Telecommunications Satellites, eSIMs and AI – with 2degrees Stephen Kurzeja

Special thanks to organisations who support innovation and tech leadership in New Zealand by partnering with NZ Tech Podcast: One NZ HP Spark NZ 2degrees Gorilla Technology

Episode Transcript (computer-generated)

Stephen Kurzeja:
Thank you, guys. Hey, folks.

Paul Spain:
Greetings and welcome along to the New Zealand Tech Podcast. I’m your host, Paul Spain. Today we’re with Stephen Kurzeja, who is the chief technology officer at 2degrees. Welcome along, Stephen. Great to have you on the show.

Stephen Kurzeja:
Thank you for having me and pronouncing my name so well.

Paul Spain:
Well, thanks for the help in advance on that one. I do try. Sometimes I get these things wrong. Polish name, is it?

Stephen Kurzeja:
It is a polish name. Yep. Half Polish. My father from Poland, Mum from England. They migrated here in the seventies to the mighty, well, Rotorua. In fact, around the bab plenty. Then to Todang.

Paul Spain:
Oh, that’s where my family came out in a similar time zone. Time frame.

Stephen Kurzeja:
Yep. That’s my hometown. Tauranga. Yeah.

Paul Spain:
Yeah. Very good. Well, really keen to sort of dive in today. There’s lots going on. Always in the technology world. The telecommunications world is the same. And I think before we started, you were saying, well, we’re not really a telco, we’re a software company first. And telecommunications is obviously a key part of what you sell.

Paul Spain:
Have I got that roughly right? You might say it slightly differently.

Stephen Kurzeja:
Yeah, 100%. And I say that being a network engineer in my background as well. But ultimately, I think of our company as a software company with great telco assets. We’ve got a really great world class network, but how our customers consume the network is through software, through service. So a lot of our focus is actually, how do we activate that and build that software capability? So it’s been a huge focus of ours for quite a few years, and that’s how we think about ourselves. It’s not a technical capability as such. It’s actually how do we wire our organization for fast feedback and how do we deliver for our customers quickly and innovatively. Great.

Paul Spain:
Well, before we jump in, of course, big thank you to our show partners, which includes 2degrees, One NZ Spark, HP and Gorilla Technology. So we appreciate their support not only of the New Zealand tech podcast, but the broader sort of tech and innovation ecosystems here in New Zealand. So I guess it’d be interesting for listeners to know sort of, you know, what’s your background? When we were chatting before the show, you’re talking about working@max.net. 20 odd years ago. Yeah. How did you get started in this big, wide world of.

Stephen Kurzeja:
Did we start from four years old? No. Yeah. I guess my journey was, in a way, it’s kind of fate, I think, getting back into telco. Cause I didn’t go down that path initially. I just like solving hard problems. Loved technology as a young kid. Started up my own bbs, you know, I was the guy.

Paul Spain:
Bulletin board.

Stephen Kurzeja:
Bulletin boards, yeah. I was the person that was in the back of a telecom exchange trying to find spare parts and that sort of thing. I never saw a career in it and never really kind of thought about that. So what did I do instead? I went down a kind of science pathway, studied physics, mathematics. I did electronics as well, started. Did a master’s thesis in quantum physics and started a PhD in a similar thing actually in quantum teleportation. But while I was doing that, I was doing some contracting on the side and doing software development, which kind of got my taste into other areas. But weirdly enough, the work kind I was doing back all those years ago at the University of Auckland.

Stephen Kurzeja:
And New Zealand’s probably not well known in terms of what it’s done on the world stage in quantum physics, but it was kind of the foundations of a quantum Internet. So the quantum teleportation piece was kind of part of that. But I’ve left that part, although I think about it a bit, what could have been down an academic pathway? I took a different pathway and I ended up in a small company called Max Net. It was one of the kind of the range of small ISP’s that started up in the kind of mid nineties, late nineties. As the market was changing, I started as a software developer. It was a great experience just being the only software developer in a small company and you get to see everything happening, you get to see your work that matters. And in a way I’m trying to replicate that again in 2degrees. So what I’ve learned in that kind of, what was, he didn’t talk about agile and squads and all these things.

Stephen Kurzeja:
You actually just did the work in a small startup. How do you replicate that on a bigger scale? So being a software company is actually, how do you replicate that on that bigger scale? But my journey from there, I left Max net, joined core plus, and then we’ve had a lot of mergers and acquisitions along the way, eventually in the 2degrees family. So we all have our different origin stories. Being at the organization for 17 years, which seems like a long time, but it’s not really. Yeah, it changes every few years. It keeps it all very interesting. And my remit is chief technology and information officer. So it’s across software, security, data, the network, your mobile network, your fixed network, and all the other parts around technology that we don’t think about often, like your laptop, corporate it things, retail it staffing, anything to do with technology.

Stephen Kurzeja:
But what I’m really passionate about is how do we connect that technology to our customers? And that’s what I think about a lot every day.

Paul Spain:
Yeah. Oh, that’s great. Yeah, lots in there if you want.

Stephen Kurzeja:
To delve into quantum physics anytime.

Paul Spain:
Well, I was going to say, look, you talk about quantum teleportation, are we going to be able to teleport ourselves? What did you land on in those regards?

