Guy Daniels, TelecomTV (00:13):
So in a moment we're going to have our first panel discussion. But to set the scene, I'm going to ask Danielle Rios to come and join me on stage. In fact, go to the lectern please DR for a presentation. Round of applause please for DR. And Danielle is Acting CEO of Totogi.

Danielle Rios, Totogi (00:35):
Well, thanks Guy. I guess a couple of things for you all. Number one, this is the last panel before lunch and so you're probably grumbling and hungry. And then number two, I have like 35 slides legit, so here we go. It's not a joke. So yeah. So to start I'm going to ask you guys a question and that is, this is not working. There you go. How many of you believe you can vibe code in an enterprise grade carrier grade application? One person, two people, two people there. You're well, three if you count me. So that's what I thought. Today, you believe building A BSS requires years of development teams of specialized software developers, deep knowledge of telco standards and millions of dollars in budget. You believe that telco enterprise software is too complex, too critical, too regulated to just vibe code. But let me tell you why you are wrong.

(01:45):
In 2023, AI could barely write a function without bucks. By 2024, cloud was scoring 34% on coding benchmarks. Today we're seeing 77% accuracy on complex software engineering tasks. But it's not just the models, it's the entire ecosystem. Cursor went from auto complete to full application generation replicate can now take a simple description and build, test and deploy complete applications. Azure AI Foundry connects requirements directly to running code in the cloud. The entire journey from product requirements document to test it, deploy application. Something that used to take your team months to do can now be done in hours. But here's the critical insight. None of these tools understand Telco. They don't know TM form standards, they don't understand telco specs, they don't realize your vendor ecosystem. These tools build generic software, not carrier grade systems. But even if these tools understood Telco, you'd still have a problem. You don't have AI talent.

(03:06):
You need people who understand how to wield these AI development tools and no telco. So Totogi built the solution, a platform with an ontology at its core that fully understands the telco domain and provides a universal translation layer for AI and a team of elite four deployed AI engineers that know how to use it, that's the combo you're missing. You might be wondering why can't the traditional BSS vendors do this? Because their entire business model depends on complexity. Long implementations, professional services, revenue vendor lock-in, they need it to take years and cost $20 million, not when you use Oggi right outside these doors. Quite literally, we have something we call BSS Magic running in our demo pod. It's not an agentic system on top of some BSS. It's a complete platform that understands telco like your best experts. Understand telco, it knows your standards, it speaks your business language, it thinks in your ontologies.

(04:21):
BSS magic isn't another platform you have to rip and replace your existing systems for. It connects to what you already have, even if it's decades old and completely closed off. It translates everything into one language. So when you build something, it works with your existing systems, immediately extend your ecosystem, don't replace it. And right now as I'm speaking to you, one of our four deployed engineers is using it to build a complete BSS from scratch. Live today, not a demo, not a mockup a functioning BSS, because if you can vibe code A BSS, you can do anything. What we're going to do today is build a CRM product catalog, order management billing and revenue management invoicing today while you guys are having coffee and to prove to you that this is real, I want you to decide what feature we add to it. Just scan the QR code and give me your feedback.

(05:32):
Should we add an AI assistant to assist users ai, first customer support ticket resolution, a new regulatory requirement, an automated phone call to welcome new subscribers or blockchain payment integration for new iPhone 17 orders. While we eat lunch, our FDE will add this feature to our BSS. So visit the demo pods during the breaks, check the progress, watch your feature come to life at any time. You can test the full VSS being built. But this isn't just about building faster. This is about fundamentally changing what's possible in enterprise software. If we can build a production ready BSS in a day, what does that mean for your digital transformation? Let me be clear about what that means. The traditional BSS vendor model is dead when you can build production ready BSS in days instead of years. When you can add features during lunch instead of waiting for the next release cycle when the gap between idea and implementation collapses, why would you ever sign another multi-year implementation contract? It's proof that the right combination of telecom context and elite talent changes everything. Our four deployed engineers aren't generalist developers, they're specialists in rapid system construction. They combine world-class software engineering with to Tokis BSS magic platform. That combination elite execution plus deep telco context is what makes the impossible possible. And by tomorrow they'll have built something that would've previously taken years to do. Ladies and gentlemen, this is the future of software. It's happening here, right outside this room today. See it for yourself. So let's go build A BSS today. Thank you. Thank you.

