Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
AI is coming: Data, networks, and the next telecom shift
1 Plays2 seconds ago
The telecom world is being reshaped by new business models, the ongoing debate around 5G’s success, and the rising strategic value of data. In this episode, I am joined by Miguel Carames, Chief Product Officer at Mobileum to explore how AI is transforming networks and enterprise services, from the drive towards autonomous networks to new AI agents for small enterprises.
AI is coming for telcos and in this brave new world data may be one of telecom’s biggest untapped opportunities.
Recommended
Transcript
Intro
Introduction to Tech Drift Podcast
00:00:05
Pablo Tomasi
So welcome back to the Telco Drift, a podcast exploring tech telecoms and other sci-fi tales. Now, today's a great episode as I'm joined by Miguel Carames, Chief Product Officer at Mobileum So Miguel, welcome to Drift.
00:00:20
Miguel Carames
Thank you. Thanks for having me, Pablo.
00:00:23
Pablo Tomasi
So as always, once we start an episode, we'd like to start knowing a little bit more
Miguel Carames' Telecom Journey
00:00:28
Pablo Tomasi
about our guests. So Miguel, can you just share a little bit about sort of your journey within the Tech & Tech world and also maybe introduce a little bit about the company as well?
00:00:37
Miguel Carames
Sure. Yeah, so you know my name is Miguel Carames. I've been in telecom 26 years, right? So actually, I went to school to do this. So if you add their college years, much longer than this.
00:00:52
Miguel Carames
um Started the my career in back in 2000, spent nine years there, Started in Madrid, but I was sent to Chicago on assignment a few weeks later and then pretty much spent my entire career in in Chicago.
00:01:07
Miguel Carames
I worked on 3G R&D, WiMAX, then the beginning of 4G, LTE. And then in 2009, I moved to Verizon and I spent 13 years at Verizon. Different roles from individual contributor building the roaming portfolio,
00:01:25
Miguel Carames
um all the way through, you know growing through the ranks up to executive director in network planning by the time I left in 2022. And then here at Mobileum I've been here for three and a half years.
00:01:40
Miguel Carames
um The first... six, seven months, I was running the roaming and security business unit. And then three years ago, i got the chief product officer role. So I'm in charge of our product strategy, ah capital planning, P&L ownership for the the entire business and such.
00:02:01
Miguel Carames
And, you know, as far as Mobileum we are a ah Highly diversified company. We're about 1,700 people all over the world. I don't have any of my direct reports in the same continent. So, you know, it goes to show that it is the new normal. um We are a software vendor in the telecom space. We mostly build sovereign based solutions for MNOs.
00:02:27
Miguel Carames
But increasingly, there are also opportunities in the near adjacent areas like enterprises, MVNOs.
00:02:27
Pablo Tomasi
Mm-hmm.
00:02:35
Miguel Carames
MVNOs are a growing area of business for us. Global IoT players, car OEMs, and government and regulators as well. And so, you know, Mobileum grew through a lot of inorganic acquisitions about five years ago. So we have five distinct business units, which I'm i'm sure we'll touch during the conversation at some point.
00:02:58
Pablo Tomasi
Yeah, absolutely. and And there are a few things that I want to to touch on based on what you said. And of of course, it great to to hear that enterprise is sort of an opportunity is emerging. But if if I take like a step back, so you work both on the vendor side and on the telco side. What would you say is the biggest difference that that you see between sort of, you know, building the the product and the solution or maybe more on the commercializing side of the things?
MNO vs Vendor Perspectives: Integration & Commercialization
00:03:24
Miguel Carames
Yeah, it's a really good question. And so they they um the roles are very different. um In a way, when you are at the MNO side, you are more of an integrator, right? So you you have to make buy versus build versus partner type decisions.
00:03:44
Miguel Carames
ah More often than not, you you buy, right?
00:03:47
Pablo Tomasi
so
00:03:48
Miguel Carames
Whether that's, you know, like core network functions or your RAN partners or, verticalized application providers and so on. And you you have a business goal in mind or a set of solutions and products that you want to offer your own customers. And again, that can be consumer, but can be also small and medium business or large enterprises.
00:04:11
Miguel Carames
and yeah And you try to choose best in class, partners for each of those building blocks that you need, and then integrate them in hopefully something that's value generating, differentiating for the company, and also sticky, right? Obviously, one of the the big priorities where are the MNO is minimize churn at all costs, right? it's ah Losing a customer and regaining the customer is very expensive. Keeping a customer is much better.
00:04:40
Pablo Tomasi
Yeah, that's a great point because it's ah it's always the case of like, I just want kind of lend the customer, you know, because because that that's the big effort, you know. Once I'm there, then even from a customer perspective, it's always harder to leave
00:04:55
Pablo Tomasi
Even if you're really unhappy with the service, it's it's much easier you know to to find a way like, the yeah, I'm going just give it another try. compared to you know starting everything anew. And since since you're covering sort of a lot of different things within the telecom space, first question I would like to ask you is is really setting the scene and trying to understand, OK, you guys are doing lots of different
AI Trends in Telecom
00:05:18
Pablo Tomasi
things. You're targeting a lot of different markets.
