actual fix. Right? And then it can tell the client without them having to, like, submit a support ticket or anything, hey. Like, here’s the status update.

Here’s the resolution. Notice this happened to you. So sorry about that. Right?

The most relatable I would say to something like this is say you’re on a an airline. Not gonna name any specific one. Let’s say you’re on an airline and you have terrible WiFi. Right?

And you get off the plane and you get that email that says, I’m so sorry you experienced, terrible Wi Fi on the plane. Here’s a credit for your next flight. Something like that. Right?

This just takes it to the next level from a technical standpoint. Alright?

The third use case. Alright. So, again, we’re dipping into cloud. But, guys, like, don’t forget, this is a really, really big deal. Okay? CX is one thing, but if the data is no good or if the infrastructure is poor, it can’t support all these other cool tools.

So let’s dive into sort of the cloud resource provisioning swim lane just just for a second. Okay? And, again, I’m not a cloud expert, but this is a really applicable use case. So let’s say, this customer.

Right? They they create they request a sandbox because they wanna test a bunch of integrations and stuff. The agentic AI component is is not gonna wait. In fact, it’s gonna use that predictive piece of the brain, and based on everything that it that the customer has created in the past, it’s gonna go ahead and generate this, what we call, infrastructure as a code, and it’s gonna deploy that environment securely automatically.

Now in the past, if you’re familiar with the process before, it it was a human. Like, the human had to do, like, a call with the customer.

They had to talk about the history and review the history of all the environment the customer previously created.

And then someone had to go in and modify, generate infrastructure as a code. Right? And then someone actually had to deploy it, and then the customer could go ahead and test integrations in a sandbox. Right? But now we can take all of that, automate it with something like agentic AI, to to take that over. Really, really neat stuff.

Okay. So what do you do? What do you ask? I even, like, start this conversation.

You could literally start it by something like, hey. I was just on the phone, reviewing some, information about agentic AI, and I think we need to have a bigger conversation about it. When are you free? Right?

That is, like, the most basic way to get someone on the phone or in a meeting to talk about something like well, anything, but agentic AI in particular. That’s what we’re talking about today. But some discovery questions for you are things like, you know, what repetitive tasks are slowing down the teams today? Do they guarantee there are many?

Okay. Even we experienced them here.

How quickly can the systems turn insights into action?

Or how are you turning insights into action today? How are you leveraging the data to make actual data driven decisions?

And then a general one I always like to include here is, hey, what are your biggest goals for this year? So it could be anyway, it could be, you know, head of marketing or head of sales, or head of operations. Right? Just kind of get to know what their goals are and what their key performance indicators are. Because as soon as you understand what’s important to them and what’s contributing to their paycheck, the more that, you know, you can get sticky and really fine tune your recommendations on targeting for them to hit those KPIs using technology.

Alright?

So that said, next steps, k, engage the supplier’s team, get to know key suppliers who are doing this stuff. And, you guessed it, a lot of them are already doing this stuff. You’re gonna you’re gonna see agentic pop up everywhere. It already is.

Just make sure that you’re staying up to date on the latest trends, news.

The neuron is one of my favorites.

CX today as well, UC today as well, you name it. But, we actually have a blog. So Kobe and I wrote this blog about this content. So definitely check that out as well.

But, yeah, we are here to help you every step of the way. We’re having AI conversations with customers every day, literally every day, multiple times a day. So, definitely engage us, and we are more than happy to help. So that said, I would love, love, love to go ahead and bring on Doug, bring on Mike b, Megan. Let’s go ahead and answer some questions or stir up some interesting conversation.

Phenomenal questions and participation in the, chat today. I commented myself in there. I’ve been doing this nearly fourteen years on this call. I have never seen the chat window blow up like this today. So much interest, so many great comments.

Mike, I wanna start kinda where you finished up here at the end and give you a chance. You’ve got a few that I know you wanna respond to. But right now, everybody’s thinking, alright. My mind is blown so much information. Where do I go for great resources to follow-up on this? And I’d start with replay this call, but you’ve got a few other suggestions as well.

Yeah. So, here at Telarus, we’ve released, I don’t I don’t know, Sam, like, five or six different AI documentations, and one is specific to use cases. And then the next slide is specific to the use cases and the suppliers who fit those use cases. Because it’s one thing to talk about agentic AI.

It’s another to say, okay. I’ve talked about it. I got them on the hook. Who do I hand the pole to to reel them in the boat?

Right? So, it is becoming a little bit of a buzzword, a little bit of a market texture word. Right? Where everyone now does, does agentic.

What I always say is if some supplier is interested is is interesting to you and saying, oh, we absolutely have agentic, try them out. Right? See if they can show you a demo or or see if you can make your own voice agentic.

So there are some suppliers out there who would do a really good job.

Just reach out to Megan and myself, and we can kinda narrow that down.

But, Doug, it’s all about, education and tools and bringing in the right folks.

I wish I could tell you that Megan and I are one thousand percent up to date, but as soon as we somebody said in the chat, as soon as I got it figured out, they introduced new things. So we really try to be on the on that tip of the spear for y’all.

But, no, we’re learning every day as well.