Guy Daniels, TelecomTV (07:49):
Yeah, thank you. Applause. Fantastic. Thanks Daniel. Come and sit down. Sit right in the middle of here and what we'll do is we'll get our next panelists up and all over the panelists up onto stage as well. So lemme click the magic button and we should see, there we go. Ai, native telco from concept duality. It's the first panel of the event this year. So this is our first panel. As I said, we're going to discuss the progress of the AI native telco from concept two reality and I'm going to ask my guests to come up onto stage. So let's bring them all onto stage. So we've got four more guests if you'd like to come up onto stage with Danielle here and take your seats and then we can start our panel. And once you've finished building this, are you going to tell us how you got on tomorrow? Are we going to see the results tomorrow?

Danielle Rios, Totogi (08:36):
Yeah, I mean you can see it right now. I think we've already finished two modules, so we kind of got a kickstart this morning right at nine o'clock. Our engineer has been working and I think I've seen at least two modules finish. Fantastic. So

Guy Daniels, TelecomTV (08:51):
We'll tune in tomorrow

Danielle Rios, Totogi (08:51):
Then you can see it. I mean it's pretty amazing.

Guy Daniels, TelecomTV (08:55):
Thank you very much. Excellent. Right, well thanks everyone for joining us on stage for this opening panel. I'm going to start and get our guests to very briefly introduce themselves, name companies, really briefly starting on my far left with Andrew.

Andrew Collinson, Connective insight (09:09):
Hi everyone. Andrew Collinson. I'm the managing director and founder of Connective Insight and also the Unthinkable initiative. I hope to see some of you for breakfast tomorrow. I we'll be counting on watching, passing along.

Nabil Lahyani, Nokia (09:21):
Hi everyone, my name is Nabil. I'm head in analytics business unit autonomous network in Nokia Cloud network services and I'm one of the firm believers in autonomous network.

Guy Daniels, TelecomTV (09:32):
Thanks, Nabil.

Danielle Rios, Totogi (09:34):
Danielle Rios, everyone calls me DR. I'm acting CEO of Totogi and also founder of TelcoDrR that's a thought leadership. Maybe you've seen my podcast Telco in 20 or read my blog that I put out every two weeks.

Eric Van Vliet, Dell Technologies (09:48):
Thanks Eric Van Vliet. I work for Dell Technologies, telecom systems, business units like Kiran, who you saw earlier. I head up the sales and go-to market organization for emea.

Guy Daniels, TelecomTV (09:58):
Great. Vivek.

Vivek Chadha, Rakuten Symphony (10:00):
Hey everyone. Vivek Chadha. I work for Rakuten Symphony. I run the global cloud platforms business. Great to be here. Haven't been in Dusseldorf for a while, so thanks for having us on here.

Guy Daniels, TelecomTV (10:10):
Oh pleasure. Good to see you all. Thanks so much for joining us on the panel. So we are talking about what it means to be an AI native telco going from that concept, and we spoke this morning about the concept and when we first came across it about just over a year ago, the nature of what AI native means, but how do we get that to reality? And what I'd like to really start us off with is how do we define AI native for us telcos? Should AI be treated as a foundational operating principle like cloud native has been treated or like a tool set on a case by case basis as and when required? So I think there's a lot of definitions going on here. Andrew, are you able to start us off and just to say yes, we'll see you tomorrow morning for breakfast, but are you going to just start us off and give us some, your take on what it means to be AI native?

Andrew Collinson, Connective insight (10:58):
You probably pick the worst person to start with. You've got a panel of proper AI experts here. In my mind though, I'd say there's probably two or three components of it. I thought Ahmed gave a very interesting technical description of AI native, which basically means you build everything. So AI can use it technically, my feeling is that there are probably two other and maybe many more, but the two that are in my mind, one is the business framework to say, are you thinking about AI as a core principle for your business? And the other thing I was going to say is the piece really missing is the natives, the people. How do we as people interact with it? I'm slightly disappointed, Danielle, you've got an engineer, not an AI building this, but I'll let you off. But quite seriously, technology works. The technology works to some extent.

