HITT- The transformative power of AIOps in network management- April 15 2025

In a recent HITT, experts Graeme Scott, Jason Kaufman, and Matt Douglass explored the revolutionary impact of AIOps on network operations. AIOps leverages AI, machine learning, and big data analytics to automate IT processes, enabling networks to self-heal and respond to issues in real-time. This technology addresses the complexities of modern networks, enhancing visibility and efficiency for IT teams. The panel emphasized the importance of data readiness and understanding customer pain points for successful AIOps implementation. As AI continues to evolve, it is set to augment human tasks in network management, paving the way for a more efficient future.

Transcript is auto-generated.

As always, your comments and questions are welcome in the chat window for a live q and a that happens both during and after today’s high intensity tech training. Your clients spend countless hours and a good chunk of change in keeping their networks up and running and optimized as well.

Now their networks, sometimes they go down, but what if your client’s network could fix itself?

Well, as it turns out, it can. And today, we learn how AIOps is transforming network operations with smart, self healing systems that detect, fix, and even prevent issues often before they happen.

Telarus VP of Advanced Networking and Mobility, Graeme Scott, joins us today, along with Telarus solutions architect Jason Kaufman, who was here last week. We just didn’t let him out after the call. It was an error on our part. But, Jason, glad you’re back. And we’ve invited Matt Douglass, senior director of solution engineering at CBTS, for his insight on this astounding new set of network operations capabilities.

Graeme, welcome back to the call and welcome to your panelists. AIOps is a massive game changer for network operations. How are you doing?

Doing great, Doug. As always, great to see you and, you know, happy tax day to everybody. I know it’s not a day that we normally celebrate, but, you know, it’s part of it being in this country. So hopefully everybody got it done and got their taxes submitted or will be today.

But we are ready to jump in and start talking about this. It’s another day and it’s another session on AI. Right? It’s such a big buzzword.

Everything, going on around us is AI. So, I know a lot of our folks out there on the call. If you’ve ever had to, you know, miss a family event or stay up real late to deal with a customer’s network issue, go ahead and drop a one in the chat. I know there’s a lot of you guys out there that work in that business and have had to take that late night call or got that, beeper in the middle of the day and ended up spending some time troubleshooting a network when you did not want to. Well, the good news is those days are coming to an end.

AIOps is a powerful tool that we are starting to see integrated within network operations, both within some of the suppliers that we have in the portfolio and within your customers’ network. So as Doug said, joining me today here, we’ve got Jason Kaufman. We’re gonna go ahead and break down some of the advantages, some of the things that are happening. And then we’re gonna bring in our good friend, Matt Douglas from CBTS, who’s gonna talk through what they’ve done at CBTS, how they are implementing some of those tools within their own network, and how they can help you and your customers do the same. So first off, let’s start with an overview. What is AIOps?

So AIOps is like having a great AI assistant within your network team to help you troubleshoot and perform network operations. It’s where artificial intelligence, machine learning, and big data analytics all meet together to create a fantastic tool for any and all network administrators. So what it does is it essentially automates and enhances IT operations.

One of the limitations, of course, of individuals is that we can only take in and act upon so much data at a time. Well, AI does not have this limitation.

And as such, AI ops allows your network operations to react to, to ingest, to track, to see all the things that are going on within your network at any given time and react to those things in real time. It streamlines and makes, networks way more efficient and helps, those of you who have taken those late night calls, you know, now able to avoid those in the future.

So the core capabilities of AI ops, data ingestion at scale. We know AI can do this across a number of different functions. Networks is no different, can bring in all kinds of data, all kinds of data points across the network regardless of where it is and use that information to make decisions.

Correlation, pattern recognition, seeing what’s normal, seeing what’s not normal, and acting upon those types of things. Anomaly detection. Right? When there’s something going on that is not standard, that is not normal, when Matt Douglas jumps on the network from Shanghai, China at two in the morning, we wanna know about it, and those kind of things are noted by AIOps.

The other thing it does, it really gives us a lot of predictive insights. Right? Areas where we know potentially there will be bottlenecks or service issues in the future, and we can address those ahead of time. And then, of course, as Doug referenced at the beginning, the ability for networks to self heal, identify issues, and fix it on its own.