00:05:21
Pablo Tomasi
If we look at the trends that are shaping the sectors, what would you say are the main trends? And I'm particularly interested in anything that you would say this has developed over the last, let's say, couple of years.
00:05:33
Pablo Tomasi
So anything that has changed recently, and that as a consequence, is forcing a little bit of of a broader change, maybe in how you guys work and how ah some of the other people work.
00:05:43
Miguel Carames
Yeah, so it's ah it's a really good question. So if i if I think about the last couple of years, obviously we'll start with the elephant in the room, AI. There's no conversation without AI anymore.
00:05:56
Pablo Tomasi
It had to happen, right?
00:05:58
Miguel Carames
yeahs ah at some point it was going to ah to come up in the conversation, but but it's a a very real trend, i like ah right? at the, I guess the iPhone moment with ChatGPT.
00:06:10
Miguel Carames
where it made the technology very real for the end customer.
00:06:15
Pablo Tomasi
Thank
00:06:16
Miguel Carames
right like It's not like AI. yeahi is nothing new to telecom. right I've said this in a million presentations, even when I was still at Verizon. There's been a lot of AI in telco for a very long time, i because it's nearly impossible to run these networks and these businesses without machine learning and technology.
00:06:38
Miguel Carames
tooling and solutions, right? That's just the way it is. The numbers are too big, they generate too much data. it was not necessarily gen AI, right? So that's the the innovation that's developed over the last couple of years, but AI, there's been plenty. um And that's a trend that nobody can ignore, right? So there there is impetus for using the technology to be more efficient, right? And that's probably the the most important trend over the last 18 to 24 months has been this
00:07:13
Miguel Carames
Renew focus on innovation, sorry, in automation.
00:07:15
Pablo Tomasi
Thank you.
00:07:16
Miguel Carames
right So you have pretty much every telco around the world has reduced the task force some, some more, some less.
00:07:27
Miguel Carames
But they all have less people doing a lot of work. and And so they need definitely a lot of help with with automation to run the networks, run the services, help their customers.
00:07:39
Miguel Carames
identify problems before they invite the the end customer, make sure they get resolved, automate all these hundreds of changes that happen in these networks every night. right like the um The miracle of the technology is that so many things are changing all the time.
00:07:55
Miguel Carames
And yet, for the most part, we, the users, really are agnostic. right All this is transparent to us. other than when it goes bad, right? But for the most part, you don't really know and that all this is happening.
00:08:09
Miguel Carames
But that's a big trend. Then on the on the technology side, you know, we've gone through... So we went through the the moment of a lot of excitement with private networks and multi-access edge compute.
00:08:25
Miguel Carames
Then it died out a little bit, as in, OK, there are thousands of these networks out there, but is there a return or not? There were the examples that everybody was presenting, like the mines and the smart ports, which are big in sales.
00:08:39
Pablo Tomasi
um um' I'm covering in my other life as an Analyst problem matter, so I know the pain of that market quite well.
00:08:40
Miguel Carames
OK.
00:08:43
Miguel Carames
Excellent.
00:08:47
Miguel Carames
Yeah, so it's ah it's a very interesting thing that I do think there are verticals where where it's critical. And I saw it firsthand in my previous life.
00:08:57
Miguel Carames
But maybe not every application is perfect for a private network necessarily. Maybe needs to be a little more maybe more intentional, more careful. not you know Like when you have a sledgehammer,
00:09:11
Miguel Carames
It seems like that's the only tool, but but for some reason, you need ah something else, i like a little more the more granular.
00:09:17
Pablo Tomasi
Yeah.
00:09:18
Miguel Carames
um But this MEC is an interesting one because with the now with Gen.AI and this push for you know keeping the data on net, protecting the access to the data, the concerns about how the large language models, the frontier labs use the information and so on.
00:09:41
Miguel Carames
if It looks like it's still like new life on this. orreka Do you need your own compute? Do you then extend that compute to your customers, maybe at least they on the enterprise side? So thats ah it seems to be picking up steam again. That's ah that's an interesting trend that we're seeing as well.
Towards Automation And Talking About 5G Developments and Challenges
00:09:59
Pablo Tomasi
And it's it's interesting, particularly on the on the MEC side, because it's what we see now, so that idea of edge compute at the edge of the network, AI factory, yeah, grid, whatever, you know, once, part is MEC 2.0 pretty much, right?
00:10:14
Pablo Tomasi
MEC was was quite hyped a few years back.
00:10:14
Miguel Carames
Right.
00:10:14
Miguel Carames
Agree.
00:10:16
Pablo Tomasi
It never really made it mainstream, in a sense. But now, because of AI, it's definitely people thinking, OK, what should I compute and where and like the distributed network network and the sovereignty angle is definitely quite big.