Stephen Kurzeja:
Yep, that’s a good question. And often you have an analogy drawn to Star Trek and teleportation, but the weird thing about quantum is it’s completely different. So it’s actually not really teleportation, it’s actually how do you transfer information from one quantum object to another? And that’s really hard in the quantum world, because as soon as you measure something and it changes. So the kind of traditional classical teleportation would be more like a replication of you, kind of mapping you out. Maybe that’s where the AI, digital twin parts sort of come in from, a quantum part. It’s quite interesting because you can then build, and these are real things now, you can build quantum encrypted channels that you can never break them, because if you do try to do a man in the middle, the entire state collapses. So it’s got the physical properties are built in to be secure. That’s pretty cool.

Stephen Kurzeja:
And those things are actually being used in defence areas across the world. Now, what was, for me, kind of, this is the interesting thing going back to the nineties, and quantum computers were talked about back then. AI was talked about back then. We had the AI winter for 50 years, but I’m now winding the clock forward to 2024. What I was studying there is actually now, 25 years later, sort of coming to fruition. So I find it quite fascinating, the things that sometimes humans can maybe underestimate or overestimate technology changes.

Paul Spain:
Yeah, yeah, yeah. Fascinating. Now, probably lots and lots that we. That we can delve into. It’s been between 18 months and two years since 2degrees and vocus, I guess, merged. So lots that’s been going on. What can you tell us around that journey? What are the things that have. That have gone well? What are the things that are probably maybe more complicated than you would have liked? I think we’ve all seen over the years varying things with challenges related to bringing software together.

Paul Spain:
I think we’ve certainly had one or two telcos over the years that have. That have got stuck years and years and years into having acquired a company, and they still haven’t managed to merge, merge systems. So, yeah, walk us through those aspects.

Stephen Kurzeja:
Yeah. So it has been nearly 18 months. So from what was the 1 June 2022, two organizations came together and we had to hit the ground running. And a lot of it is things that you wouldn’t even think of. Some of the basic stuff to start with. We made some really key decisions early on, but things like just how do you have a single identity on your laptop? How do we get people in the same building under the same communication platforms? And we had to do that really quickly to get our teams actually connecting and getting the culture right. So those sorts of things you don’t think about are actually really important. We actually physically relocated people in offices and got all that working quite quickly.

Stephen Kurzeja:
And then the key decisions around systems and technology stacks that were made very early. And we’ve got a. We do have, I guess, through our journey and my journey for over 17 years in the organization, we have had a lot of mergers and acquisitions. We do have a playbook and how we do it, but every time it’s different. So there are always different aspects. But probably the key theme is it’s kind of a disagree and commit piece because there’s always emotional attachment to certain things as well. So make decisions quickly, communicate the why, and then just get on with it. Get on and get it done, and do it incrementally.

Stephen Kurzeja:
So rather than what would be a traditional kind of big bang, multi year digital transformation effort, and we’ve probably seen many in the news around those that don’t work, it’s, how do you do it incrementally? So a lot of the thinking is, how do we. This is the really interesting part, and one of my hats was an architecture hat for a long time, is how do you take legacy and modernize it? Where do you look at the fracture plans in technology to do that, to glue bits together? How do you do transitional architectures? So that’s the technical part. We’ve got some really smart people in the team that have worked through that to make these things more seamless. So we are 18 months through. We’ve already had some good wins. Part of it is also belief. So we’re building a lot of our platforms internally. I think our competitive advantage is actually, we do have a lot of in house capability.

Stephen Kurzeja:
And my team, I’ve got around 450 technologists in some way, over 100 of them are software developers. So we’ve got quite a sizable software development capability here in New Zealand. But in saying that also it’s democratized there are a lot more people that do software as well, so we try and democratize it. So yeah, a really great team that are building this capability. So we’ve had some, yeah, early on it was around how do we create belief that we can do this in house for a lot of things? And we had some wins early on and deliverables that we delivered as we said we would and the belief grew, but we still got a lot more work to do. So a big part of it is how do we unify these systems so we can have a great customer experience. But we still have two main stacks that our customers sit on and we’re quite close to actually getting these together right now. So we’ve got probably still another six to twelve months of really hard work.

Stephen Kurzeja:
And it’s not all glamorous stuff. There’s certain people that like legacy and working through that, but we’re doing in a way that is always customer first. So how do we create a seamless customer experience? When you’re migrating from one system to another, we’ve got to unify the customer experience layer and then can we add more onto it? So while we’re trying to maybe do a like for like in terms of customer experience, can we actually improve it along the way is what we’re always looking at in terms of opportunities. So yeah, a lot of work that’s been happening and alongside of that is things that you don’t talk about as well. Data modernizing, modernization. When you’re moving systems, you need to modernize your data warehousing and all that stuff. Cybersecurity is a huge one as well. While you’re doing this work, it needs to be secure.

Stephen Kurzeja:
Threats are always persistent and growing. So there’s a lot of aspects that we’re trying to do and that’s the integration piece. But we’re also growing. So we’ve been really growing in the market across the last 18 months as well, in all segments. We are in the consumer segment, mass market in the business segment for corporate governments, medium sized enterprise small, which I think often New Zealand may not know that segment that we plan because we are serious about business and also wholesale. So those are our key areas and we do work in all of those spaces. And one of my key passion areas, and I work a lot alongside Andrew Fairgray, who’s our chief business officer, is how do we grow in that business segment? I just see so much opportunity in terms of productivity for New Zealand that we can do. So that’s a big area that we’ve been focused on what can we do as a telco to help lift up productivity for New Zealand and help make small businesses more productive ultimately and just make telco an easy thing to consume.

