AI meets network management.

A Hello, everyone, and welcome to this edition of Network IQ, the topic for today is IQGeo’s exciting acquisition of Deepomatic, the telecom and utility industry's leading AI provider of computer vision software. And to discuss this exciting topic, I'm pleased to have David Cottingham, the CTO at IQGeo, and also Augustin Martin, who is one of our new employees and the former CEO of Deepomatic. Let's back up a little bit. For people who aren't familiar with the acquisition and maybe not familiar with Deepomatic, let's start by explaining a little bit about what AI computer vision software is and the benefits that it brings to the telecom and utility industries. So, Augustine, I'm going to put you on the spot. Give us your elevator pitch. What's AI computer vision software? Yeah, thank you, Steve. So, basically what we do at Diplomatic, we turn every camera, every sensor that is capturing information in the field. So, mostly field workers' cameras into a smart sensor. So, we add brain to the camera to make sure that the device understands automatically the situation, the context and what's happening in the field. they capture all the information, the data available in the field to do several things. The first one is to automate the reporting of the jobs from the workers or the technicians, the field technicians. And the second one is to verify, we verify with AI that the job is done accordingly to what is expected to the quality standards and to the statement of work. And so, if anything goes wrong, we will tell the worker immediately to make sure he can course correct its action and complete the job right the first time they are in the field. it increases the first time right, the productivity, overall quality of the network. It has several financial benefits. So, that's what we have been doing in the past. Five plus years. And we are going to do much more now that we are a part of IQGeo. Excellent. So, right first time field operations has been the focus to date. and that's what your existing customer base is. Those are the benefits that they're seeing. Yep. So, how did we get here with this acquisition then? You've been involved in the process from the very beginning. and there's a lot of history between Deepomatic and IQGeo. So, Deepomatic's been a technology partner for years now and you've been at our events and we've done things. But from your perspective, how did the relationship begin and how did we get to this new partnership? initially, we were starting to develop relationships with the ecosystem, but more on the field app side, FSM side. But like three years ago, it became quite obvious that we could add a tremendous amount of value by feeding the system of the record system. And in parallel, we met you guys almost three years ago going to your event. That was in Denver, I think, at the time. And we met a great, also a great company that was growing, that was healthy, that was sharing our culture. So, little by little, it became quite obvious that we simply could grow further and quicker together, joining a company that was growing together, joining a company that was significantly bigger and with a lot of funding to keep growing and conquer the market. So, the journey was very exciting. And as we had known each other for two years, it was quite natural to merge and to be bought by IQGeo. In addition, you were one of our partners that genuinely, without asking for any sort of compensation, allowed us to meet your customers. And within this first year of collaboration, we were not even together. We had no letter of intent of being acquired. You allowed us to meet several prospects, such as Connect Fiber, who became a customer even before this acquisition was completed. So, the synergies was already proved from our point of view, even before the completion, the acquisition of the acquisition. And the joint partners, the joint customers, that's proof of that synergy and how the technologies work together. And we're seeing more of those coming to the fore, aren't we? So, there's more joint customers appearing. So, great. So, David, with that context in mind, and Augustin has touched on some of this, but can you explain how the Deepomatic AI technology fits into the IQGeo portal? and what it means to our customers? Because I think we see this as quite a compatible fit for what we're doing. And the other observation that I have is about why AI and why now? I've been to a number of shows recently, fiber broadband shows, and AI is just everywhere. And everyone is talking about AI. What's happening? Why is this the moment for AI? You know, it's funny. I think you're right, Steve, that everyone is talking about AI, and it's the best thing ever, and it's going to slice the bread and make the tea. I think there is obviously huge value in AI, and we see that not just in Deepomatic, but in lots of other places as well. What the tricky part is for, if you're in telecoms or utilities, you're going to deploy AI, and it's going to be great, except that you realize the thing you've always known, which is that poor quality data to start with, you're feeding into this machine, and you're not going to get great outputs at the end. And so we talk about digital twins, and digital twins needing to be accurate and like good stuff. And it's true, because when you are feeding these data to an AI, you need to fix your data to start with, and your digital twin to make it accurate. And so really, where D-Formatic is interesting is we come in and we can say, aha, you want to deploy all this AI, we can help you fix all that data you've got to feed that AI with AI. And so Deepomatic is our, if you like, secret weapon to go and fix all of those data quality issues that your utility or telecom provider might have, bear in mind, they have telecoms that have made 20 acquisitions in the last, you know, 10, 15 years, and they've kind of pushed all these data together and crossed their fingers and hoped it was fine. Or you've got incumbents that have been there for 30, 40, whatever years, and some of the data is that old as well. It's