00:10:32
Miguel Carames
grant
00:10:33
Pablo Tomasi
So one way or another, it seems that the AI is the big thing that is changing. and and and ah And something other you mentioned, so you mentioned that everything is becoming so complex and and big that you need AI.
00:10:47
Pablo Tomasi
So it' is that sort of the case? So we can expect more and more autonomy in how we manage the network. and And you can only ah achieve that pretty much with AI or ML. Is that sort of the direction they we're going?
00:11:04
Miguel Carames
Yeah, so I definitely think that' um you know that's an unstoppable trend, right? so And I think it makes logical sense. I think we are probably, in most cases,
00:11:18
Miguel Carames
a few years out from of the the what the team forum defines as level five and so on, the click and forget.
00:11:26
Pablo Tomasi
Yes.
00:11:29
Miguel Carames
We might never get there. There is also a point of friction between the technology and the culture of these organizations too. right like they um The fact that you don't want when you're running these networks and you're responsible for mission critical services for millions of customers,
00:11:48
Miguel Carames
you cannot afford to lose the sense of control. ah You have to feel 100% comfortable that if something was to wrong go wrong, you'll be able to root cause it nearly in real time and bring it back in control.
00:12:01
Pablo Tomasi
Mm-hmm.
00:12:02
Miguel Carames
And and I think ah that ultimate vision of you know you click and forget, like a big, red, easy button, right like ah
00:12:10
Pablo Tomasi
second
00:12:11
Miguel Carames
that you still have in the commercial. Probably it's more aspirational, but definitely you are starting to see it in pockets, like ah functional areas, maybe parts of the network or specific services.
00:12:24
Miguel Carames
And I think it's unstoppable and it makes sense. where you Get the machine to do what the machine does best. Use the technology to help you. like like yeah Probably a bad analogy, but like the Waymos or the you know like whatever Tesla calls the self-driving mode, i the the technology is there.
00:12:42
Miguel Carames
Why not use it? You still have a human in the loop, in the example of Tesla, and you will need it for a while. But you know you're i I'm a big believer that we'll see incremental progress towards many of these processes parts of the network will be very highly automated. And and yeah, for that, you need you know the capabilities that the AI provides.
00:13:06
Pablo Tomasi
And so what I find fascinating in a sense is that if if you look at sort of, you know, kind kind of branching out a little bit, some sci-fi stories or whatever we want, and and you have sort of worlds that are very much polarized in a sense with very advanced things at the same time is very kind of retro things. And when you're talking about, you know, you were just talking about the autonomous driving.
00:13:32
Pablo Tomasi
I never think about it because I live in Italy and that's really is not a thing. But thinking about the US, that is something that, you know, small or large steps, it is happening.
00:13:43
Pablo Tomasi
And I can imagine also with the network, it's going to be awesome. People really going full autonomous or as close as they can, very automated and others going to be kind of at at a different stage. So quite a quite a complex environment, definitely. And particularly when you know different networks, different autonomies will need to collaborate in one way or another. But one thing that you you didn't mention directly, i mean, you mentioned private networks, you mentioned MEC and those are some flavors, if you want, out of the 5G dream, opportunity, whatever. um what's What's your take on it? Is 5G done, delivered, what it was meant to do?
00:14:24
Pablo Tomasi
Is it too early to to kind of call it a day on success or failure? What's happening there for you?
00:14:32
Miguel Carames
Yeah, this is a very tough one. And so if you look at the, you know, the IMT 2020 definition or, you know, even um i remember we used to talk about the currencies of 5G, like all the value that it was going to deliver and this also, you know, many different dimensions.
00:14:52
Miguel Carames
So there was definitely the throughput, but there was the all this sort reliable low latency. There was like the high density of devices and so on.
00:14:57
Pablo Tomasi
Thank you.
00:15:00
Miguel Carames
And you step back I mean, I think we need to admit that we have not delivered all that value as ah as an industry overall. by like This is not blaming anybody, but I don't think the value has been fully delivered.
00:15:13
Miguel Carames
um I think it's honestly the case where um the moment we had the two evolution paths, so had the non-standalone and the standalone,
00:15:24
Miguel Carames
And the non-standalone was easier for obvious reasons. It's much faster to market.
00:15:27
Pablo Tomasi
Mm-hmm. Mm-hmm.
00:15:28
Miguel Carames
You could reuse more of the network that you had already built for 4G with you know some minor changes here. And there, obviously, you know a lot of work on the radio side.
00:15:40
Miguel Carames
But in the core, you could pretty much reuse the entire core and so on. It made it hard for the industry to convince themselves that the move to SA was that critical because you already had some of the value delivered, right? Like the additional spectrum could be put to use. You already had the, you know, the icon for the end customer. You had better throughput.
00:16:02
Miguel Carames
you know, we lost some momentum there. And I think it's, you know, that SA looked harder and harder, right? Like to get there. um we We will see.
00:16:13
Miguel Carames
I think now as more and more networks deploy SA, some of the promise can be unleashed, right?
00:16:14
Pablo Tomasi
Yeah.