Paul Spain:
Yep. Yeah, I think it’s interesting because. Yeah, 2degrees. Yeah, I think a lot of people would have probably think of you as more that sort of consumer focus. So, yeah, it’s interesting to hear and I guess for a lot of the vocus business was business enterprise and I guess some government sort of customers and so on in there. Right, yeah.

Stephen Kurzeja:
And that’s growing and the network itself is considered a critical national significant asset. If the network’s having a bad day, New Zealand’s having a bad day. So we really focused on the reliability of our network. We’ve got all sectors, local councils, different industries, emergency services, so reliability is hugely important. Having the uptime is significantly important for us. So we invest a huge amount in reliability and diversity of the network, but that is what we consider bread and butter. But a lot, it’s a bit of a secret in New Zealand. I think that we don’t, a lot of people don’t know that we have this network that we own over 4000 fiber and the mobile network as well attached to that.

Stephen Kurzeja:
So I guess our challenge is how do we change that perception a bit and activate that more? Because once we start talking to people and customers, they actually understand it, they love us. Yeah. That’s our challenge going forward, I think, is we have the telco assets, we have the software capability. How can we actually just change perception a little bit in the market?

Paul Spain:
Yeah, yeah. So how would you use the fiber today that, you know, that 2degrees owns. What are the. Yeah, so it’s because I think for most of us, if we think it’s sort of home or business Internet connection.

Stephen Kurzeja:
Yes. This is a little bit different. Yeah.

Paul Spain:
You know, then that’s got its particular role, but there’s also the importance of, you know, connecting assets and the country together. So, you know, the backbone is really important.

Stephen Kurzeja:
Yeah. So the backbone is exactly what it is, it’s connecting cities together. So that’s the, that’s the fiber network that spans from, you know, top of New Zealand down to the bottom. And there are probably, I guess you’d say four major national backbone fiber networks in New Zealand today. This one is one of them. And interestingly a part of our story, I think some people know the 2degrees effect and what happened overnight when 2degrees launched mobile and overnight it created a huge economic benefit for New Zealand. Quoted by the Reserve Bank $15 billion of economic value. Our heritage and fighting for fear has been across all these different origin stories.

Stephen Kurzeja:
We had slingshot that kind of created an ISP overnight and was fighting David and Goliath, but the fiber network has a similar story. So FX networks, which is probably a name that maybe not a lot of people have heard of, they actually was one of our origin companies, built the fiber network in a time when it was dominated by incumbents and that’s a huge investment to do and it requires a lot of backing and leadership to do it. So, yeah, FX networks actually changed the game and that backbone backhaul environment at the same time and that was before chorus existed in terms of that separation. So just another part of our origin story and how we’re trying to drive competition. Yeah, and there’s many examples like that. So that’s one that’s often forgotten about, but that’s part of our heritage. And yeah, the network itself is used to just transfer huge amounts of data. A lot of it still is kind of Auckland origin, but it’s, you know, it’s bandwidth.

Stephen Kurzeja:
We have optical networks that sit over the top that are running in the terabytes per second on a single fiber. It’s all great, fascinating stuff, you know, high speed lasers and electronics and it’s also secure. So we have these fibers that are highly secure. So we have customers that buy direct fiber because they want direct fiber to a data center and another data center. So the security aspect is really important and other different use cases on the fiber network, but predominantly it’s to connect cities together. Yeah. So, yeah, fascinating stuff. And connecting mobile towers as well onto the fiber network is another key aspect.

Stephen Kurzeja:
So it’s not just your local fibre companies. This fiber asset is used for a range of things like that.

Paul Spain:
Yeah. Now with varying weather events and so on, we have found ourselves with mobile towers that were not connected going back. Was it a year ago now? Two years?

Stephen Kurzeja:
Cyclone Gabriel. Yeah, was just over a year ago.

Paul Spain:
So yeah, looking back there, what have been the lessons and sort of changes from there? And of course part of this crosses into a time where access to satellite connectivity has much improved over the last couple of years. So how has that changed for your network to provide increased resiliency going forwards?

Stephen Kurzeja:
Yeah, the event was, I guess, something we’d never seen before because the extent of the damage wasn’t localized. So we’ve had other events that have been more localised. Obviously we had three national emergencies, but the other two, Christchurch and COVID, was another one. It was a different sort of issue, but this was really spread. So it did spread our field services quite a lot as well. The people on the ground that were doing the hard work. From the backhaul perspective, that was a quick turnaround on how do we use Starlink business grade backhaul on a mobile cell tower. And that was something we had already been testing in the lab and we rolled it out relatively quickly.

Stephen Kurzeja:
The team did an amazing job to get that out. But, yeah, many factors around access and discussions around the battery life in a cell tower generator use. How do you get fuel into a site? It’s all around access. How do you get to basic needs? So, yeah, there’s been a lot more discussion around that and also with our, particularly our bigger corporate customers, they’re very interested in getting resilience for their own services. So our Starlink business broadband package has been working really well for our business customers. Well, that want to, they might have a primary fiber or even a secondary fixed wireless access. And then a third option with Starlink. There’s been other discussions around community hubs.

