The history of Comsof and the foundation of automated design software.

Hello, my name is Steve Tongish and I'm the Chief Marketing Officer at IQGeo. And welcome to Network IQ. I'm really pleased today to have one of my colleagues with me. This is Raf Meersman. He's our General Manager of our European Operations here at IQGeo. But for the sake of the conversation today, we'd really like to talk about Raf's previous role, where he was the CEO at Comsof. So welcome, Raf. Thank you, Steve. For those of you who don't know Comsof, they've been around for 22 years, for a very long time anyway. And Raf's been with them for 22 years. And Comsof has made their reputation providing really the industry's leading automated design software for fiber and utility networks. And as I said, Raf's been with the company a really long time, first in technical roles, and then later in management. And so what we thought we'd do is to do a lot of work. And what we thought we'd do is take a look at the history of the company and the technology and give people a little bit of a background of what Comsof is all about. So let's just jump right into it, Raf. With 22 years at Comsof, you've seen a lot of change. Maybe you could back up a few years and explain to us how the company started and what the technology is about, because you've got some really very bright people, mathematicians and analysts in the organization doing some amazing software. Yeah, yeah, indeed. And it's been indeed an interesting journey with quite a long history. In fact, when we started back in 98, as a company, the idea was already to bring smart automated planning for telecom networks to the industry. So to develop these new software algorithms. And where back in the late 90s, the industry, telecom industry was focused still a lot on the development of a backbone network, we had Comsof switched pretty quickly, around 2004, maybe, with a full focus on fiber to the home. So the access part of the network. And that was really because we saw that that was where the big challenge for the next decades of this industry would come from. And so we have indeed build out a very strong team with mathematical background. But I would say what maybe differentiates us is that it's also a very strong engineering background. Because what we discovered pretty quickly, if you want to automate the design of a network, a telecom network, you really need to know what you're designing. So you're designing. So you because the end result should be a buildable network, people will actually construct this network. And therefore, it was very important to actually combine mathematical algorithmic experience and background with an engineering view on this. And so we really, in the end, we developed more than 15 years. So I guess more than 150 person years of development went into our own in-house algorithms to solve this kind of complex design problems for fiber to the home networks. It's a very sophisticated environment and very complex. So today, what does the typical fiber network design process look like? What does the typical fiber network design process look like? We are now really in a major wave of infrastructure rollouts. And that is indeed where this process is hard to maintain and where you're running into the limitations of this kind of planning process. So scalability is clearly a big challenge in a manual environment. Plus, I imagine also linking it together has to be really difficult. So what does the automated planning technology bring to that environment? Yeah, so everybody will associate automation, of course, with saving time from people, from the people who would actually design the network. That's what you think when you say automation. But actually, what we found in this process is that what we're bringing into this process is actually even more benefits on the consistency of network designs and the quality of network designs. Making sure that every design on every area is following the actual engineering rules and that it's without any errors. Because any error in the design would actually introduce an extremely high cost at the construction side. So it's actually especially those kind of benefits which we found to be extremely high. We actually did an interesting test already maybe seven, eight years ago. We had this one neighborhood that needed to be designed and we handed it out to three experienced manual planners. And they each had the same guidelines and they designed each a network on this area. And it will not be a surprise. We got back three valid designs on a first base, but very different designs at the same time, because they interpreted the rules a little bit different. They started a little bit differently. And the surprising thing was when we would actually calculate the cost to construct these three designs, we found a variation of five to six percent between the cheapest and the most expensive ones. And that is also where, thanks to our automation, we can actually focus and consistently get that cheaper design, make the choices that actually make your construction cheaper. And that is actually found to be one of the biggest benefits that we can bring into this process. Yeah. So really optimizing that design in a way that's consistent can, as you say, can be built, because that's the important thing. Having a design doesn't help if you struggle to build it properly. So as people are making this transition from a more manual design environment to an automated environment, that's a big change. Is there any piece of advice that you would offer to people who are starting to embark on this journey? What's the one thing that they need to think about? Absolutely. There's definitely some advice, because like you say, it is a big change. The process does change a lot. And I always also say it's not magic in the sense that it's not removing all the hurdles that you have. In fact, the biggest change and the biggest advice when you want to go to automation is that you have to be very careful and spend probably some more efforts in the preparation and the setup phase of the whole project. Because automation obviously also relies on input data. And the algorithms need to have an accurate representation of the real world, of where streets and address points and existing infrastructure are positioned in the area. And so this quality of the input data is extremely important, because garbage in is garbage out, we always say. So obviously, in the process, you may need to spend a bit more time on data preparation. But then obviously, you save the time in the actual design process. And especially in case, and that's what you often see, that you would need to start iterating designs and revisiting a design. For example, a pop or a central office, if it would change the location, well, in the automation and reiteration is very quick, while in the manual process, that would be very lengthy. And so that is extremely important that there's enough time spent on this preparation stage. And that is something where, yeah, it's also important when you step into this to step into that with the right expectations. And overall, again, the benefits are at the end. So on the cost side, more maybe than even on the time side. Right, right. So it sounds like really kind of like any big automation process. You need to spend that upfront homework, getting things right. And then you can really benefit from the automated environment and the automated capabilities. Yes. Yes. And maybe adding to that, just that, for example, the size on which you apply this automation is therefore also, of course, important. You could compare it with like, if you want to go to a shop in your own town, yeah, you won't take a helicopter. Because by the time you take out the helicopter from the garage, you already went back and forth with your bike. So you need to fit it to the challenge and the size of a project. Yeah, I'm almost careful taking the helicopter out myself, Raph. So good point. Okay, so let's wrap it up. So as I said, ComSov joined the IQgeo team earlier this year. Maybe you could just quickly, as a final question, tell us how the automated design software fits into IQgeo's network lifecycle product suites. Yeah, we naturally fit very well into that lifecycle, obviously. And mostly initially on the early stages of the lifecycle, when you look at the planning and the design stage of the network, then obviously this automation comes in and adds a lot of value. And of course, we're working on the integration of the software tools so that that process can also be very smooth and very efficient from a data exchange perspective. And so that's very natural. But the very nice thing for ComSov is that now, thanks to this integration and coming together, we can actually also expand the usage of our smart algorithms to operational stages in the future. So we still see a lot of opportunities to build further on that. And the other thing that is very interesting is that we come, as I mentioned, from a fiber and a telecom background. But we also had started developing and expanding into other utilities and electricity networks, for example. And also there, there's a very good match with the IQgeo product suites that is also available for electricity and utilities. And so we also see there are a lot of potential future developments from a ComSov side that fit very well into the IQgeo portfolio. Yeah, and I know, and of course, we know, Raph, that we were working together on customer projects before the acquisition. So it was a really nice fit. And from my own selfish interest from a marketing point of view, it's been a really great addition to our story and the capabilities that now the greater IQgeo team can bring to our customers. So super. Well, thanks for taking a few minutes today to go through this with us, Raph. My pleasure. And really, the history lesson and the technology lesson. And if anyone has any questions about this, you can reach out directly on our website, and we'd be more than happy to follow up with you. So thanks very much. And thank you all for joining Network IQ.