00:16:20
Miguel Carames
Like slicing and so on. And maybe we'll see a second wave of life behind 5G and 5G's advanced type capabilities.
00:16:31
Miguel Carames
But but it's ah it's a really tough question because right now, if you look at the financials, there was a lot of investment made. and arguably very little return, right? Because customers were already paying, i mean, they live here in the US, so all the customers have unlimited plans as is, right?
00:16:49
Miguel Carames
So you have different tiers of unlimited plans with different caps after your which you'll be throttled and so on, or maybe different resolution for your video and stuff, but it's already unlimited, right?
00:17:01
Pablo Tomasi
Mm-hmm.
00:17:02
Miguel Carames
And it's hard to explain to the customer at times that you are, paying more for more unlimited, right when they feel like what they have is good enough. right So finding that killer app is still a little elusive for 5G.
00:17:18
Miguel Carames
ah But I think it's critical because otherwise, who how do we get to 6G without some value being generated 5G? This is a tough one for the industry.
00:17:29
Pablo Tomasi
and And for me, that that is a little bit of the of the challenge because let's say you don't have 6G, right? So you only have 5G. You have time to make it work.
00:17:39
Pablo Tomasi
Whatever you know it takes like 2, 3, 4, 5, 10 years because that is sort of as good as it gets for now. and that's you know And then you can really put your best brain to decide how to monetize it. But because 6G is coming sooner rather than later, and I think a lot of the attention is is kind of shifting, least the hype in a sense, just like, okay, you know, 5G done and dusted, but 6G is going to be this and that and everything. And I think that it from a pure monetization point of view, that is a little bit detrimental because people are just saying, oh, no, we need to focus on the newer thing.
00:18:14
Pablo Tomasi
Is that the right strategy? No.
00:18:18
Miguel Carames
Yeah, 6G both excites me and and worries me right a lot. because they So obviously for for the industry overall, these 10-year cycles are very important. right like The major generational upgrade cycles are are key to find the R&D world that gets done and and so on.
00:18:43
Miguel Carames
But it maybe for the first time my career, it feels a little bit early. It feels like it got here before we really took full advantage of why we built in 5G. So it's going to be interesting in the next couple of years. i mean, obviously, right now, is not specs are not really implementable yet.
00:19:02
Miguel Carames
So we'll see when we get there in 2028 and then closer to 2030 where things are. And then how much of the reusability of the ready-made investment in 5G is a factor and those in those specs. It's going to be very interesting, actually. Very interesting. I'm i'm wondering.
00:19:21
Pablo Tomasi
Yeah, definitely. you know For like a techie person like yourself, it must be like quite quite cool to to see what's happening, what's not going happen. When I'm looking so the from a business perspective on strategy, strategic if you want, I'm always wondering, like it some let's say you're the first telco that starts pushing some 6G out, whether it does a lot or not, yet to be seen, right?
00:19:42
Pablo Tomasi
But is that going to force again everyone to run into sixty because someone is doing it, you know, kind of before it's fully cooked? So that that's going to be interesting to see. But, you know, we're going to have to wait a few years, I guess.
00:19:57
Miguel Carames
That's right.
Leveraging Data in Telecom
00:19:57
Miguel Carames
We'll talk again in 2028.
00:19:59
Pablo Tomasi
that But if we instead of go back to to you know nowadays, and so you have AI, you have 5G in one form or the other, you have telcos that are building these massive network complex network
00:20:14
Pablo Tomasi
and And I always hear people saying that telcos have a lot of data, what is like their own data from the netto, data from customers, and that one of the bottlenecks possibly for telcos struggling a little bit in in some of the growth area ah was in the fact that they haven't really been able to capitalize on the data that they have.
00:20:33
Pablo Tomasi
is Is that the case? Do telcos really have that much data and could they do more with with what they have? What's your take?
00:20:40
Miguel Carames
So this is very interesting. And I'll mention a conference that I attended a couple of weeks ago with the Rocco guys. um So telcos do have gold mine of data, right? Because that's your really your main point of engagement for your digital life, pretty much for all of definitely for the younger generations. I look at my kids, and you know their laptop is an afterthought. It's something that you use for school.
00:21:08
Miguel Carames
the phone is the primary means of engagement, right, for everything digital.
00:21:10
Pablo Tomasi
Yeah.
00:21:12
Miguel Carames
So they they do have a lot of data. they are um they They are very constrained by what regulation allows them to do. And in a way, um if you look at the market in 2026, it's quite interesting how ah how many limitations are there for the telcos.
00:21:36
Miguel Carames
and how few are there for the over-the-top providers. right And i think I know the argument is always going to be, well, many of these over-the-tops are providing the service for free. And the moment you accept the terms and conditions, you are given all the the rights for them to productize your data and whatnot.
00:21:45
Pablo Tomasi
Thank you.