Pretty crazy. So, Chandler, let’s go ahead to the next slide. I’m gonna go ahead and bring our man, Jason Kaufman, in here. Jason, one of the things we are seeing a lot today as we build out networks and do network, architecture, and I think this is something you talked about a lot at the, IT expo down in Florida, is that traditional tools, the tools that we have used in the past are really struggling with today’s modern networks and everything that’s going on.

Modern networks today are highly distributed. We’ve got cloud. We’ve got hybrid. We’ve got edge. They’re very, very complex.

And the tools of today are struggling, hence the need for AI ops. So talk a little bit about what you guys are seeing within the sales engineering team.

Yeah. A lot of it is just around what you’re talking about, distributed networks and many different components and overall complexity on how to manage all this stuff. You have strategic vision coming down from the top on, hey. We need to implement this, or we need to have this cloud strategy, or we need to do this to keep the business going and allow us to scale.

So IT teams are trying to keep up. And how do you manage all this effectively? I mean, we don’t most of them are gonna get the budget to add more headcount as they’re trying to scale and add all this infrastructure and complexity. So how do they do this and manage it to where the business can still have continuity no matter what they they implement within the infrastructure.

So one of the biggest complaints we get is, hey. I wanna have visibility into everything under one single pane of glass. But then also, I need help to see what’s plugged into what, what’s the dependency here, and then, ultimately, what can handle that low hanging fruit to where when that issue goes wrong, either can it can it self heal on its own and perform the remediation and I come in and I just know about it so I could tell my superiors, Or does it give me some troubleshooting steps? You know, traditionally, when you have an IT network, the you know, when they invented the OSI model, it was basically a guideline on, hey.

What are the things you wanna attack first to get through layers one through seven, given order of operations to get back up and running so you have some form of methodology, you know, talking about the physical layer. Is it plugged in? The data link, are you getting lights? Are you you know, the layer three, are you getting the IP addressing?

Is Is the network able to talk to each other? We’re going all the way up to the presentation application layer. How do you do that effectively on every single issue, and where do you start? I mean, if you have something that’s so distributed, it may be something you can’t get your hands on.

You have to walk somebody through it. Have you ever tried walking somebody through something that requires a little bit of technical competency from, you know, multiple hundred miles away? Kinda frustrating.

Yeah. I have to do that with my, parents a lot even just trying to get the Netflix thing going. And so I can only imagine how hard that would be as a network engineer trying to walk somebody through, you know, a retail establishment or something like that. So these things have a ton of advantages.

Obviously, chief among those, Jason, spotting issues faster than humans. Right? I mean, you know, we’ve got, you know, the the AI can watch this stuff twenty four seven, three hundred sixty five days a year, always on guard, always watching, always looking for anything unusual, slowdowns, weird traffic products, and there’s no waiting for a ticket. Right? A lot of times, it’ll open that ticket immediately and go ahead and address the issue without human action at all.

Yeah. Create the ticket, perform the action, document the notes, close the ticket so it’s all referenceable. But doing that on behalf of a human that, you know, is not up twenty four by seven or, you know, is gonna be on holiday, there’s a lot of things that having a machine handle this stuff on somebody’s behalf is make things way better for the environment.

Yeah. And and as we know with a lot of AI models, right, its ability to learn over time, I think, is a huge advantage for this thing. It starts to know what is typical within your, organization’s network, the patterns, the things that are happening, And it starts to look for things that are unusual and can address those. And also areas where there may be a breakdown or a future breakdown by understanding where things are. Hey. Sometimes it can even reroute traffic around problem areas and address those.

Jason, talk a little bit about natural language models because I think that’s sort of what you were touching on at the beginning when you talked about trying to troubleshoot across a geographic area. This to me is really, really one of the coolest aspects of this. As a guy who’s not a big coder or or not really overly techie, this has got a lot of appeal. So talk a little bit about that.

Yeah. Just imagine you walk into a new environment or you get a new piece of equipment and you’re not you’re not familiar with the command line interface that comes to it. You know, usually you have to type in all this different, you know, you have to type in the the commands and then you wanna tell it what what you needed to do. And it requires a lot of competency on how to communicate in a Linux platform.