(11:50):
I mean, let's discuss and debate with ai. The real challenge in telecoms is getting change made is getting stuff done. And I was talking to Philippe in earlier and he'll talk to George later and we're talking about cloud native and it still hasn't really happened 10 years later. Why? Because you haven't got the natives on side. So for my book, there's a piece of this jigsaw, which I always come back to, which is the people. How do we take people with us? How do we change the industry so that it embraces this? So three layers for me, technical business and people.

Guy Daniels, TelecomTV (12:21):
Nabil, what's your take on AI native?

Nabil Lahyani, Nokia (12:23):
Thanks. You really made me think about this native component of it. I tend to agree. I think we need to differentiate between AI native and AI assisted because we tend to fall into the trap of just adding on AI and whatever we are doing. And we say, okay, this is AI driven kind of tool of or so in my opinion, AI native means you have to start from the foundation bringing ai, that mean network has to be driven by AI operations. They have to be driven by ai, the business processes also, they have to be driven by ai. Last but not least, to fix these 360 transformation people. Part of it, the organization, part of it also has to be AI adopting ai, not AI driven. And if I simplify it from let's say the telco process perspective, that means when we design a network, the design has to be driven by ai, not by human looking at the map or running drive test.

(13:26):
And then you say, okay, I'm bringing AI to run a macro on top of it. So the design of the network must be driven by AI based on the data we have about the area, about the network, about the demand of the market, the forecast. If we move, then we have the deployment. The deployment must be also driven by AI about how do we do an automation, really automation and the deployment. Then if we move forward, how do we tune that network with ai? How do we optimize the network with ai? How do we play with capacity in a smart way? How do we understand the behavior of the subscriber impacted by the network? How do we predict the traffic? Last but not least, how do we maintain the network or do we need still the call center people running? So in a nutshell, if I summarize it in my humble opinion, AI native means you are really conceiving the process, the product, having AI as part of it really native.

Guy Daniels, TelecomTV (14:18):
Thanks Nabil. I'm going to go down the line and DR.

Danielle Rios, Totogi (14:21):
Yeah, I mean I think everyone will kind of answer it in the same way, meaning it's foundational and you're using it as you design things, but kind of focusing on the second part of your question of like is it as foundational as cloud native? And I'm actually going to just flip that around and just say, let's pretend you didn't treat it like that and your competitor down the street did and started to redesign every single role, every single job thought about it being AI first redesigned the telco from the ground up and started getting some of these benefits that you see like we're demonstrating today and we estimate we'll generate 500,000 lines of code today on site. The average developer does about 4,000 lines of code and that's not totally committed into your repository. So if you think about it, if your competitor starts doing this and treats it as foundational and starts and it's hard, we've been talking about it all day, it takes work because it's not figured out. There's not a blueprint, there's not a roadmap. Vendors are figuring out themselves, sis are still figuring it out, but someone's going to crack this nut, someone's going to figure out how to do it and what side of the line do you want to be on? And I'm like, I think you want to be on the line that you're in the game and competing versus on the sideline waiting to see, well, is this going to work and is it foundational or not?

Guy Daniels, TelecomTV (15:43):
Absolutely. Thanks. Eric...

Eric Van Vliet, Dell Technologies (15:46):
Well, we heard we need change of operations, change of thinking mindset. Yes, AI native is a design principle, but I think it's much more the risk we have and telcos and us in the infrastructure we're very good at focusing on the technology. Is it feasible? Can it be done? But at start with business benefits, I would love to see a future where telco goes, it costs me X amount of tokens to operate my network and now I'm looking not to reduce power, but I'm looking to reduce tokens, therefore I will reduce power, I'll reduce infrastructure costs, et cetera. So I think that's where it starts because building A BSS is great, 500,000 lines of codes, but how much did it cost? Maybe not humans, how much did that acceleration cost and what does that mean for operating my network? And I think this change in our behavior comes down to how do we measure success and a fully self-functioning operating network, it'll still cost money to do that. There's business investment needed. How do we measure it, how do we improve it while we still deal with SLAs, penalties critical infrastructure, the fact that we deal with humans that buy the network, which always is at the back of our mind and then also keep up with society who's also consuming AI and monetize that or have enabled the telcos to monetize on that.