00:21:53
Miguel Carames
But the reality is a very uneven playing field right when it gets to what one part of the market can do versus the other. And I think that's um In a way, and this is where I wanted to go back to that conference that I was mentioning, it was a very interesting conversation that followed a hackathon. kind So a few vendors showing some solutions on APIs and monetization and so on.
00:22:19
Miguel Carames
And then there were a lot of questions on privacy and regulation. And this was a very global audience. So you got to see how the differences in the different market, but the underlying commonality that was coming very clearly across is every MNO that was represented was very concerned about getting in trouble from a privacy perspective.
00:22:31
Pablo Tomasi
Mm-hmm.
00:22:41
Miguel Carames
right and And then call the the very philosophical question that followed was, well are we as an industry maybe taking it too far like by being worried and trying to do the right thing?
00:22:55
Miguel Carames
Maybe we're limiting ourselves versus you know Pretty much in every market, there is an opt-in ah you know set of rules and regulations that apply. So if you can create the trust with your customers,
00:23:09
Miguel Carames
to go to this opt-in and and make the value feel you know clear for the customer as well. why What's in it for me? What do I get for if I give this information?
00:23:18
Pablo Tomasi
Yeah.
00:23:21
Miguel Carames
Then could we unleash the opportunity? Again, as an industry, right? And and I think it's something that that we really need to to think about very carefully because that's where there is some more opportunity.
00:23:32
Miguel Carames
if there If the customer is not going to pay more for more unlimited, because they already feel they get everything they need, then how else can you get the incremental revenue that allows you to keep investing in these networks? Because you cannot stand still. That's that's ah that's a given.
00:23:49
Miguel Carames
You need to keep putting more spectrum. You need to keep you know upgrading your network. Every operator will tell you the same thing. that You can deploy as much cap capacity as you want in the network. and yeah A month later, two months later, you're already at capacity again because everybody's using the devices more and more. So that need doesn't go away. So how do we make the customer feel that this is not a they are being abused, but this is the way that the telcos have to give them more services without charging more for it directly, directly right? and I think that's where the opportunity lies.
00:24:23
Miguel Carames
not Not a simple problem. Otherwise, I would be retired by now. But you know I think it's one that we we got to tackle as as an industry.
00:24:27
Pablo Tomasi
Thank you.
00:24:32
Pablo Tomasi
Yeah, and and and and it makes sense. It's definitely a larger problem. and And I think the mindset, that as you were mentioning, maybe the telco industry becomes a little bit too conservative in a way compared to some of the other guys that sort of they do something and then don't ask for permission, but ask for forgiveness type of thing.
00:24:52
Pablo Tomasi
Right. And sometimes I'm not saying every time, but sometimes they can really help.
00:24:53
Miguel Carames
Correct. Yep.
00:24:57
Pablo Tomasi
again, opening up those new opportunities. And then because of new opportunities, you have a better up chance of shaping the market and shaping the regulation that is coming. But again, jumping around a little bit, we we mentioned AI.
00:25:11
Pablo Tomasi
And I think this is quite connected with all the data that the telecoms have and everything that's
AI's Role in Telecom Efficiency
00:25:15
Pablo Tomasi
happening. Where do you see the biggest impact for AI within the telecom industry?
00:25:23
Miguel Carames
Yeah, so i mean we see so right now, i guess you can break it into value for you know internal use. and then a value facing the customer.
00:25:33
Pablo Tomasi
Mm-hmm.
00:25:36
Miguel Carames
So in in the areas where we play on the internal use, we see a lot of efficiencies, right? So if we look at the some of the products that we we build and solutions we provide, like for example, on the on the fraud management or the revenue assurance side,
00:25:56
Miguel Carames
we we already have the product out with the Gen AI capabilities and it's all about explainability, right? So there is not just, listen, you know this particular MSISDN or IMSI, it has crossed some thresholds, so you should really look into it.
00:26:15
Miguel Carames
The analyst gets the verdict from the our agent already, right? So this with a very high level of confidence, It's a fraud case. it's a fraud this scenario. This is all the evidence pack that we create that supports that something should be done about that particular line. And this is what we suggest gets done next.
00:26:33
Miguel Carames
Today, still human in the loop, right? So the analyst to review and agree and take the action. For some of these cases, like the 80-20 rule, right? So many of them will be Self-explanatory, you have maximum confidence that the action can be automated. So you could do the blocking, throttling, you know canceling the account on an automatic basis with very limited risk.
00:26:57
Miguel Carames
There'll be some that you'll still want a human to confirm and maybe help improve the logic as well. right So they'll they'll have an opportunity to say, yeah, agree with analysis. Or no, maybe you understated this particular signature. And and you know this one is more important for this particular type of fraud. So we see that that that's a use case already in production.
00:27:21
Miguel Carames
We've seen one another one that we've also already brought out commercially, which is on the testing side of the house, where we have these products that are very technical.
00:27:34
Miguel Carames
right so it So it's a solution built for engineers by engineers.
00:27:36
Pablo Tomasi
Thank you.
