Fiber network planning & design: Modernizing workflows for faster planning.

Welcome to Bitesize Fiber, the podcast where we break down the biggest challenges in fiber and telecom one byte at a time. I'm your host, Ellie Pulse. In Episode 3 of our Planning & Design series, we're exploring how outdated workflows and manual processes can stall your fiber rollout, and how AI can help modernize planning for speed and accuracy. I'm joined by Stephen Baker and Niren Van Averbeka, who share what it really takes to streamline design workflows, why disconnected tools create unnecessary delays, and how AI-powered design automation is helping operators plan and scale networks faster than ever. If your team is ready to move from spreadsheets and manual tasks to a modern AI-supported planning process, you won't want to miss this conversation. Let's dive in. Are companies and teams are still using the same workflows or do you think those have gone updated over the years? What would some best practices be if you were going to change your workflows? My recent company experience was to build really over-complicated workflows that required a lot of data to be downloaded, uploaded to the workflow application. And it was also built around, I'll call it, I'll call it, management by metrics, because each task had an SLA and you were measured against the SLA for your work versus the quality of the work that you were going to deliver. So companies continue to change them. Do they do them the right way all the time? Probably not. I think there are most companies would lean towards more efficient workflows to get the job done better. It's just some of these companies, they're so big, they're silos, people have different agendas. You don't always get the job done better. You don't always get the best workflow in the end. It's easy to fall into the rut of this is the way we've always done it. But not taking into account the business is changing, you're not. So how are you going to keep up efficiencies if the company is changing, which is being driven by customer demand as to how you have to change and adapt to provide services to them? And looking at a typical network rollout, historically, there will be a couple of different systems that are used for different types of actions. Like you have a planning tool or a planning software, then a system of records and a couple of other ones as well. And I think one of the best practices that I've definitely seen in the last couple of years is that the companies who succeed in linking those applications, interfacing the different parts of rollout, do get to success. And even better, I think if it's possible to have a single software solution in the middle that controls planning, design, rollout, maintenance, and so on. But all of those activities centered around one single software in the middle, that really is a good basis for success. AI is obviously playing a growing role in our businesses and our workflows, network planning as well. Have you guys seen any examples of where automation or AI driven tools have helped speed up or improve the design process? Yes, there are probably more than a few tools out there that do this, but obviously Thumbs Off brings that to the table with AI. I'm going to call it AI. It's probably a little bit more algorithm logic, which is the root of AI to begin with. You have to train something. You have to train artificial intelligence. But I think it's been instrumental in the planning process to determine how much is it, you know, putting together an estimate. If I'm going to fiber overbuild a city with 10,000 homes that I'm going to service, I don't just go out and start designing and building. I've got to plan for it. I've got to make sure I've got the financial backing as well as many other planning processes, working with utilities, working with municipalities to be able to go out and do it. So if you're getting that up front, but then on the backend, doing high level designs and then into low level designs. So in a previous role, I left a program that I put together, which was leading towards design automation. And the goal was to virtually eliminate a designer from drafting in a very complicated system of record GIS that required a lot of manual input data to input it in there. It was going to take a lot of work. Right. Transitioning data models into a containment type model and then implementing the ComSoft products with this other system of record and the business case. But the senior director, executive director over the design team signed up an almost 70% cost savings in the process. To my mind, there's actually two main components of AI. And the first is one that we already mentioned at the beginning. That's the data. So for any AI to work, there needs to be data. That is available. Large sets of data that have been tagged and can be used as training for any AI. And then the second bit is actually the training. So the fact that a planner or a network expert or whatever is tagging the data saying this is a good decision. This is a bad decision. So in the end, it all boils down to making sure that the data is available in large quantities. That based on this data, certain general trends can be recognized and tagged and then applying that to the new data. So the fact that any network rollout generates so much data because you're covering a huge geographical area. The GIS data is in there. It needs to be parsed and tagged. I think there's definitely an opportunity for AI to start with, but also while doing a major rollout to increase the data and the analysis that you can do on that data. So I think the further you go along with a rollout, the more data there is, the more it can be tagged and the more efficient the AI predictions will be moving further along in any network rollout. And then there's also the as-built data. So pictures that have taken of splicing trace of the internals of cabinets and so on, that then can be parsed by AI and seen. Are there any changes here, any difficulties, anything that could result in a degradation of network quality, splices that are missing? The first part to be able to use AI is generate data, centralize that, tag it, and then see how it can be used to increase the quality of network. Yeah, so that's really cool. Yeah, so that's really cool. I know we've seen some new tools come out. One of those is Diplomatic. Have you been able to get familiar with that? Can you tell us a little bit more about Diplomatic? Yes, indeed. Diplomatic is an important partner that IQgeo has been working with for a couple of years now. What they allow to do is have, after the deployment has been constructed or a piece that has been constructed, like a splicing tray, a cabinet, a manhole, to take pictures and then parse those pictures on the fly and identify in the field that something may have been mislabeled, misrepresented, misspliced, and so on. So it's first a way of validating the rollout on the spot, but also afterwards having a digital trace of what has happened in the field. And that reverts back to a previous topic that we've mentioned. The fact that once constructed, you need to make sure that the as-built data is brought back into your data systems. So it actually plays a double role there. You are validating the rollout and at the same time making sure the records of your inventory system are as accurate as possible. That's cool. I'm excited to see the role that Diplomatic will play with IQgeo.