A lot of them are a little bit different to where not everything runs in the Cisco language. Palo Alto has their own. Juniper has their own, but they kind of they kind of mimic a little bit so somebody could sit down and fit holistically across them. But how about if I just come in and I use human language and I say, hey, I want to segregate this network and put this many VLANs and I want to use this IP address scheme and I’m just using it by typing like I’m sending a love letter to somebody and then I submit it and now this thing builds it on my behalf to where now I’m instead of putting hours and days within building an entire network and ensuring that everything is configured correctly, now I’m doing it within, you know, a matter of minutes and having the system design that for me.

So not only from using the natural language models, the natural language processing of building something, but it’s also taking that anomaly detection and giving that, hey, here’s our recommended, steps you could take to troubleshoot to where now you don’t have to have somebody with multi years of experience managing a network. Now you can have somebody that’s pretty greenfield come in and they get everything spit out to them to where it’s easily too congested and understand.

So it’s really revolutionizing network management holistically as we’re starting to see resources get constrained and what type of resources can be used for this stuff.

Yeah. And that’s great. And so what I’m gonna do here, just go ahead and bring in Matt Douglas. So Matt, joins us, currently in the Gulf Shores of Alabama, but by way of Indianapolis.

And I think we’ve also got Katie out there working the chat. So if you guys have some questions, I think Katie’s out there answering a few of those. So welcome to you, Matt. Even though you didn’t award the victory to Sam and I in the karaoke contest in Vegas, it’s great it’s great to have you here.

So thanks, Ash. Chandler, go ahead and switch to the next slide. Matt, you guys have done a lot of the heavy lifting for the AI ops within the CBTS network.

Talk a little bit about what that process looked like, how you guys integrated some of these technologies into what you guys are doing at CBTS.

Yeah. No. It’s across the board, not only within our internal systems, but then the products that we represent and the managed services we build on top of those.

You know, really, I think all of us are starting to take a look at our customers kind of for a consultative approach of what are the applications they’re running across these networks, what are the platforms they’re running on, where those platforms are in their data center, in a, you know, IS structure, if they’re in a public cloud somewhere.

And then, you know, what are the operations? What are the customer benefits that they need? Right? And AI is affecting all of this, and and you guys kind of have mentioned quite a bit of that of AI ops ability to ingest this massive amount of data, detect anomalies, build playbooks, kind of predesigned playbooks. One of the great things about these new technologies, and we’re gonna talk a little bit on the network side, is that you can literally build quality of experience metrics.

Same thing happens on the SD WAN side now and the SaaS side with some of these newer platforms of really being able to to build a quality of experience metric because it’s it’s not so much about that my network can route a packet across particular, you know, net or or network, but how is that application running for that user whether they’re working in the office, they’re working from home, or they’re working from Starbucks. And so these AI ops abilities across network, across security, across CX, even personal productivity is just really changing how quickly we can detect issues and how quickly we can respond. We talk a lot about the mean time to detect and the mean time to respond.

Yeah. And I think that’s, I mean, a game changer for your network operations team. Right? I mean and and obviously has has made the processes so much more streamlined and efficient.

Oh, there’s no question. I mean, we’re gonna talk a little bit on on the Powell side, some of the tools, but literally the ability to, you know like in our network, you take a look at CBTS. CBTS is a couple thousand employees working from a, you know, a variety, probably ten different countries.

We have embraced the entire AI IOPS model in our SOC and our security operation center that not only manages our core but also manages for customers.

And we literally for our instance, we see some, like, five million events that happen a day that are automatically correlated down to about a hundred and fifty issues that that need to be looked at, but looked at by AI comes down to somewhere typically five to ten issues that need to be in investigated by a human, and those are solved in a couple hours where it used to be that amount. You know, you you mentioned earlier, there’s a massive amount of data and the data coming from everywhere. Like, you know, we as a user now, every user is a firewall. I’m working from I’m working from Gulf Shores here, but I’m working on a CBTS platform that AI ops is watching everything that I’m doing, the application I’m going to, etcetera.

And, these are really game changers on how quickly companies or managed service providers like ourselves can can deliver the results, the business outcomes customers are looking for, and it’s just slash the meantime to to to detect in the meantime to repair across a variety of different, you know, environments.