Stephen Kurzeja:
We’ve done work with network for learning recently, connecting schools with Starlink. And those are examples where you could have a hub where people can get to that still has a connectivity pathway. The great thing about low Earth orbit satellite is it goes up and it comes back down to a different ground station, a different area. So it is truly diverse. From that aspect, you just need power. So having a community hub aspect is something we’re working on with a few places as well. But, yeah, it’s an evolving thing around resilience. And that discussion that we’re actively working with our industry partners on as well.

Stephen Kurzeja:
Yep, yep. Yeah.

Paul Spain:
That’s a bit of a journey to a point where you’d be able to say, hey, we’ve got satellite backhaul options on a very high percentage of remote sort of cell sites. It’s not something that takes.

Stephen Kurzeja:
Yeah, permanently is a bit challenging, but having it in a deployable states is what we’re doing, so they can be deployed quickly. So, yeah, part of it is how do we respond quickly? And then your other aspect is the direct to cell technology that is evolving as well. In a state of emergency, just making a phone call or a text message was really important when Gisborne and Hawkes Bay were, in some cases, some parts were cut off. Just being able to contact a family member. I had family members in the region, how do we contact them? So having that instant emergency response is really critical. And that’s where the Leo direct to satellite cell tower in the sky, I think will play a big part going forward as well. So kind of complement the resilience discussion.

Paul Spain:
What does that journey look like? You’re working with Link, you know, I guess they’re sort of a smaller firm if you put them next to, say, SpaceX, right?

Stephen Kurzeja:
Yeah, the gorilla.

Paul Spain:
So you’ve got SpaceX with all their own launch capabilities and Starlink, which we’ve mentioned that they’re able to kind of, I guess, started tagging onto some of those Starlink satellites, some of that direct to cell type of technology. How do you see that sort of journey playing out? Where will we be in twelve months, 24, 36? Is it a decade away until we’ve got complete everything on our mobiles when we were in those distant rural locations.

Stephen Kurzeja:
How do you see that? Yeah, like I said at the beginning in quantum physics, hard to predict the future, isn’t it? Sometimes, but yeah, I guess complete accuracy.

Paul Spain:
No, no, just your kind of thoughts on how it plays out. And we know that look, you can’t control what other firms do and so on, but what would you sort of expect it to look like?

Stephen Kurzeja:
Yeah, I mean, the first thing is space is really hard. So, yeah, like you say, there’s all the caveats in there, but it is evolving. So, yeah, our partnership with Link, they’ve got five satellites up now. Our CEO Mark Callender actually went over to see a space launch two months ago. I didn’t get invited.

Paul Spain:
I was a little bit jealous on that one.

Stephen Kurzeja:
Me too. Me too.

Paul Spain:
Someone invite me?

Stephen Kurzeja:
Yeah. No. So they are launching. They’re launching with other launch partners and it is improving. So we took quite a pragmatic approach to it. We’re not going out bullishly saying this is going to solve everything. So we know space is hard and this stuff is hard and you’re not going to get necessarily in building coverage. It’s going to be certain times of day when it’s happening.

Stephen Kurzeja:
We did our first text message and voice call last year, which is really exciting stuff. Who would have ever thought technology could get to this point? When you’ve got satellites whizzing around 500 km in the air, it’s fantastic. But I still think you’re in the horizon of probably two years to have something I think really compelling. And you’ve got other players in the market as well. In the market and trying to in this race, Amazon, AST, Space mobile. So there’s a range of activity going on. There’s a lot of funding from different areas. So I don’t think it’ll be a one horse race necessarily, with SpaceX and Starlink.

Stephen Kurzeja:
They do have a lot of launch capability, don’t get me wrong. And they’re getting their new big rocket ready to go where they can launch a huge amount of satellites. But it still will be an ecosystem, I think. Yeah, we’re really excited to be part of it. Link have been a great partner for us because it’s just been quite an easy way to work with them and just get some learnings and actually see it working. We did some more trials down in Nelson, so it’s, yeah, great for emergency use cases right now. So it’s not a, you know, you don’t have 100% coverage everywhere, but it’s going to evolve and we’ll just be keeping updated as it progresses. We’re not making bold claims around this.

Stephen Kurzeja:
It’s just, you know, progressively as it improves, we’ll just be communicating with our customers and the market around what’s happening. Yeah, but, yeah, fascinating. There’s also many fascinating aspects, one that we don’t talk about a lot, which is not the direct to sell, but it’s the satellite to satellite space lasers. So creating new low latency pathways which were traditionally tied to or tethered to fiber optic subsea builds. Imagine like a case of New Zealand and South America being communicating a low pathway. Those are interesting, and that’s probably going to open up as well. But that’s tied to some regulation aspects as well because you have to have your connection coming back down the same geographic area at the moment. So.

Stephen Kurzeja:
But that could open up whole different use cases. Bandwidth might be still lower than a terrestrial subsea, but you’ve got this low latency aspect. I know the gamers would love that, for sure. Financial institutions, I’m sure there’s other parts.

Paul Spain:
Yeah. So rather than a subsea, you know, subsea fiber optic cable from here to, let’s just say, chile or wherever. Yeah.

Stephen Kurzeja:
You could basically have greater dynamic satellite.

Paul Spain:
Connection then laser from one satellite to another. Or would they go through sort of multiple.

Stephen Kurzeja:
They go through multiple. Multiple satellites.

Paul Spain:
So you’d be able to do those kind of hops with laser.

Stephen Kurzeja:
Again, bleeding edge technology, but it’s all possible.

Paul Spain:
And then back down.