00:27:40
Miguel Carames
oh extremely flexible, extremely powerful. You can really simulate and emulate anything you could possibly want to on a global basis. right and And I was a customer of the solution and yeah I'm an engineer, so I really like that flexibility.
00:27:56
Miguel Carames
But now when you fast forward 10 years later, many of the power users of this system are no longer engineers, are business analysts, are roaming managers that just want to see you know, what's the quality that their customers enjoy it when they travel to a particular network in a particular country, right?
00:28:12
Miguel Carames
And so now they don't necessarily want to engage with this very technical platform.
00:28:12
Pablo Tomasi
you
00:28:16
Miguel Carames
And so we have a, you know, like a natural language-based bot interface that allows them to do feature discovery, pull reports from the platform in another few months, get root cause automation as well. So when they, you know, they are running their testing and something failed,
00:28:36
Miguel Carames
they'll be able to tell what failed, what to do next, what actions to follow, and so on. And so that's the second area of efficiency for us. We we want to make some of the products that we already have out there easier for more people to use at the customer side. right So those are a lot about efficiencies.
00:28:56
Miguel Carames
And then when we look at products for the end customer, there are some solutions that we've done with a partner tackling the small and medium business.
00:29:06
Miguel Carames
just you know If you are a small flower shop and you have your website where you can publish your information, also get queries from customers and maybe orders as well, we can very easily help that flower shop create their own agent either that then would be integrated into the platform.
00:29:26
Pablo Tomasi
Mm-hmm.
00:29:27
Miguel Carames
They don't need to be super technical. They don't need to learn you know some of the frontier model APIs and so on, you know in an hour, they are up and running. right so that we see that as an opportunity again for these segments. Obviously, if you are a large enterprise, chances are by now you already have your own AI team and so on, so maybe less applicable.
00:29:48
Miguel Carames
But for small a small and medium business, we see that they want to participate in the trend. They might not have the capabilities. And then and then we see a lot on the customer-facing scam and spam protection. right So you know I'm sure you get a number of calls that you don't want to get, that you just ignore. Yeah.
00:30:08
Pablo Tomasi
I think that's it everyone's most annoying thing about the smartphone right now, you know, all these scammed calls.
00:30:09
Miguel Carames
and
00:30:15
Pablo Tomasi
And some are really good, to be honest.
00:30:18
Miguel Carames
They are very good. And i so it becomes harder and and harder to detect them and prevent them, which is why I think we feel very strong that there are Gen AI-based offerings that can be added both for text and voice and data as well to help protect customers.
00:30:37
Miguel Carames
And those could be revenue generating. So it's not only an efficiency thing, but
00:30:40
Pablo Tomasi
Thank you.
00:30:43
Miguel Carames
if If you were able to trust your the voice channel again, or you were able to protect your elderly parents or your kids, would you pay for it as a customer? right and And I think the answer in many markets is going to be yes.
00:31:00
Pablo Tomasi
Yeah, I would definitely think so. It's like kind of making life easier, right? And and you know that's ah that's a big problem. But a quick question on, so you you mentioned your solution helping create an agent for for like, say, a small flower shop. Would that be solved for a teleco, I would imagine?
00:31:16
Pablo Tomasi
Or would you go sort of directly with your?
00:31:19
Miguel Carames
Yeah, so this in in our case, we we always, you know, it's a partnership with the telco. So the telco sells to the, and then after that, the model can be a rev share, you know, but but yeah, we it' it's through the telco, yeah.
00:31:34
Pablo Tomasi
Okay. Yeah, that that makes sense. Because like again, like ah when you're thinking about as as small small business, theyre their knowledge of IT is probably limited, their knowledge of AI, even less so.
00:31:48
Pablo Tomasi
And and maybe someone that is not particularly young, right? So definitely, it that it does need to have that that that easier way than you want.
00:31:56
Miguel Carames
For them, some of these solutions are are you know like if you don't respond or don't respond on time or don't send the proposal on time, it's an opportunity gone.
00:32:06
Miguel Carames
right There are plenty of flower shops that somebody else will pick the business. So I think the motivation is high. But yeah, to your point, it's not like they have IT teams lying around. So making it simple for them is ah it a big thing.
00:32:20
Pablo Tomasi
And then the ah the other thing that that you mentioned that sort of I found very interesting. so you were mentioning about sort of the the fact that now non not only engineers, but also broader set of people may be using the data for business purposes. So in a sense, there is that democratization, if you want.
00:32:40
Pablo Tomasi
of what used to be a traditionally extremely complex environment. And I found that that fascinating because it kind of shows that because of these new tools, then it becomes easier for people that probably wouldn't understand, you know, what it was happening 10 years ago, but now they have the the opportunity to interface in a different way then can use the data for their own you know benefits, KPIs, new products. and And I think that's sort of the the way forward, but kind of connected to this,
00:33:10
Pablo Tomasi
And we sort of talked a little bit about it, but just wanted to try to get in it clear answer. Ever going through autonomy, is there a an ideal scenario in terms of how much the human element you will like you expect to see remaining in the loop versus how much?