Yeah. That’s fantastic. Now, Chandler, if you go ahead and move us to the next slide here. Matt, there’s a there’s a process that you guys went about when you started when you made this change, when you started implementing these types of tools within the network. Maybe touch on that a little bit and what the tech advisers here on the call are gonna be sort of some of the things they wanna look for within their customers when they start to have conversations about maybe incorporating AI ops.

Yeah. Absolutely.

The the AI conversation, it I mean, it’s exploded for us. Right? We’ve all been trying to figure out in the last year how do we respond to the need. Right?

And we have kind of, put it in three categories if you would. One is talking about what are the best of breed AI embedded products out there. Right? On the security side, we’re a big power shop. We’re gonna talk about that. On the network side, we’re a big juniper shop. Those are really platforms that have really, embraced the AI ops capabilities.

And then, of course, you know, there’s all sorts of UCaaS and CCaaS providers that the new kind of AI tools around sentiment analysis and automated AI agents really change the customer experience. So what are those best of breed solutions out there that use a or AI driven AI ops to deliver better results for customers?

Then I think there’s mic the the whole Microsoft Copilot and personal productivity thing, is a huge opportunity, I believe, for for partners out there to at least explore and talk to with with their customers.

People are trying to figure out how am I going to use Copilot.

I’m gonna use Copilot to either improve my own productivity, let’s say, as a as as a sales engineer or salesperson? How am I gonna use it to improve my productivity in the finance department? How am I gonna then maybe use Copilot to build customer facing tools that are gonna increase the customer experience?

And then when we’re talking about doing those things, AI acceleration and data readiness is a is a huge issue. Right? And you’ve got you know, I think in the Telarus portfolio, you’ve got a number of providers that can help people start to take a look at their data, organize it properly, tag it properly, and get it ready for implementing some of these large language model solutions, these automatic AI agents.

And there there’s a there’s a lot of work there. One of the things that we do, and, Mike, we’re gonna talk about a little bit is we do this one day, AI readiness assessment where we come in and show the customer kinda the art of the possible. Oftentimes, it’s around Copilot. Do that the first part of the morning.

And then the next part of the afternoon, we’re then taking our what’s the low hanging fruit either internally or externally? We can help your company. What’s interesting about this when you start to open the Microsoft to this AI readiness conversation, we hold these sessions, and it ends up like three to five opportunities within that customer to help them accelerate. So getting in this conversation of AI ops, of Copilot, of AI readiness and data, bringing the right partners in, which you’ve got a number of great vendor, vendors, it really opens up a a whole plethora of new opportunities inside that customer that was hard to see before.

Hey, Matt. Quick question. So, like, when you guys are doing that out of the possible conversation, what are some of the biggest pain points that, you know, transition the conversation into an AI ops deployment? Like, what are you listening for for that type of engagement?

Sure. And I think there’s kind of a almost a a difference between looking for the AI ops kind of conversation and then looking for the AI enablement of of workflows conversation. You know, deploying maybe AI agents that when a customer experience environment or deploying AI and Copilot from a a productivity standpoint. So on the AI ops side, it is really understanding and kinda listening for how overwhelmed they are in their in their help desk, their internal help desk, how they’re handling these problems.

Even simply asking things like, well, you know, okay. So you guys are managing your security environment, etcetera. What is your mean time to detect for problems out there? What’s your mean time to respond?

Oftentimes, there’s like, I have no idea what that is. You know, do you have a full time SOC that that it it has a full time SIM that’s gathering all this information has has built playbooks to respond automatically, etcetera. So, on the AI ops side, it’s a lot of times about how are you running your network, what are the issues you’re having, even something as simple as Wi Fi. You know, one of the things we’ve seen recently.

Right? COVID, everybody moved home.

We’re working from, you know, gigabit connections at our home. We got great bandwidth. And all of a sudden, all of us are coming back in the office in the in the last year, and and we’re coming back to networks that were built pre COVID, networks that are five, six, seven years old. And so what we’re finding out is people now are having to revamp their Wi Fi networks. Well, that now there’s a whole group of new AI enabled tools in how Wi Fi networks are managed and built out and build quality service metrics, etcetera. So listening to, I think, just asking what are some of the challenges even coming back to work has been these new AI ops tools can really effectively help people run their networks better.