Stephen Kurzeja:
Yeah.

Paul Spain:
So.

Stephen Kurzeja:
And you know. Yeah. Speed of light, it’s faster in a vacuum than it is in a cable too, so it’s actually faster.

Paul Spain:
Yeah, I was just kind of curious about that. So doing it that way, going up to space and back down, you potentially would have a quicker communication, lower latency than going via an undersea fiber optic cable.

Stephen Kurzeja:
You would. In theory, yeah, yeah. Okay. In theory, bandwidth would be the biggest challenge initially because it would be how do you line these things up to actually communicate with space lasers and do it reliably? But that’s what I see as definitely something that’s. I’ve been fascinated about the subject for many years. Like ten years ago. Like this is a potential. Leo’s coming and it’s here now.

Stephen Kurzeja:
The future is not far away.

Paul Spain:
Yep. Yeah, we had spoken with him, George Lee, who Kiwi guy who ended up getting. He was at NASA in the US and I think he’s ended up probably in Australia, but he was working on, I think, the space lasers and sort of communication from satellite to satellite and so on. So, yeah, quite an interesting field. It might be. It’s time for us to catch up with him and see where it’s at.

Stephen Kurzeja:
He can tell you what’s really going.

Paul Spain:
On, where some of those things are going. So you think there is the potential to have enough bandwidth to make that sort of connectivity something that would be. That it actually be able to take more than a fraction of a percent of communication between, say, New Zealand and South America?

Stephen Kurzeja:
Yeah, I think so. It’s going to be another avenue. And latency is a critical thing for certain applications. So everyone wants to. You’ve seen probably the financial institutions that are running direct fibers to do trading. So there’s many different scenarios where I think latency would be more important than actually bandwidth capacity. So, yeah, another avenue to consider in this whole Leo dynamic that’s happening direct to sell. We’ve already got the Starlink broadband services that are really great in New Zealand because we have a really good position in the world to actually get line of sight to satellites.

Stephen Kurzeja:
So that’s really going well. We connected the Chatham Islands probably, we’re going back 18 months ago actually, with Starlink. So these are making tangible differences to digital equity already and we kind of see it as complementary to what we’re doing. You’ve got UFB, you’ve got fixed wireless access, you got Leos all coming in just to have good choices for New Zealanders. So yeah, yeah, exciting times in the telco industry.

Paul Spain:
Yeah, yeah, yeah. Now esims onto another, another topic. So this has sort of been, I guess, you know, progression for mobile sort of carriers around the world. And I guess smartphone makers have moved to esims being quite an important element. And in fact, is it the US where now when you buy an iPhone that they only take eSIM? So I could be jumping to a conclusion there, but I’m pretty sure there’s.

Stephen Kurzeja:
That’s. We’ll fact check that.

Paul Spain:
That has started to happen. But here in New Zealand, you buy an iPhone, but one physical sim in and one eSIM. So things seem to have moved in that direction. Which, yeah, is interesting, can be very, very helpful from a flexibility perspective. You even see when people are traveling, they can go and buy an eSIM effectively for any. I don’t know about any country, but you can go and line those things up before you get somewhere if you’re trying to sort of save on roaming costs and what have you. But there’s that aspect of the wearables and being able to have sort of one number across a smartwatch and a phone where’s 2degrees on that front. Seems these things are a little bit more complicated than just ticking a couple of boxes on screen and turning on all the possibilities.

Stephen Kurzeja:
It would be nice if it’s the case. Yeah. So we launched eSIM back in August 2022. I’d say we were a little bit behind. We’re highly self critical of ourselves. We always want to do better. And, yeah, we’ve been working hard to get the next part of that, which is the wearables piece. It should be only in a couple of months.

Stephen Kurzeja:
In July, we’ll have a wearables component with Samsung and hopefully we’re working hard just to get the rest of it done. So I can’t mention timeframes on the other parts, but all the other kind of basic mechanics of things that you don’t think about. I originally thought sims are like this 1980s technology that you’re putting in a smartphone. It’s like putting a cassette tape in a smartphone. Right. It’s like, why would you do that? But going to eSIM actually opens up all these other aspects that you didn’t think about. How do you transfer an eSIM from one phone to another when you’re doing a phone swap? That was easy with a physical one because you just pull it out, plug it in. How does that journey work, converting from physical to eSIM? So we’ve done all of that work.

Stephen Kurzeja:
So you can. We have eSIM capability. You can go on an iPhone and go from a physical to an eSIM. We’ve got a whole lot of sign up journeys. The wearables part is something that we’re just on the final parts of getting sorted, which. It’ll be really exciting. I see you’ve got your watch on your cell phone watch lists at the moment. Yeah.

Stephen Kurzeja:
So imminent is what I would say. And, yeah, I’m really keen to see it in action myself. Yeah, it’s great freedom to have your wearable independent from your phone. So, yeah, coming really soon.

Paul Spain:
Yep, yep. It’s good to hear. Now, artificial intelligence, especially sort of generative AI, has really been, you know, almost always we seem to talk about when it comes to tech over the last 18 months.

Stephen Kurzeja:
Surprised we haven’t been talking about it yet.

Paul Spain:
18 months, you know, where does it fit in for you and for. For 2degrees at this point?