Well and, Megan, you really brought this out at the beginning when you talked about the, development of the evolution of the various types of AI that has come around. And I think for a lot of us who look at this, as something we’re not very familiar with, but we’re learning ourselves, we finally reached that point, and we all knew it was coming where now AgenTek AI has gone beyond just what we teach it and tell it. It’s setting goals. It’s learning. It’s adapting.

It’s reasoning.

And how do we continue to properly train and manage it given its new capabilities?

Was that for me or Sam? Sorry.

I was throwing it out to Sam. I’m sorry. Did I say Megan? I’m sorry. Either one of you.

Well, I think, Megan, go ahead.

Just needs to help me with this.

No. No.

I was just thinking, yeah, training you’re right because you always have to check.

You can’t just assume you’re launching something and it’s good to go. It it you know, Sam had mentioned you have to continually train it. I think we all know that.

Either getting feedback on the responses it’s providing or whatnot, but you just have to always train it and make sure it’s it’s good to go. And like Mikey was saying, put guardrails around it as well before you actually publish it back out. So it’s just a constant monitoring of it, I would say. I mean, even though we say agentic is very autonomous, it has to get to that point.

It doesn’t start off like that right away. Right? So we do I feel we we still have to continue to monitor and make sure we were keeping it in check, and then it can run on its own. But it’s it’s not something you can just turn on and maybe do a check once and then assume it’s gonna be good the rest of the time.

For those that have a a little bit of a concern about the guardrails portion, how how safe are the guardrails?

Does does AgenTic AI and its continuing evolution have the power, authority, or just the ability as it continues to learn and adapt to say, I don’t like those guardrails, and it would be better if I did something else. What restrictions are in place to ensure that those guardrails remain firm?

You sound like, Marvel co comics. Who’s watching the watchers? Right?

So I’m just planning my weekend.

Yeah. Yeah.

Anyways, I mean, this is a nerd show, so we might as well throw some comics in here.

But on a serious note, I think each supplier needs needs to, and I’m sure they already have, their own security document.

As we do as we walk into the AI world, I’m seeing conversations actually starting with infosec. So once once once the CX leadership team says, yes, we wanna move towards an AI model, they are we bring in the infosec team right away to get all all that conversation off the table, then we can focus on the solution and the better customer experience and employee and agent experience after that. So, it is important to get the supplier on with the InfoSec team straight away. I did mention one supplier, Luminal.

Madison, you owe me, you owe me a cup of coffee.

But what they’re doing is is pretty cool. It’s it’s not so much a a CX tool per se, but it’s for kinda like Sam said, how are you using ChatGPT?

Well, what if I use ChatGPT?

Hey, create a a voice that sounds like Doug Miller who lives at one Main Street, Sandy, Utah, x y z. And all of a sudden, that goes to ChattGPT, and then your address is written up there, Doug, for everybody to use. That’s probably a bad thing. Right? Or if you’re a law firm, hey.

You know, Megan Tai was was cited for, rioting in the streets of downtown, Spokane.

And now you how do I defend her? Now that gets written up to chat g p t, probably a bad thing. So what we like about Liminal is it’s not too much a CX product, but it really puts those guardrails up automatically for when I use chat g p t and I use real people’s names. And I would love to see the video of Megan rioting in the streets of Spokane.

So, somebody can I’ve got her real drowned if you’re not gonna be able to find it.

I’m pretty sure it exists somewhere. We just need to ask chat QPP.

Yeah. Exactly.

Yeah. And and just to kind of dovetail off of what Mike b was saying is that, there’s a big kind of security concern around this, obviously.

People willy nilly using the free version of Chat GPT, and they don’t realize that all that data is just out there for everybody to use. It’s almost like a big Wikipedia. Right? Do we even use that anymore?

But, yeah, it’s like a big Wikipedia.

And so you’ve got several suppliers now kind of jumping on this because there’s this kind of, path secure path to LLM, right, that a lot of companies are looking into because they know that that people are using it. I mean, I took a poll. Everybody’s using it. Right?

And so how are you making sure it’s secure? Right? So you now have, suppliers like Liminal. You have suppliers like Expediant.

You have so many out there now who are kind of finding ways to allow employees to use LLMs or leverage LLMs securely.

So that’s definitely a big theme that we’re now starting to see. And then you start building agentic AI in there, where it starts, you know, making autonomous decisions, it takes it to a whole new level. But that’s kind of like that that first entrant on the security standpoint.

Sam, you’re VP of CX for Telarus, and and we’ve spoken with you many times over the years about the evolution of the customer experience. We see this branch into other disciplines. We’ve talked a lot in recent weeks about the patient experience in medical, for example.

As you look at this and you see so many different disciplines now becoming more and more embedded with AI, what do you hear from customers, from our advisors about overall satisfaction with AI in CX from those users who encounter it when they call in or when they’re otherwise interacting with AI.

Yeah. I like to say that it’s a blessing and a curse at the same time because I’ll tell you. I called into a company the other day, and it just let me only pick one or two from the keypad. It didn’t ask me what I wanted, and it didn’t route me immediately. And it was infuriating.