Yeah. And I think we have a really good example coming up here, Matt, that you’re gonna talk through, a a Wi Fi type deployment that you guys did. But, you know, folks on the call here, it’s very easy to see how all of this stuff starts to blend together. Right?

We’re talking a lot about data readiness here. And I know for those of you who’ve attended our AI and Tech Summit, you know, Kobe does a whole section on that really, really good. He just came out with a data readiness playbook, that you can get out there. So that’s very important.

And then we heard a lot about security as well. Right? You know, that network kind of blending with cybersecurity and all the things going on there. So AI is really driving us into a larger conversation that crosses so many of these technology swim lanes, and I think that’s really cool.

Chandler, go ahead and move to the next slide here. Matt, touch on this a little bit because I think we’ve got a couple of different models here, how we’re looking at AI internally versus externally. So maybe touch on this a little bit, and then let’s just transition right into a couple of those case studies, that you have there.

So I think this goes back to your partners out there as you’re beginning to have an AI conversation with the customer.

We’re really understanding or talking to them about what where are they see some AI opportunities internally focused? Like, you know, I mentioned Copilot is a tool. Right? Copilot in our world is just an incredible new tool for our sales folks, for marketing, for finance, legal.

It’s it’s improved our efficiency. It’s improved the ability for us to deliver accurate answers, etcetera. Right? And so those tools around Copilot and and and the internal focus of that’s been a been a pretty game changing thing for us, and we’re gonna see you’re gonna see that across your customers.

Then when you talk about it, this internal focus as well, which also kinda fits externally, but in the operations of your network, your IT people, when they’re servicing that network, when they’ve got their socks, their knocks, etcetera, what are some of the business outcomes they’re trying to achieve that AI and AI ops can help solve? Right? But then you go external focus. Right? What are the things that they’re trying to build as far as giving them competitive advantages against, you know, against their competitors? What are the changes that they can make using AI, you know, to help deliver a better CX experience, let’s say? Right?

Or what are the changes that they could make inside their website or tools that they could build AI bots to be able to answer questions quicker for customers.

So it’s really I think when you talk to them we went through this internally. We went and basically went department by department and said, okay. What are the internal things you think you’d like to get done? What are the external customer facing things you’d like to get done? Let’s prioritize those and then start to work down the list, and we’re in the middle of that every day.

Yeah. And that’s a great approach. I think that we can use with our customers as we have these conversations and start that AI journey. What are we trying to accomplish internally, and what do we need to do outward facing?

And, obviously, here’s where you see a lot of that blending. You know, CX is more of an external facing tool, although some of the insights it provides are good internally as well. Chandler, next slide here. Matt, quickly touch on, what’s going on with the Jupyter Mist product.

To me, this is a pretty cool product, and we do have a number of suppliers within the portfolio that offer this. But this is something you guys have leaned into fairly hard. So just touch on this real quick, and then, obviously, I think you got some stuff on Palo here as well right after this.

Absolutely.

So for us, we’ve always been a big managed network house. We started delivering Meraki going back seven, eight years ago in the channel number, but service providers did. Basically, Juniper started because a group of people left Meraki and said there’s a better way to build a mousetrap. Let’s build a better microservices cloud environment. As they started to do that, then they purchased a company called Mist that had really kind of perfected an AI ops solution in the Wi Fi world. And Juniper now, when you look at the last four years of Gartner’s, they are the complete upper right hand leader of anything to do with wired or wireless networks.

And and Jason had talked about that earlier. They’ve basically taken AI and proprietary large language models around this and been able to build an interface that says, hey. I wanna deploy six switches and these voice VLANs, and I want you to fix that, and they do it in a natural language.

This AI and machine learning intelligence then is constantly monitoring the network, and I talked about kind of building quality of experience alerts. And so it’s taken a look of, wait a minute. I got a DHCP pool that is starting to fill up, or I’ve got a certain number of users that couldn’t access this that couldn’t get this Wi Fi access point. And And now we can do some repair, but if it can’t do repair, it can automatically do Wireshark captures and open a ticket.

Jason was talking about that. So the ability to basically design a network from the get go using some quality of experience metrics and having this AI ops overlay be able to fix. If it can’t fix, open tickets and do Wireshark captures. And then when people want to to, you know, manage that network, being able to even do it with with natural language.