Stephen Kurzeja:
Yeah. Are we deep fake? I don’t know. Are we actually real people if we’re people watching us on the podcast, or are they real voices? We’re on a journey, on a learning curve. Yeah. So the excitement, obviously, with chat, GPT launching, it was everyone’s. It’s become the most accessible thing. Right. It’s been, from a kind of worldwide perspective, just this easy way to get into AI.

Stephen Kurzeja:
We’ve been doing AI generally for a little while, machine learning, churn, prediction models, all of these things, probably for a bit of over a decade in some ways, but taking a fresh perspective with generative AI, we’ve been doing a few things, but taking it as a way, it’s not an AI strategy. We’re not doing it for the sake of technology. It’s actually, how do we improve and execute on our strategy quicker as a turbocharger? How profound is it? Is it all hype? Is it actually something real? I’m on the side of an optimist and I think there’s quite a profound change happening. But there is definitely a huge amount of hype in the market. I talk to my peers, other CIO’s, ctos, and there is a bit of this fomo going on as well. It’s like, what are you doing? What are you doing? Because they hear things, because you put your best Instagram post up. This is what we’re doing. So we’re taking, again, quite a pragmatic approach.

Stephen Kurzeja:
It’s been, in a way to bring people on the journey, so there’s a whole bunch of enthusiasm. So we’re using the simple things, which I think will be table stakes around employee productivity. We’re utilizing Microsoft copilot in a small group, doing things like team summarization, all of those good things around, generating actions out of meetings.

Paul Spain:
How do you find those things? Because I guess experiences I’ve had with some of these tools is the hype is often way better than the actual sizzle isn’t sometimes there’s not much of a real kind of result that comes out. And you mentioned teams meeting, summarization. I’ve seen things there where whole segments of a meeting, whether it’s a particular word or subject or you know, someone’s voice or accent and so on, can just, you know, if you’re completely relying on that kind of. Yeah. You’re gonna, you’re gonna be in a big hole if you haven’t actually been, you know, taking the human kind of notes as a, as it were. Right.

Stephen Kurzeja:
Yep.

Paul Spain:
Yeah.

Stephen Kurzeja:
We’ve done measuring, sort of trying to measure it quantitatively and qualitatively with our groups of users, but overall I think we had around 70% said they improve quality. We’ve just on those kind of what I call table stakes, things that’ll be around employee productivity. The top users kind of pragmatically saved around 10 hours per month of. Yeah, but you know, you’re improving a five that up.

Paul Spain:
That’s over two weeks a year, right?

Stephen Kurzeja:
It is, yeah. Yeah. What do you feel?

Paul Spain:
It doesn’t work quite like that usually, does it?

Stephen Kurzeja:
What do you fill that with? Yeah, so you can improve a five minute task or a 1 hour task and get a better quality outcome. There’s a great study by BCGX around this jagged frontier and it was about cyborgs and centaurs around how to use AI. There’s different people that use it in different ways. One more as a copilot is co intelligence. Like a coach actually. Give me this information and then iterate, iterate, iterate. And there’s others that use it just as a machine and go do this for me. Bring it back.

Stephen Kurzeja:
So great research if anyone wants to look at it. There’s a really good book called co intelligence as well, which goes through all these different aspects. I’d highly recommend it as well. So there’s that part of it which we’re kind of growing and just learning, testing and learning because I don’t think anyone really knows. But Microsoft’s products are getting better and better as we move forward. Things are getting iterating and interesting cases of things that you may not expect as well where you have the sentiment of a teams meeting where employees are talking to each other and goes, this person doesn’t like this person. So the teams is wrapping up some sort of sentiment in there as I. Interesting, but the other area we’re using it is in our call center.

Stephen Kurzeja:
So we’ve got, I’d say it’s still a trial phase, but it’s progressively moved into production around when an agent has a call. We have an auto summary that’s generated by an AI model. So it’s effectively similar to teams meeting concepts where it’s generating the transcription, but then it’s also analyzing that and pulling out the sentiment and intent of the call. And then we feed that back into our kind of our data models and can get some learnings out of it so we can learn. Okay, customers are calling in for this reason. Should we add more automation or more self service, or should we be training our call center staff in a different way? So that’s been, I think, quite a massive improvement for us because traditionally, sentiment intent’s been really hard to analyze. Yes. So that’s been really successful.

Stephen Kurzeja:
It does mean you can wrap up calls quicker as well because there’s a lot of time spent actually wrapping up calls and typing out notes. And yeah, our frontline staff love it as well because it makes their lives easier.

Paul Spain:
How does it help your frontline staff?

Stephen Kurzeja:
Just removing the laborious work of typing out notes in the wrap up so they get a head start on it. It’s all transcribed, it gives them some indications. Gotcha. There’s this thing called next best action and next best experience as well, so you can kind of evolve it to the next piece. Could you have an agent with a copilot that could be kind of analyzing what’s happening and actually give some coaching to the human and go, okay, you should be thinking about this, talking about this. So that’s kind of interesting scenarios moving forward.

Paul Spain:
Yeah, it does seem like there’s a lot of opportunity when it comes to call center and customer service type of.

Stephen Kurzeja:
Yeah, a huge amount. We’re on a journey on that, but taking people on the ride as well, because there’s so many ethical aspects to it. There’s the bias, hallucinations. We’ve got a, you know, pretty early on, we put a responsible AI use policy in place. We’ve got a data AI governance forum. We, we talk about this stuff because it is, there is definitely an ethical part, there’s a safety aspect as well. And the other area, which I think is probably the more profound because it is around software development, productivity. So we have, as I said, we consider ourselves a software company with great telco assets.

