Episode 12: How AI is Shaping the Next Generation of Telecom

Bob: Hello and welcome to another episode of Bob Friday Talks. Today I'm joined by Tim Fell, VP of Networking Technologies and Services at TELUS. Today we're gonna be looking at AI as a next step in the evolution of automation and how it fits into TELUS's strategy. So, Tim, before we get started, maybe a little bit about your background and maybe a little bit about TELUS and how, when did AI really become real for you at TELUS?

Tim: Sure. Thanks, Bob. And great to join you today. Excited to be able to chat with you. And my background, I've been at TELUS for a few decades. And mostly I've been in the networking and broadband and TV area. So, I spent a lot of my career building TV platforms and all the networks that support it.

And then most recently, I have, taken over the core network for TELUS and the automation associated with it for IP and transport and, broadband networks. So, AI kind of, or sorry, I guess a little bit about TELUS. TELUS is one of the three largest operators in Canada. So, we have a national network for our, all of our services, whether it's mobility or broadband services or enterprise services, etc. And, we really started on the AI journey fairly early. And so it has kind of been woven into the whole fabric of TELUS, I would say. And so we've enabled, Gen AI tools for everybody at, tell us to use.

So, any employee can go on to the, the hub or portal and use, AI, in their daily work, if they like. And we have also used it a lot in, different organizations to enable them to look for efficiencies, in their own teams as they, as we all learn more about AI and how it can be effective in our day-to-day, operations.

Bob: Yeah. So, I mean, I always look at AI as kind of that next step in the evolution of automation, and as you kind of mentioned, right. There's kind of internal operations. I think almost all large corporations are looking to leverage Gen AI to improve, make more productivity, is TELUS actually looking at using Gen AI for both internal and customer facing, you're the biggest operator up in Canada, are the customers in Canada going to start to see the benefits of Gen AI in their experience?

Tim: Yeah, we've definitely, we've started using it already in our, customer facing operations as, as a copilot for, agents and other employees that are serving customers. So, we've got it in use in that regard. We also make use of it in a lot of the different operational areas in TELUS. For example, one of the areas is our field operations team.

They have used it quite extensively and it's quite remarkable in the last two years to see teams like that, that have really embraced it and have been able to use it themselves to find opportunities to make themselves more efficient, to be able to dispatch technicians, for example, and, find outages faster. So yeah, it's really, getting broad usage across TELUS.

Bob: Yeah. So, if you look at AI, nowadays, AI and Gen AI are almost synonymous, right? They almost, I think we say that now there's other AI involved. Computer vision is TELUS looking beyond Gen AI is tell us looking at other opportunities to leveraging large models to make a better customer experience.

Tim: Yeah, definitely. There's a couple of parts to that, in one case, we've used specific machine learning for being able to for example, if a fiber gets cut, a fiber might have thousands of services on it and thousands of customers with different endpoints. What we've been able to do is implement some machine learning to generate that list of customers and what the impacts are immediately as a first step in being able to, able to find new routes or new paths, not routes for customers.

And, yeah, that makes it, much better for customers to, reduce outage time and, hopefully, prevent outage time entirely.

Bob: Yeah. I mean, when I look at, the internet, it's kind of right up there with, power and water, right. I think everyone's become very dependent on making sure their internet is up and running at all times.

I'm almost, you know in my household I don’t know about you, but, if the power goes out, we can live with that. The internet goes out. We, that's, you can't really live without the, the internet nowadays. Maybe, I look at the, big mobile operated, carriers like yourself, when my internet gets slow.

Are we starting to leverage AI to start to figure out why I have slow internet? Usually it's either at my Wi Fi in my house or somewhere in my, either that I'm calling up my carrier to find out what's going on.

Tim: Yeah, I think we're just scratching the surface of that. And I think that's an area that we're doing it. We're really focused on in my team and at TELUS. So step one is to get the data aligned in a universal inventory system so that we can, yeah, accurately find out what's going on. So that's an initiative that kind of crosses both the network and the IT teams. And so that's kind of a first step. But then, once we have that in place, we'll be able to take advantage of the, start to look for outages, using machine learning to look for patterns in all of that data that will, enable us to, find those, outages or prevent outages that are that are coming. So, we have a number of use cases enabled already on that.

But I think that's the big promise for us on the network operation side. Also on the provisioning side, but really from a customer differentiation perspective. It's, as you know, as you mentioned, I think I saw one survey that that said people actually rank, reliability of the internet above running water.

So, it is very important, for not just my family, but all the businesses we support as well. So, yeah, there's a lot of work to do on that, and I think there's a lot of opportunity, not just at TELUS, but to be able to work with other operators and companies such as, Juniper and others to federate that data so that you can, incorporate more learning and, resolve those issues much, much faster.

Bob: Yeah, you can't talk about dating. I'm famous for saying, like great wine starts with great grapes. Great AI starts with great data. I think everyone's starting to realize there's working on automation and AI that, their data strategy is kind of the foundation of all this.

If you don't have great data to work with in any words of wisdom for the audience on, big operators carries, you guys have tons of data, how's your data strategy and streaming, and, I think we're going from kind of a pulling data out of devices to streaming data, anything you get help with on how that's working at TELUS.

Tim: Well, it is a big, like I think every operator, it is a big, issue. We've, have a lot of history of building databases and different systems that, make assumptions on, they either take data from the network or they build more data to put into the network, or they make assumptions on what's in the network.

I think that the starting point for us is to assume the network is the source of truth and synchronize all the data to that. I wouldn't say we're there yet, but that's the first point in the strategy. And then there's a lot of, we have a lot of legacy systems, so there's a lot of collaboration across teams to put in place to agree on how do we, you know, what is going to be the source of truth.