And radio management’s becoming bigger and bigger. Radio management, the Wi Fi world used to have to be a pretty manual process. Now they’ve overlaid machine learning and AI to be able to take a look at radio outputs and realize, wait a minute. This particular access point has it has got too many people on it.

We need to move people or it’s got too big, you know, power on it. We need to reduce power. So even the management now of these Wi Fi networks can be happening in real time.

And, Matt, one thing I wanted to add is, one of the first, demos I’ve seen of the the Juniper Mist platform was a live customer view where they had an issue with communication between two devices, and it actually pulled up a picture of the device’s connection and the wire connected between them, you know, the Ethernet Yep. Cord. And it would show it would break down the Ethernet cord into the specific wires in there and show you which one was bad. And it said, hey. Replace this wire, and that resolved the issue. Just imagine how long of troubleshooting an IT team would have to do in order to figure out that one specific wire within an Ethernet cord was the was the culprit.

So it’s just really cool how how much telemetry you get with these platforms.

And it it’s drastically reduced ticketing, so it makes us easier as a managed service provider to to to support that customer, etcetera. I mean, little there’s literally stories out there of organizations seeing a, you know, ten time reduction in tickets across these kinds of environments. So it’s pretty impressive.

Chandler Chandler, let’s go ahead and move to the next slide. Touch quickly on what you guys are doing with Palo as well, Matt. And I know, Jason, you’ve got a lot of experience with this as well. You’ve been digging into this model quite a bit.

Yep.

And you you mentioned this earlier. Right?

The massive amount of data that needs to be correlated to build automated playbooks to respond quickly and to resolve quickly, it just can’t be done by a human anymore.

And there’s a number of providers, I think Palo’s way up there in the lead, that has basically built a model that can take inputs from not only Palo, you know, Palo software or Palo devices, but also from a variety of different inputs, other firewalls, other EDR platforms, etcetera. And the idea is being able to take all that massive amount of data, stitch it together, correlate it together, and improve drastically improve the mean time to deck to deck and the mean time to remediate those problems.

As an example, Palo sixteen thousand employee company, Palo, globally. Right? They run their entire SOC with thirteen people.

Okay? Jason, I think when you and I did an event, we were talking about eight. I found out later that they’ve added a couple more as they continue to grow. But it’s amazing that you could build a SOC of that size, that many employees, and be able to run it with thirteen, you know, thirteen folks.

And so the idea that these platforms now can take this massive amount of data, stitch it together, bring it down into actionable things that humans can respond to quickly is is revolutionized how you run a SOC and how you respond. And you guys mentioned earlier these distributed networks, I think I said it earlier, we we are all now everyone of us a firewall. No matter where I’m working from, it’s a firewall, and we better get that data correlated from everywhere and be able to respond quickly. And AIOps has been been able to do that in the security world, you know, pretty impressively.

Yeah. I love that. And and this is, you know, again, such a great product. And, again, where you see the kind of merging of network security, see all these different aspects of a business all come together.

You know, the network is the underallowing underlying foundation for all of this stuff. And, you know, it’s really impactful when you can get AI operating in the right ways and helping that team, streamline. Matt, a couple minutes here. I want you to spend a little bit of time.

You mentioned that you sort of teased it earlier. Chandler, next slide. Talk about this real quick. This is available to our partners on the call here and, of course, any of our tech advisors.

Something that you guys can help out with. You know, we had a lot of questions in here about, hey. Where do we start? How do we have this conversation?

You guys actually do a little one day seminar where you can walk them through. So talk quickly about this, and then I think I wanna open it up, Jason.

Let’s talk about where we think we’re going with this. Like, what’s the next iteration of this look like?

Yeah. So this this is a a one day seminar that we’ll do with the customer either in person or, virtually. We do it at no cost because in the end, it actually opens up so many opportunities for the partner and for us.

Typically, it’s focused around Copilot. You know, mostly people at Microsoft shops. Yes. There’s some Google shops, but mostly people in Microsoft shops. They’re trying to figure out how to use Copilot because it’s not cheap. It’s a thirty dollar, you know, per person ad typically. And so the first part of it, the first part of the day is showing them the entire Microsoft environment, some use cases, kind of the art of the possible up top of what are these tools and how are people using it, you know, across the Microsoft portfolio.