Stephen Kurzeja:
So our software developers are using a copilot to help write their code maybe sounds pretty boring on the surface of it, but their productivity uplift has been significant. And if you consider, I think, what’s holding back not just 2degrees, but any organization in the world is, how do you develop software more productively in this age? Of digital that we’ve been in from the 1970s to now, it’s all been around software production. That’s ultimately what we’ve been producing. And software, in a way, is a negative scale industry. I started off at Maxent as one developer, and that was really productive. When you get to five developers or 100 developers like we have, how do you scale that up as something we think about a lot, and it is negative scale. The more people you add, it does get harder. Communication channels get harder.

Stephen Kurzeja:
So I think probably one of the most profound things about AI is actually how does it change and lower, lower the cost? I just mean, in terms of, how do you develop software better, quicker with a copilot will be fundamental because that’ll drive the next layers of innovation. So that’s exciting. And we’re seeing about a 20% uplift in productivity and quality. So it’s better code, better peer reviews, all these things you don’t think about as a software company that you need to do. It’s not just magic that happens. So, yeah, I think that’s quite a fundamental one for us, and that’s going to keep improving these large language models that we’ve only on the infancy. I think you’ve got. I saw in, I think it was Twitter around the delivery that Jensen Huang gave to Sam Altman, the new Blackwell chipset for the training of the next version of maybe GPT five.

Stephen Kurzeja:
Who knows?

Paul Spain:
Yeah.

Stephen Kurzeja:
Yeah. So we’ve got Moore’s law, which we had in the past, which was doubling microprocessors every two years, is the Jensen Huang law, I think, which is we’re seeing a significant increase in compute power, an entire change in the industry from what was traditional compute to this accelerated compute, which drives other aspects. You know, how do we power it? Data center capacity, all of these things. So I think it’s just at the very beginning of these large models, and then you think about language and what it is. Can you train these models on other things, like cellular language? What does that mean for, you know, in the biotech field and health science? So there’s language and text. There’s other things that these models can be pre trained on. So there’s a lot of areas. I think that’s why I’m more on the optimist side.

Stephen Kurzeja:
There is obviously all the caveats around. Is it safe to use? What happens if you get an unsecured AI model? And can you get to artificial general intelligence when the AI is actually thinking for itself?

Paul Spain:
What’s your pick on AGR, general intelligence?

Stephen Kurzeja:
Yeah, when is it going to happen?

Paul Spain:
Are you someone that sort of thinks it’s a long way or a never or this is sort of something we.

Stephen Kurzeja:
Might see, depending on the definition of it. I listened to a podcast that Sam Altman was on about a month ago. His definition is kind of a median intelligent kind of person. He wasn’t talking really aspirational Agi. And what we’re playing with the models now kind of feels like someone’s thinking it feels a little bit magic, but it’s not magic, it’s just predicting text.

Paul Spain:
That’s right.

Stephen Kurzeja:
Yeah. So I don’t know. The predictions I’ve seen vary. Like there’s five years to ten years, but I couldn’t. I’d say maybe it’s in that realm, but you just don’t know. And again, it’s in its infancy. Like the amount of investment we talk about telecommunications capital investment, the amount of investment going into these AI factories, data centers is going to surpass telco investment. So they’re talking $7 trillion of investment.

Stephen Kurzeja:
There’s an arms race going on, so it’s going to be fascinating and possibly five to ten years. That can be. Sounds a bit scary. What happens if you’re the optimist or the other side? Is it Skynet taking over or is it going to benefit the human race? I’m on the benefit side, and as a co pilot, definitely. But there’s so many facets to consider in this space and, yeah, we do still need regulation and all these things to play a part. So it’s a huge component is different to previous technology where this can potentially go out of control.

Paul Spain:
Yeah. Yeah, yeah, yeah. I guess. Yeah. My. My leaning on AGI is it’s. It’s maybe not. Not anytime soon, but I could be completely wrong.

Paul Spain:
Right. It’s like, that’s. That’s not my era of specialty, but it just seems to create that sort of intelligence seems sort of too mind blowing. But then again, with just how good in some areas that certain aspects of AI is today is pretty surprising. And if you’d asked me ten years ago, would I have been able to get my head around what we can do now? And, yeah, maybe not sure. Some of these things, you can definitely join up the dots and kind of see. See how they will happen and so on a reasonable number of years in advance. But, yeah, some things end up being much, much harder.

Stephen Kurzeja:
Yeah. There’s a great report that Amy Webb, she’s a futurist that she publishes every year that tracks trends, and I definitely recommend it it tracks AI trends, all sorts of trends. And she says that we actually are in a technology supercycle at the moment. There’s a combination of AI, biotech, and connected things all coming together at this moment in time. And she makes certain predictions about what that means in the near term. So it’s definitely one I think the viewers could check out. It’s really good material. Thanks for.

Paul Spain:
Yeah, we’ll put those into the show notes. So people that are wondering afterwards if you haven’t been able to write that down now. I’m just thinking what we can squeeze in before we finish up. What’s on your mind that we haven’t covered that maybe you think listeners might find interesting?

Stephen Kurzeja:
Yeah, I think we’re doing a lot of work to create some amazing customer experiences with our integration onto one, onto one stack. But the key areas, what are your.

Paul Spain:
Technologies that you sort of built that on? For those that are interested, we use.