not very accurate. Being able to deploy Deepomatic to help fix that is really, really important. That's why it's the time for D-Formatic, because we see all this talk of AI, we're actually enabling other AI use cases by helping to fix the data. And so where we see it fitting in with IQGeo is very much in the first chapter of where we go, is how can we take that right first time message that Augustine talked about, and fit it onto, correcting all those data that you have in your system of record, and then going forward, can we help you fix further things in other areas, right? And there's lots of talk about there as well. But that's really how we see it fitting in, helping you to deploy AI to fix data to help AI. This is a, in a way, David, it's not a new story. We've been talking about network data quality for a long time, haven't we? And I know that the marketing team has done our own research about some positioning activities that we're doing, and poor network data quality came up as the leading concern in both the telecom and utility industries that we work with. And so it's, yeah, it's, it's that, it's that kind of elephant in the room, isn't it? It's really, you're totally right, Steve, it is the elephant in the room. And some providers know how big the elephant is, and some have no idea how big the elephant is. And when, you know, we've gone into customers, and we've shown them, in some cases, the quality of their data. And there's been a kind of very scared expression comes up, because they didn't realize quite how big that elephant was. And so now, yes, we have a story about how we can help them, you know, digest the elephant to take that analogy one step further. Yeah, this, this kind of goes to my, to my final question here. Not only are we, are we helping with what the diplomatic software can do today, to make sure that there's right first time in the field. But the next step is to look into the future, and to say, how can we influence that data going forward? Because simply putting in place digital digital processes won't be enough, there needs to be something behind that doesn't doesn't there. So so my final question, and I'll pose this to both Agustin and to David, is, what is the future hold for this technology? Kind of let's not look too far in the future, but sort of in the longer term. Agustin, do you do you want to go first? Yes, sure. So first, we have seen like that for our customer, the simple premise of having a full accurate digital twin is already gigantic. But when you have that, you can start to install a lot of automation layers. And automations for several additional use cases, so predictive maintenance, payment of contractors, and all these processes that are becoming automated, are becoming automated, obviously generate a lot of savings for the customer. And this is like, very soon. Then, in the midterm, you can apply everywhere as a MLOps fundamental layer, in order to generate overall, a network that is self acting, self predicting, suggesting some actions, and some triggers to the back office users, which is obviously the next step and the big thing and the vision overall of agentification. Yeah, it's that agentification phase. It's that it's that a word that we keep hearing all the time, isn't it? David, what would you add to that? I mean, I love that Agustin brought up predictive maintenance. That's one I've gone on and on about. I think that's massively valuable to people instead of dealing with emergencies, you kind of predict and deal with them on a sort of non-reactive basis. I do think that we have the opportunity to not just look at images, right? But we can look at video, and we can look at all sorts of other sensor data. Because at the end of the day, looking for patterns is kind of what, you know, deep matter is really good at looking for patterns. And it doesn't just have to be images. And so loads of things, sensor data of all varieties that we can fuse together. And yes, predictive maintenance is one scenario where we could enable that. But to Agustin's point, there are lots of other things you can do where if you take all these sources of data, which individually don't get you very far, but when you put them all together, you can start saying, ah, I think there is this that you should do. You should expand your network. You should fix it, whatever it is. That is super valuable to providers who today are very much reacting. Thank you. Well, so we've been talking, as we've been positioning this about this first phase being the integration, which we're doing today. And we talked about the existing customers that are already using the product together with our network manager solution. Next step is to take in and embed that AI technology into the network model, right? To make sure that we're constantly monitoring what's going on with that network model. And then the third stage as we move forward is about the agentification of that, taking action, as Agustin said. So that's a really exciting future. Any final words from either of you? I just wanted to say that this vision and action plan is already shared with all the employees and everybody is super excited by what is ahead. So I can't wait to start and accelerate. Okay. Well, I think we have started. So keep accelerating. Absolutely. David? I just take what we've said really that it is, I know everybody says it's super exciting. It is really exciting, right? There's so much that we can do here. And I think we're just scratching the surface. So really looking forward to doing more. Wonderful. Thanks to both of you for taking the time today to join us on Network IQ. This story will keep going and there'll be more information. So check out our industry events. Augustin mentioned that, our meetup events, There's one coming up later in the year in the US. So yeah, stay in touch. Check out our blogs. And we'll keep you in touch about what we're doing and keep you up to date on our evolution. of the Deepomatic and IQGeo strategy. So thank you for joining Network IQ today. And guys, thanks for your help. Thanks, Steve. Bye-bye. Bye-bye.