And Microsoft continues to add, like, they’ve said in a product called Fabric, which is all about kind of this data correlation and organizing data. Right? So the first part of the day is showing them some of these tools, and then the last part of the day, the second half of the part of the day, if you would, is now let’s sit down and talk about what are some of the the kind of that that other slide we had or what are the internal things that I wanna work on with, kind of internal efficiencies and operational efficiencies? What are the external facing things that maybe I wanna figure out?

Can I use AI to make more money? Right? And help them spend a couple hours, you know, three, four hours identifying what some of the maybe that low hanging fruit they can attack next, and then we give them back a report. We give them back a report on what that is.

And as I said earlier, we’ve done this now probably, I don’t know, twenty times in the last six months.

We’re literally finding three to five new opportunities in some way, shape, or form when we sit down and kinda each of, you know, each of us open to come on and take a look at what’s there. So this has been very successful for us.

Yeah. This is good stuff. And and, folks, if you, if you guys wanna sign up for this, Katie’s in the chat, as I mentioned. Just go ahead and let her know.

You get you some information on how you can participate in this and get this going. So, Chandler, let’s, take the slide down here. Just go full screen. Couple comments here.

I mean, self healing networks. Zach made a comment in the chat about robots fixing wires.

Kaufman, where are we going with this stuff? Right? What’s the next iteration of this look like? It it’s moving so fast. Like, what what what’s what’s next?

I mean, the AI ops is continuously learning. So we’re gonna see more playbooks, more automation from a software level on not only what the system can self heal, we’re gonna see that adapting, get get more, you know, black by the word, healing. But, ultimately, there’s still a human in the loop. And, if you follow the news, you see all these companies that are trying to push out robotics.

They’re all getting super cool. But we’re probably a year or two out from actually having you know, removing that human in the loop and having an actual robot come in and do all the hardware components. So if you need to unplug a cord and replace it, cool. It can do that.

If you need to rack and stack a new switch and replace the hardware because everything went down, cool. It it will physically be able to do that. So I I mean, say in one or two years for for that holistically may be a little, you know, a little aggressive. But, I mean, I think that’s the future iteration of it is the software is gonna work with the hardware component and completely augment a human that had a human requirement in order to fix a lot of these issues.

No. No.

Let me add a little bit to that too.

I think we’re gonna see more and more focus on application performance.

These networks now are starting to be able to keep track of the app actual application I’m using, where it’s running across the network. As you mentioned, you know, you saw Juniper being able to say, hey. It’s the problem with that cable. These technologies could even now start to say, okay.

I’m running that application from my desktop. I’m working from home, and I’m trying to get to, you know, you know, such and such website or SaaS application across the other end of the earth. Being able to detect, no. The problem is actually my laptop.

The problem is in my Wi Fi locally. The problem is in the middle mile. You know, Powell’s got a product called ADEM that can tell that Cisco’s got a thing called ThousandEyes that can do a lot of that work. And so it actually be able to keep track of application performance even when I’m going to the other end of the earth on the SaaS and being able to alert on it, you know, have intelligence, fix things, etcetera, in the middle wherever those parts are controlled by AOPs.

And so I think part of that journey here is we’re gonna see more and more focus on application metrics and these AI and AIOps tools being able to fix things to guarantee performance of applications.

Well, we’ll know we’ve reached there when one day it’s a robot Doug hosting the call and not Doug himself.

So There you go.

Doug, what do we got out there for questions? Obviously, a a lot of information we threw at the folks here today.

A lot of good questions, a lot of good interaction. What what do we got in the chat there? Anything we wanna highlight here with the team?

Absolutely. And there’s not a robot out there yet that’s got hair like this, so it’s it’s just not going to happen.

Matt, great presentation, Jason and Graeme as always. And I wanna thank Katie too for expert work in the chat. One of the themes that keeps coming out of some of the questions is, you know, it seems like we’re seeing another one of these shifts in the way that we work. IT administrators, network administrators, seems like much of what they have done in the past can now be done by AI, but it seems as though the shift in what they’re doing is to watch the AI.

So what do you recommend beyond the technology that network administrators and others involved in this can do to help prepare themselves and continue to qualify themselves to be able to function well in those new roles?