Stephen Kurzeja:
A lot of open source. In fact, I’ve always been on the curve of trying to do things a little bit different with innovation. So we’re using technology that was actually born out of Uber. When you book your Uber app and the workflow engine behind the scenes, the orchestration engine, we’ve actually utilized that bit of technology to orchestrate our network. So when a customer can go into, we’ve got a portal called flex where a customer can log in, particularly for medium sized enterprises, they can purchase a circuit and a service and a network as a service model. So not traditional telco. The orchestration technology behind the scenes to go and provision a piece of equipment in our network is all done through this engine that Uber invented, which I think is pretty cool stuff. And for those that are techno inclined, we use a lot of go Lang, which came out of Google programming language.net, a range of other SAS platforms as well, around the edges, and a whole range of different databases, document storage, traditional relational databases, a range of other things.

Stephen Kurzeja:
And it’s kind of an interesting one when you’re in technology, you’re trying to create guardrails around it and standardize, but you actually want to give freedom to your developers to make decisions as well, but with enough guardrails in place not to have a kind of a spaghetti mess. So I think we’ve struck the right balance there. So we’re giving. Our goal is to have teams that are autonomous, self organizing. They can make most of their own decisions within the guardrails of our architectural roadmap, which means we can deliver fast and we can do innovative things. A lot of our innovations actually came out of inspired from a developer’s idea. We run hackathons, we have innovation days. One of our big ones was a thing we call the Tahi app, which is an internal portal for our colleagues and our frontline staff to actually interact with our systems to serve our customers really well.

Stephen Kurzeja:
Telcos, and probably many in the industry, focus a lot more on the customer self service, but they forget their dear colleagues that are running on like seven different internal portals. You invest all this energy and customer experience at this end, and what matters is that actually your digital experience for your colleagues and your customers. So that was one innovation that came out of an innovation day that we just said, this is great, let’s fund it, let’s build on this. We just want ideas to come from anywhere. So it kind of touches on the team culture, piece around innovation. A culture of psychological safety and learning is what we strive to have, but the technology behind the scenes reflects that. There’s a lot of open source technology, a group of different things that we use. So keeps it interesting, keeps it fun.

Paul Spain:
Yeah, yeah, that’s good. And what’s your approach to cyber security? Is there kind of a. A simple way you can boil down how you keep things in check on that front?

Stephen Kurzeja:
It’s a journey. So, yeah, it never stops. We keep investing in it. We’ve got a 24/7 security operations center that we run both on our network and our security we connected in with all the industry. We kind of lead in some areas as well. So we’re scam frauds and those sorts of things that we see our own internal security posture. It is a journey. So we’ve got your traditional stuff that you’re protecting your perimeter, your internal systems, removing legacy vulnerability management.

Stephen Kurzeja:
Yeah, I’d say my simple version is it’s a journey, you just need to keep investing in it. And we do consider we’ve got certain regulations and accreditations we have to adhere to. We serve government customers, we’re on certain panels we do have to ensure we’re secure, but we’d happily talk to anyone around what we’re doing there because we like to work out loud and share stuff like what is good practice for us, what’s endpoint protection mean and how does that work? We offer managed services as well, for firewall as a service and a range of other things. So happily, can I talk to anyone around that? And we do sponsor a little hacker, not hackathon, a hacker conference down in Christchurch called Chitchcon. So if anyone’s down in Christchurch that do go to that. They’ll see my head of security there, Ivan, every year and some other events. We sponsor the OWAsP event as well for developer security. So a lot of it’s actually how do you shift security left and have everyone think about it? Because everyone needs to be concerned about it.

Stephen Kurzeja:
It’s not just one security team. So our developers are thinking about how do you build things securely? Our frontline agents thinking about data and how do we, our customer data is like, it’s critical asset that we can’t expose. So ensuring that’s always protected and people are thinking about it and that way invest a lot in awareness and training. So as a. Yeah, full. And our board are really interested in it, obviously that’s good. There’s been a lot of high profile cases in Australia, like Optus, for example, and a range of other things, so. And, yeah, we got customers that we want to ensure we can work with as well to help protect them.

Stephen Kurzeja:
So, yeah, we’re very engaged in it. That wasn’t a short answer, Paul. Sorry.

Paul Spain:
No, that’s good. Oh, it’s been really good, Stephen, to sort of, you know, delve in and to look inside 2degrees a little bit from your perspective, wearing that sort of chief technology officer and chief information officer, those two hats. Yeah. Really, really insightful. So, yeah, thanks so much for joining us on the New Zealand tech podcast.

Stephen Kurzeja:
Yeah, thanks for having me, Paul, appreciate it.

Paul Spain:
Good stuff. Well, thanks, everyone, for listening in and of course, thank you to our show partners as well. 2degrees, One NZ, Spark, HP and Gorilla Technology. If you’ve been watching the the live stream which is available across LinkedIn, Facebook X and of course, YouTube, then do make sure you subscribe or follow us through your favorite sort of audio app, whether it’s Spotify or Apple podcasts. And if you’ve been listening to the audio, then make sure you’re following us on those platforms, YouTube X, Facebook and for LinkedIn, that streams out from my profile so you can follow me, Paul Spain on LinkedIn. Thanks, everyone. We’ll catch you next week.

Paul Spain:
Thanks again, Stephen.

Stephen Kurzeja:
Cheers, mate.