Guy Daniels, TelecomTV (17:11):
And this could so easily fall into a trap of a marketing term. It's convenient. It sounds good. AI native sounds fantastic, but there's real deep meaning here. There's purpose for that, isn't it? What does it really mean for a telco and its partners?

Vivek Chadha, Rakuten Symphony (17:24):
Is this your way of reminding people how much I moaned about the term cloud native all these years?

Guy Daniels, TelecomTV (17:30):
You have to put everyone into context here. Yes indeed. We've had an ongoing discussion about the misapplication of cloud native I think over the years and it's still going on today, let's face it.

Vivek Chadha, Rakuten Symphony (17:38):
Well thankfully AI native isn't as enmeshed in controversy yet. And I think there's a lot of salient points that the rest of the panels made. But just reflecting back on the essence of your question, should AI be treated as foundational or an add-on? I'm going to use a metaphor and coincidentally I'm talking a little bit more about the same topic later in the afternoon after lunch for those of you who probably will still hopefully be with us. I think the fundamental difference that we see as Rakuten in the change from how industry and vendors looked at cloud native versus how we all have to look at AI native, I can sum it up in a tagline. A cloud native was essentially what's the best way of doing deployment? How do you deploy infrastructure in the most automated, scalable, low cost manner? And that goes with the entire cultural process tool set change includes the cloud native people journey as well. So what's different in AI native, the focus of the question has changed. You no longer think about how do I deploy? You think of who decides what do they decide and who controls the decision.

(18:56):
And if your business is using a significant element of AI to do these three things, you probably have just claims to say you are more AI native than the others. The example that you gave with the BSS magic is definitely leaning towards that kind of mindset and cultural change. Nobody, and I just refer to the DR example for BSS magic, but this could apply at any domain in your network actually it could apply to any enterprise. You could be a bank, a car manufacturer, a retailer, how you think about CapEx, opex investments, you currently have a template, a blueprint and industry guideline, a board that defines that. What if that was no longer the holy grail and source of truth? And it's a journey because humans don't like losing control. You're not going to have your stock price be subject to a fancy algorithm, at least not anytime soon. So my response to this, and I can see Philippe smiling, so I'm sure he has a point of view on this. My response to this is I think the fundamental question that we were focused on for cloud native is changing when we think about AI native.

Guy Daniels, TelecomTV (20:05):
Very interesting.

Nabil Lahyani, Nokia (20:08):
Please, can I comment something?

Guy Daniels, TelecomTV (20:09):
Yeah, course.

Nabil Lahyani, Nokia (20:09):
Vivek said that the three questions they need to be answered have AI in most of the operations or processes. I would like to challenge that we answered the question about AI native, right? Or if it should be AI alone, we can resolve business problems like Eric said, without having ai, we should not be conservative making this statement. I mean there are many business issues we can't solve still without ai. So yes, if we aim to be L four, L five in autonomous network, definitely AI has to be native there, but there are many use cases. There are many business issues that can be still solved without ai. I just want us to emphasize that it sounds like we cannot break without ai. We are forced to have AI.

Vivek Chadha, Rakuten Symphony (20:58):
I think we are in agreement because the example that you gave would fall into the category of, and I think it was Bejoy if I'm not mistaken. He was talking about the known unknowns.

(21:09):
You can solve a lot of things today because you figured it out. I agree. What is it that we don't know? What is it that's not visible based on our petabytes of databases and CDRs and all the algorithms we've been running, the wow factor that comes when you have an aha moment, when something either using RAG or contextual learning gives you an insight and your network engineer or a board member goes, I didn't realize we could deploy this in a different way and have an impact on ROI five years down the line. That's an unknown unknown today, and I think that is what the potential of AI is all about.

Nabil Lahyani, Nokia (21:44):
About. I agree with you,

Vivek Chadha, Rakuten Symphony (21:46):
The low hanging fruit is exactly what you said. Autonomous networks is a lot of engineers in this room regardless of our current job titles, right? By trading. So we like predictability, we like deterministic approaches. So the frameworks we've evolved, let's create a maturity model or an index that shows level one to level five of autonomous networks because it helps us put things in place and there's a grade that we can chart a journey on it. Who says it's only level five? What if in five years we're talking about a level six of the network, which today is not even feasible. So I guess my deep rooted belief on the promise of ai, not the premise of AI, is how it'll start revealing unknowns.