00:33:34
Pablo Tomasi
Is this sort of a gold standard, let's say, that the industry has? Because I know that everyone says full automation and everything, but we already discussed likely not going to happen. So what are you talking about? and and know Maybe 1% human involvement, or does it even make sense to think in that way?
00:33:50
Miguel Carames
So to be honest with you, I think this is ah an evolving situation still. I think there'll be parts of the network and specific services where you'll be able to get to, you know, like high night is midnight is automation ah because they are a lot more a lot more deterministic, a lot less can go wrong that you haven't been able to predict and so on.
00:34:19
Miguel Carames
But there'll be there'll be other things where, you know, like if I if i think it's, ah you know, the network is the pyramid with, you know, the bottom, you have the highly distributed, right? Like RAN and edge data centers, and then you have your edge core, and then your centralized core.
00:34:35
Miguel Carames
The higher up you get, there the higher the impact of something going wrong, right? Like in that it can really show on a nationwide basis. So those are the areas where probably you you need to be hyper careful.
00:34:50
Miguel Carames
even Even when you automate, even when um you allow the machine to make most of the decisions, you might not feel comfortable with all the decisions. i in um
00:35:00
Pablo Tomasi
Okay.
00:35:02
Miguel Carames
if you know If it's something simple like, OK, this particular, say it's a Kubernetes-based deployment already, and one instance is approaching capacity, and you've defined your orchestration such that in that case, you can you know create a new instance and recreate the connectivity and so on. Maybe you feel comfortable with that. But if if you are in ah in a different and scenario in which the the orchestrator is saying, well, this particular network function instance doesn't seem to be healthy.
00:35:34
Miguel Carames
and the recommendation is to reboot it or kill it and instantiate the new one. you You probably want a human to validate, at least at this stage, just because the stakes are too high at that point.
00:35:47
Miguel Carames
So I think that that's what I see.
00:35:47
Pablo Tomasi
Mm-hmm.
00:35:49
Miguel Carames
it's you probably and And when you see some of the practical use cases that the TM Forum presents, you see the different MNOs targeting service areas right or or a particular product.
00:36:03
Miguel Carames
And I think that makes sense. right Experiment, learn, decide if that applies to other areas, and then deploy similar parts of the network.
00:36:13
Pablo Tomasi
So and a bit of a side question, but then it's, so the responsibility, where does it lie? So if I have an automated software and and it gives me wrong recommendation, eighteen and me as a customer, do I assume that responsibility is from the software side or how?
00:36:36
Miguel Carames
So this is a very philosophical problem at the end of the day. So I think they um the operator runs the network, right, one way or another.
00:36:47
Pablo Tomasi
Thank you.
00:36:48
Miguel Carames
so they'll they'll the Even in the extreme, you'll still have operations teams. They will be responsible for building their automation, deciding what level they feel comfortable with. And that's where the responsibility lies at the end of the day.
00:37:03
Miguel Carames
um This is a little bit of the the internal friction they still have. right if they are If the goal is to drive a lot of automation, there is always going to be risk in the early days.
00:37:18
Miguel Carames
right And this is one of the things I always
00:37:19
Pablo Tomasi
Yeah.
00:37:21
Miguel Carames
remember earlier on, so this would have been probably about 10 years ago, we had ah an internal conference and my previous employer and and they brought some people from what's now called Meta, right, from Facebook.
00:37:34
Miguel Carames
they And there they were sharing kind of how they run their network. And it it was fascinating, right, for for us because it was so different and it's much smaller things, much more software oriented.
00:37:47
Miguel Carames
But it came with this very deep belief of calculated risks. Calculated risks being very different from risk-free, right?
00:37:53
Pablo Tomasi
Okay. Yeah.
00:37:56
Miguel Carames
That's saying, yeah, every once in a while, things do go wrong, right?
00:37:57
Pablo Tomasi
lela
00:37:59
Miguel Carames
And at that point, you might scramble some, but without taking the calculated risk, you'll never get, if you want to be 100% safe and you know have kind of the same Modus operandi we used to have where somebody writes a mob and then three people reviewed it and then somebody goes and implement it. It's not like that didn't cause problems too, right? Like ah you can fat finger command or, you know, you missed a precondition or whatever. So these mistakes can happen.
00:38:28
Miguel Carames
But I think it's that acknowledging that you are taking some risk. You're going to make some mistakes along the way. But that's an unavoidable part of the journey to get to that automation that you want to.
00:38:40
Miguel Carames
Because that that's how you learn, too. And nobody likes the mistake when it happens.
00:38:43
Pablo Tomasi
Yeah.
00:38:45
Miguel Carames
But you know how do you do it otherwise?
00:38:48
Pablo Tomasi
Yeah. and and yeah and And that is pretty much sort of the way that we learn anyways. Like you cannot learn anything without making some mistakes. Of course, easier said if the mistake is trivial, much harder if it's like a nationwide network and something goes down.