The network is the source of truth, but then that data also gets replicated in different points across the company. So, yeah, there's a, there's a lot of, collaboration and human involvement to make it happen.

Bob: And you look like you're almost my age, maybe not quite, but you mentioned your, this looks like you've been, you know, tell us for over 23-24 years now, it sounds like you've seen automation change in other ways. I'm always interested in your journey. When did AI, what year was it when AI got on your roadmap? Cause you know, we've been automating things for decades, this AI trend become relevant to what you're working on.

Tim: Yeah. I don't know when they, when I would start calling it AI, we we've kind of been working on the machine learning for different aspects for a few years or using big data, to make better decisions and, and try to do analysis.

So, I would say the last three or four years, in some forms, we've been, really starting to see benefits out of it. But yeah, probably depends a bit on your definition.

Bob: Yeah. Which is probably another good discussion. The definition, for me, I always tell people, we've been doing machine learning for years, it's really when we got into these, very large network, these very large neural networks with large data sets.

I mean, that seems to be what's disrupting the industry. You know, we look at the automation, what's different about using AI for automated versus what we've done or, what is your definition of AI?

Tim: Well, I think, when I think about it, it’s capabilities that more mimic human behavior. So, and how does it, how does it, do that? As you said, we've been doing machine learning for a long time, but I think it boils down to, the capability for pattern recognition or guessing. And the other one is the group force and ignorance method. We've got like a lot of data to get through, but you need a lot of computational resources in order to do it.

And if you can accomplish that at speed, which now you can through the use of, large language models, then you really start to, to see benefit that can extend across all the different areas of the business that we, that we talked about.

Bob: When I look at, large carriers in general, they're not, I wouldn't say large carriers are known as a speed rabbits or, being fast to draw when you look at the adoption of AI, is it a culture component or people that, is the teams at TELUS actually adopting AI really, or are they reluctantly adopting AI or do they see it as a threat?

Tim: I think there was a little bit of reluctance in the beginning and then I'm sure there's, pockets, everywhere, but we, we did embrace it, quite early on. And we do have our own division called TELUS Digital, which was called TELUS International, has a lot of offshore development and a lot of capability. And so, we were early into building tools that everybody could use in the company. And, like I say, that ended up, two, three years ago, we would have, a lot of the operations teams in order to realize efficiencies would have to, like, log a systems project and wait for resources or budget or scheduling in order to make any improvements.

You've got to try and get a business case to get IT resources, and start making changes to systems. Now I've seen a lot of those teams are really literate in AI and implementing things on their own, and they're going to the IT teams and my team and saying, I need API access for this so that I can make this improvement in my business.

So it is really liberated them to enable them to be able to take control of their own destiny and start making efficiencies in their own business without having a reliance on a central IT team that might have been the case four or five years ago.

Bob: So, the teams are starting to learn to trust for me, I was in my first Waymo, Uber, self-driving taxi there last month. And it's like, it has a certain level of trust, right? When you start to trust technology, do the right thing. It sounds like same thing there. It's not like teams are starting to learn how to trust their AI assistants.

Tim: Yeah. Definitely, I think more so than self-driving cars where I'm from, because there's a lot of snow on the ground here right now, and I'm not sure I would trust the self-driving as much. So, I'll go with your experience on that one.

Bob: Yeah. Now maybe, Tim, the other thing, so, like, I'm an internet addict. I'd rather have my internet than running water most of the time. Where do you think, for mobile wireless users, broadband users, where do you think the biggest impact AI is going to have for the everyday consumer out there?

Tim: Yeah, I don't know about the everyday consumer. I think the, the nature of, from my standpoint, looking at it from core networks and, the automation of that, I see very large changes in what's happening in consumer behavior.

Most recently, example was the Mike Tyson, Jake Paul, fight on Netflix, a live event on Netflix for the, one of the, I think the largest over the top streaming event. Certainly in North America, if not, if not the world. And so when I look at that and that trend toward those types of events taking place, and the reliance on the network, I think that we're moving to an era where the network is going to be much more cloud like, and we're going to, move to this, we've all talked about, probably the move towards autonomous networks.

But in order to get there, you need AI to be able to analyze the network to help you do all that automated provisioning and so on and to be able to scale resources up and down. So, I think that in some sense, consumers are going to pleasantly not notice what's happening. They're going to be able to partake in all of these different live events that are going on and virtual reality and so on.

And the network is going to be able to adapt to those demands for the peak demand and also be able to scale back down. Do things like save power, when the demand is not there, as one example.

Bob: Yeah. I mean, any power company can have the same problem, they have to design for the peaks and adapt quickly to that. But so maybe to wrap this up, Tim, any words of wisdom for other people you've been, you're working on automation right now? What are your one or two top priorities right now that you're focused on delivering?

Tim: Well, one thing we say here, especially when it comes to AI. So, we're putting the data in place and we're going to focus on, improving our systems to, be able to diagnose the network problems and reduce outage time and so on. But one thing we talk about with AI at TELUS is, fall in love with the problem and not the solution. So a lot of us engineers, especially want to figure out the problem ourselves. And, with AI it's a little bit more about defining the problem and using AI to help you solve the problem.

So that's where we've seen in some cases, you know, operations teams or frontline teams able to embrace AI a lot quicker than a highly trained network engineer who might be, reticent to use AI to solve the problem. So, that that's where we say yeah, fall in love with the problem, don't fall in love with the solution.

Bob: I can't agree with you more. I always tell people not everything requires fancy math. I mean, there's definitely certain things where fancy math tells, but not everything requires fancy math. Well, with that, Tim, I definitely want to thank you. Look forward to a better internet experience coming down the day.

And I want to thank everyone else for joining us on this episode of Bob Friday Talks and looking forward to seeing you on the next one.

Tim: Thanks, Bob. Great talking to you.