I would say Go ahead. I’ll by oh, go ahead, man.

No. No. Go ahead, Jason. Go ahead. Let me think about it.

So I I think the the first thing is comfortability. So learn that AI is not a hundred percent here to replace. It’s here to augment a lot of the the tasks that you need to do that are mundane, that, troubleshooting that takes you out of the strategic mission that you have to keep business continuity. You know, you get pulled away to do something because an incident happens or something like that.

Figure out where AI can help you solve these problems, whether it’s using the human language to create a network or optimize it or looking at the user experience on the application level. Learn what tools you have access to to make yourself better and more efficient because the tools are there and somebody is going to be using it, so you might as well do it as well. It’s the same thing that we talk about whenever we’re using, like, large language models for Copilot implementations and that type of stuff. It it transitions really well into the network management, realm too.

I think it fits incredibly well with what we’re what we all do in the indirect channel, what partners and what we’re doing for customers. We’re walking into customers and saying our job is to bring you a group of managed services that alleviate the day to day quackmire you’re in to elevate your conversation to start working on business analytics and other things that make your company money and give you a competitive advantage, let us get out of the day to day quagmire work. So that’s what we’re selling with managed service as well. So these capabilities are only gonna help us manage service providers build better products that can be sold well in the channel.

Yeah. And I think it’s it’s just about doing more with less. Right? I mean, if you think about so many of us here work in that SMB space and a lot of these tools and functions, you know, complex network engineers, that’s something a lot of SMBs can’t afford. But maybe, you know, they can bring in AI here to help them with that. And so what you used to have to do with ten, twelve employees to Matt’s point earlier, you can now do with eight, you know, or with six. And I think that’s really the power of this, of this model and these technologies.

Jason, you mentioned something earlier on that I wanted to touch on as well. You talked about, it’s great to have the visibility under a single pane of glass, but you also need the reporting of the issues. And this is where I tie it into the last question, the troubleshooting steps. It seems to me that the AI can assist those network IT and other managers in knowing what steps to take to be able to discover further or find more information that they need or what steps and processes they need to follow to ensure that they’re completing their responsibilities accuracy. It really is a tool to be able to help them.

Yeah. So as this thing continuously learns, it it verifies issues across every single customer base within these platforms. And it learns, hey, this is a common issue and we’re seeing it across here. Here’s what the steps have done in order to remediate it.

So now we’re going to create a playbook or a recommendation on what to do next to optimize that employee’s troubleshooting steps. We’re seeing the same thing in the cybersecurity space when it comes to like the sword now adapting into the SecOps to where you’re automating that triage and the steps to perform in order to remediate everything. That’s where that the troubleshooting comes in is because instead of coming in and saying you got a blank slate on, okay, I see this maybe as a potential issue. Here’s the data I have on it.

Now you’re getting steps one through five that you can try that the you get this list of the steps that the system already has done, but then now you have steps one through five that you can try in order to remediate this that the system wasn’t able to perform. Yep.

Great answer.

Guys, we are out of time. I’ve gotta move on. But, Graeme, finish it up, summarize this up for us, and thank you to all of you for a terrific presentation.

Yeah. Thanks so much, Matt. Really appreciate it. Jason, as always, appreciate your input. Folks, you know, this is an evolving space.

We’re gonna see so much more coming down the line here. You know, we could probably come back six months from now, talk about AI ops, and it would be a completely different conversation with new things in there. So stay tuned. This is evolving.

The great news is these these technologies are here to help your customers be better at what they’re doing. And as Matt said, get them out of the quagmire of doing some of these things. And And as I started the thing, we wanna make sure that you’re not taking those late night calls anymore, missing those family events. AI ops is here to help you do that.

Thank you, everybody.

Last word, Matt on CBTS. Great job in, Vegas with the karaoke idol.

Let’s bring Graeme and Sam out here one more time for one more chorus.

Come on out, Sam. That’s right.

There’s a few here that dispute the results on that, but, great job down there. It was a lot of fun to see back here on our end. Great.

Yeah. I’d always say thank you. Yep.

Thanks, everybody. And thank you, Katie. Thanks, guys.

Graeme and, Katie, will, continue to be in the chat window along with Jason if you’ve got additional questions.