So, you know, the basics of AI are great, but, also, sometimes when AI asks me what I want, it completely misunderstands me, and it throws me into the completely wrong area of the business.

So I like to take the first because, and and this is actually the big reason why this opportunity is so massive for technology advisers today, is that, believe it or not, less than fifty percent of AI opportune I’m sorry, AI projects last year in twenty twenty four even made it to production.

Less than fifty percent.

Alright? And that’s across all the swim lanes, not just CX related. So what does that tell us? It tells us that, yeah, certainly people are interested.

However, I think a lot of people are still struggling to figure out exactly how it applies.

A lot of them haven’t figured out that you need to constantly train it, that it needs to be secure.

And so as we start having these AI conversations across all the swim lanes and and by the way, like, it’s not just me. It’s it’s Kobe Phillips. It’s Jason Stein. It’s Graham Scott.

Right? All four of us have kind of banded together to say, hey. We gotta help customers and technology advisors figure out this AI stuff because there’s this, again, third wave, fourth wave, fifth wave’s coming. Right?

And whatever that is, we gotta be ready.

But we are having to really arm a lot of TAs with to have to be able to have those conversations around, hey. We can’t just throw AI into the technology stack. We can’t just go buy a box of it and install it. Right? That’s not how it works.

And so there’s this emphasis now of having to incorporate what I like to call the born in AI, approach. Right? And then that’s a whole separate talk track, and I have it. Follow-up with me after.

For those of you who stayed on, congratulations. You heard it. But it says born in AI approach. It’s a completely different approach than selling your typical SaaS solution.

Right? It’s not just the discussion, the evaluation, it’s the implementation. It’s the ongoing activities after that that will make that will ensure that customers are using AI effectively.

My, case to marketing to extend this call to ninety minutes a week still hasn’t gone through, so I know we’re gonna get a little bit short on time. Too much stress, Doug. I not for me. I love this.

Megan, I wanna give you a quick chance to respond to anything else that you saw in the, chat that you wanted to answer. And, Mikey, same for you. And then Sam, at the end, I want you to finish up. We’re gonna talk about, the blog here in just a second as well, but I want you to go ahead and promote that as well. Megan, quick quick q and a that you wanted to do.

Yeah. One of the things I wanna key in on, you know, Wes was asking about the SMB market, and Mike, you mentioned this earlier. It it is challenging for us. We get so many opportunities every day.

Everyone wants AI. Why shouldn’t the SMB market be able to tap into AI resources? Right? So it’s been cost prohibitive with the integrations with, platform fees and the like.

So, just know that Sam is working on this. And Mike and I and Sam, we’re always talking about what is the latest and greatest, who can we bring in that’s gonna give us longevity, that’s gonna give us what we need, that’s gonna be cost effective for SMB players. So we are still evaluating that. But I was saying to Wes, I mean, it all varies.

So what we need to know is what what size is your customer, what vertical, how are they gonna use this, how do they wanna use this, and what is the volume as far as interactions, whether it’s, you know, a chatbot, a a conversational AI, the voice AI that Mikey and I are we’re seeing so so much demand for. So we just have to understand what the the use case is. We’re gonna do our best to try to find that that good fit, give a couple options. Because there are some vendors out there that are, flexible, you know, instead of mandating, no.

We’ve gotta have this platform fee, period, end of conversation.

We we can tap into some that are more flexible and can work with us for the SMB.

Awesome. Thank you. Mike, what you got for us?

Mitel is in bankruptcy.

So, there’s no AI in Mitel unlike Megan’s name and my name.

So that’s always fun.

Avaya, no AI in Avaya, in their name. So, Avaya we mentioned this, I think, Sam, last month that Avaya is only supporting their top fifteen hundred customers globally, the g fifteen hundred, and those are typically direct customers for the most part, big BPOs, big universities, massive, companies. So this is really an opportunity right now to, you know, add FUD in to any Avaya IP Office or Mitel customer to go after that.

Right?

Friends, Romans, Telarusians, we are not here to praise the on prem PBX. We are here to replace it. Let’s go.

Other than that my gosh. Just just when I thought he couldn’t top himself. That’s amazing.

Thanks, Mike.

Sam, terrific presentation today, as always. Love the technology. Love the content. Talk about your blog a little bit, and, where can everybody find out a little bit more information?

Yes. Absolutely. It’s claris dot com slash blog, but we will also send a follow-up to this session for those of you who are in attendance. We’ll send the link.

You can check it out, but it is, co written by myself, Coby Phillips. We went ahead and just kind of outlined everything that you sort of saw here, but in greater detail. So definitely check it out. Obviously, Mikey made it very clear that the entire industry is just right for disruption.

Okay? Whether that’s moving from on prem to cloud or even having some of these conversations around agentic AI, this third wave. Right? And by the way, if you want us to do kind of private maybe private sessions for some of your clients around this stuff, like, we’re open to that.

We love doing that kind of stuff. So let us know. Reach out to us. Engage the team.

Our solution architect team and and SEs are on it, and we will help you with these conversations.

That is a fantastic offer. The Telarus resources that are available to our advisers just continue to amaze me every day.