Eric Van Vliet, Dell Technologies (22:32):
But then you're talking much more about being AI ready than anything else because in the end, yes, we have unknowns, but if you now make decisions today and in three years or two years one of those unknowns pops up and your data is managed wrong and therefore you can't leverage AI to solve the problem and there is no solution to solve the problem with AI or it's too expensive, that is then the easy. So maybe the questions as well as what does AI native mean? Maybe it means being AI ready. More importantly, make sure your data is AI ready, so that's when the time comes and you want to implement AI to solve a problem better leveraging ai, then you can do it. So maybe that's part of the equation here to think data. How can it be leveraged by AI when it needs to be? And sometimes solving a problem with a certain technology might not be the best, but it's better because coherently as an enterprise or as a telco, you're moving the mindset forward of using ai and it is a mindset that we all as engineers and in marketing or whatever role we have leveraging this technology, we get better at it. Same as that, leveraging other technologies, 10 years ago, 20, we got better and better and then we answered those unknowns, how it can help us and deliver those business benefits.

Guy Daniels, TelecomTV (23:51):
It's an important question, isn't it? Because I do feel unless we get this right from the beginning, we're fooling ourselves. We're just going to waste time and effort going down the line. Andrew, you want to come in on this?

Andrew Collinson, Connective insight (24:00):
Honestly to say a couple of things. One is his hard, horrible reality from outside the rim of engineers. Nobody cares if you're AI native, right? They will care if Danielle's right and you kick ass basically in the market, you share value goes up, the investors are happy, everybody's happy. That's 0.1. And I'm not seeing enough proof. I want to see proof use cases are great and I like to see theoretical proofs from use cases, even better case studies. Show me the money, show me the results. You need to be factoring in more of that into this discussion and I'm saying that generally to this room, kind of raise the game with your stakeholders. The second thing I wanted to say, sorry, Daniel, is go back to the people question. AI is the technology that is probably least people friendly of all the technologies I've ever seen potentially.

(24:56):
You have to be honest about that. You have to deal with that. You have to cope with the change in the transition that this may mean to your industry. You've got to deal with the stakeholders. You can't just hide in the technology wonders of whether we get an event stream in this piece of data or not. You've got to deal with some of the harder questions here and I think if you do that, you can make progress much quicker. If we just keep talking about the technology, we're just going to be going around the same thing. Sorry, I'm being really hard on you guys today, but I feel it's really important if you want to make a different industry, you've got to grapple with the deeper questions.

Guy Daniels, TelecomTV (25:30):
Absolutely. DR, want to come in?

Danielle Rios, Totogi (25:32):
Yeah. Well I forgot my point, but that's a good point too. I think a lot of people shy away from that very tough discussion and as a previous HR leader, it's so important to be honest with your people in terms of the impact of their jobs. I mean, AI is going to impact every single person's job and at first it's a little bit scary, but when you fight through it, everyone in our organization, if I went back and said, okay, we're not doing AI anymore. I think they would fight me tooth and nail. Their life is better with ai. They work better, they work faster, their quality, it has improved. And so we've kind of gone through that hard part. But yeah,

Andrew Collinson, Connective insight (26:13):
You are an outlier in that regard.

Eric Van Vliet, Dell Technologies (26:14):
I don't think she is. I think if you look at ai, if I go back to my days in high school or primary school, AI is like a calculator. It was banned at the beginning. It was not allowed to use it because I needed to have certain skills to survive in math. And then it was introduced and they got smarter and smarter. And if you look at the first calculator I got versus what you now get at high school is hugely different. It's the same with ai. So I disagree, Andrew, that yes, it's scary for everyone because change scary, but it makes us better. I was reading a report at the start of the year and it felt very foreign at the start of the year. It was a business ratio of agents per human. It feels very cold if we're going to measure businesses, how many agents per human do they have? But the reality is what we're measuring, how effective is your human? So what can you do more with the other humans that you are employing?