00:39:02
Miguel Carames
Correct, correct. right
00:39:03
Pablo Tomasi
but So the same principle, let's let's put it this way. ah So it seems we're running, as always, like around a little bit longer than expected, but just a couple of quick last questions. First one is that, okay, go back to the beginning.
00:39:18
Pablo Tomasi
Your company is doing a lot of different things, exactly on the rolling side of things. I also saw you wrote a blog on on the D2D opportunity. You mentioned some private networks. Where do you see the low-hanging fruit for you with all the AI that is happening?
00:39:31
Pablo Tomasi
Which segment of the market you think, okay, that's going to be the faster
Mobileum's Growth, AI, and Plans
00:39:36
Miguel Carames
So to be honest, we see it across all all the, so we have five business units, right? So roaming solutions, we have risk, revenue assurance, financial assurance, we have testing and service assurance, we have DPI and analytics, and we have security solutions, which for us is signaling firewall, SMS and voice firewall.
00:39:36
Pablo Tomasi
one for now?
00:39:41
Pablo Tomasi
Okay.
00:39:57
Miguel Carames
All five have opportunity when it gets to adding Gen AI capabilities and making the solution. you You use the democratization of the data, and and that's a big one that we see. there there There is the Gen AI component, and there is the being easy to integrate into Gen AI. right And I think we we have to play on both. i We have to be platforms that are easy for our customers to integrate into whatever their vision is.
00:40:31
Miguel Carames
and And hopefully, we can help them there. But if they decide to go with somebody else and our our solutions are data producers, then that's OK too. rea You can win every single opportunity.
00:40:44
Miguel Carames
And then where we see kind of growth potential for us is combining all these discrete products that we have into sub solutions, right? Because these are very distinct areas, so like very technical to very business-oriented.
00:40:58
Miguel Carames
And we're spending quite a bit of time and effort in assembling those into solutions that provide value to deploy once and solve for many use cases, kind of being the philosophy.
00:41:10
Pablo Tomasi
Mm-hmm.
00:41:11
Miguel Carames
And I think that's the, I don't know that's a low hanging fruit. There less and less low hanging fruit. Most is being picked already. So you need to climb a little bit, but I think that's where we have an opportunity as a company.
00:41:23
Pablo Tomasi
OK, but it seems quite very, very positive anyway. It seems that sort of this, you know again, the trend that we discussed in expansion of the I, the complexity of the net, but it's also kind of bringing it all together. So very good to see and a lot of momentum there.
00:41:36
Pablo Tomasi
And then literally changing completely top, but the last question, because I'm trying to, you know there are lots of podcasts, so trying to be slightly different, let's let's let's be honest is there.
00:41:47
Pablo Tomasi
um Do you have any recommendation within the world of sci-fi? And I can definitely see a couple of gadgets on your wall that would indicate any book, TV series, movie. What do you think the audience should be keeping in mind, having a read and look?
00:42:05
Miguel Carames
Yeah, so it's a good question. So I have to say i got more into sci-fi stuff when my kids got into it. I would say so i haven't watched Mandalorian yet, so that's in the in the to-do.
00:42:20
Miguel Carames
um One TV show that I did like a lot, and I i saw the first season, and I thought it was OK, but I really liked the second season, is The Fallout. on um or an Amazon Prime, and it's based on a video game.
00:42:32
Pablo Tomasi
Oh, yeah.
00:42:36
Miguel Carames
And at times, a little over the top, i was watching I was watching it on a plane to Europe, and at times, I was a little bit worried with the person sitting next to me. youra ah Some of the episodes can be a little bit gory, but I thought it was really nicely done. The sequence system was very good.
00:42:52
Miguel Carames
So that's probably the one that I would say. And then... ah I'm excited about Dune 3 when it comes out. So I thought Dune 2 was really, really nice. um Didn't care for Dune 1 so much.
00:43:03
Pablo Tomasi
Thank you.
00:43:03
Miguel Carames
Dune 2 I thought was very good. So I'm excited about Dune 3. So that's one of them I'm looking forward to.
00:43:07
Pablo Tomasi
Oh, really? thought the one was also quite good. I also remember the the older movie, Dune, don't know, from the 80s, 70s. That was like slightly different terms of special effects and everything, but it but a good one. And Fallout, yeah, great video games. And I really like the series as well. Definitely a little bit on the gory side, so probably not for ever ones. But ah I thought particularly because it's based on a video game,
00:43:35
Pablo Tomasi
It was really entertaining, like and because sometimes when something's built on the video games, it may be little bit skinny you know plot-wise.
00:43:43
Miguel Carames
yeah
00:43:43
Pablo Tomasi
Sometimes there is that that that he issue, but yeah, definitely good recommendation. And I think, yeah, at this point, we can probably close the the podcast. So Miguel, thank you very much for your insights. Really appreciate you taking the time to join me you me on the show. And for everyone listening, thanks again. and feel free to comment, and share and discuss with others and talk next time.
00:44:08
Miguel Carames
thanks to everyone you appreciate it