Andrew Collinson, Connective insight (27:07):
I love the fact you disagree because I love disagree, man. I think it's a great way to open out conversations and dialogue. And I'll just say I work with some people in agent AI who talk about AI producing new workforces, a completely new workforce of ai. That's how some people are thinking about it. I think there's a spectrum from the use and the Daniels of this world and maybe some of the people in your organization are on the far ages of thinking about this. Others are not. You've got to take them all with you and just go back to my cloud example. Yes, there's plenty of issues with cloud, but if you're going to get AO to be as powerful as it can be, you've got to take the folk with you or you've got to find a way to move your organization in that sense. So we agree, but we disagree. I think you're right, but

Vivek Chadha, Rakuten Symphony (27:52):
Also not, I'm not sure if I'm going to disagree, but I just needed a clarification when you said it's probably the most unfriendly of technologies, what stakeholders did you have in mind unfriendly for who?

Andrew Collinson, Connective insight (28:02):
What I meant, maybe I phrased it wrong. What I meant is I think it's one of the most fear inducing from an employment point of view. It sounds like it's going to take my job. Like the steam engine. Sorry, like the steam engine.

Guy Daniels, TelecomTV (28:13):
Yeah,

Andrew Collinson, Connective insight (28:14):
It's that kind of thing. Yes, it's like a transitional industrial technology.

Vivek Chadha, Rakuten Symphony (28:18):
So I will both agree and disagree and that's not me hedging the response. I'll disagree on the fact that if you look at consumers, I'll take some very simple, probably trivial examples to substantiate what I think I firmly believe in my limited few decades of experience. I'm not going to get into numbers now. This is the single most profound technology change

(28:43):
For the consumers that is yet to manifest itself properly. We are at the tip of the iceberg. My father's a retired army officer, he's a little over 75 and he struggles even on his iPad with some stuff. He's technically literate, but I mean at that age, there wasn't even a mobile. By the time he retired, mobile phones weren't even invented. And he's now quite conversant with Chad GPT, and it took me 30 minutes to get him hooked onto that, right? We do a little bit of home finance through an assortment of credit card statements and stuff, et cetera, all using a chat GPD interface that is better than anything else we've ever had right Now these are very basic, simple examples. You could say, what's that gut through with a telco expert or a telco network, et cetera. And I'm just saying if you split the stakeholders you were referring to in different categories for the end user, depending on how we create use cases and impact using ai, they're probably going to love some of the outcomes. Some of them will just be, like you said, it's the same thing with how many of us,

(29:51):
Well, maybe not in this room, but on the street, how many of you really care if it's a 4G signal or 5G signal, you want your phone to work. So from that analogy, I agree. If it's AI ready or enabled or native, none of the consumers really care. Does it really improve my life in a certain way?

Andrew Collinson, Connective insight (30:07):
Look, again, you're completely right. I, and let's face it, there's so much transformational stuff you can do as a consumer or as an enterprise. I'm talking about inside the industry. I'm talking about what makes change happen or not is whether people want to do it. And if there's some very good research, I think from Google on this saying that it's people that are either enabling or stopping it. So I'm really sorry. You are supposed to be the guard rails here.

Guy Daniels, TelecomTV (30:32):
Yeah. I'm going to jump in because I love how our TelecomTV panels evolve and snake around the subject and go into interesting areas. But I've got to bring it back on track because we are looking at moving from the concept to reality. I was going to talk about use cases and real world examples. We've had some use cases shown today. We're going to hear more later this afternoon, we're going to see more tomorrow. But as we move from concept to reality, how do telcos balance this need for quick wins? There is a need to show low hanging fruit, quick wins in the application of ai, but how do they balance that with this vision for a more longer term holistic native cross-domain approach? How do they scale from maybe just pilots and concepts to full deployments? Tricky.

Danielle Rios, Totogi (31:23):
Yeah, I mean any new technology, and not AI specifically, but I mean you're going to run, as Andrew's been talking about, this is huge change management with your employee population. And so number one, the executive team has to be 100% aligned that this is where the company's going. It's not just that you're doing AI for AI's sake, but you've set out your business goals. They need to be big and bold, but then you've decided that there's some way to use AI to achieve those goals. So tie it to the businesses you were talking about. But what's really important about change management is it's really easy to get on top of a stage and get everyone together and rah, rah, ai, we're going to be a better company, more profits, lower cost, blah, blah, blah. But it's super important that you start to show quick wins to your employee population to show that you are serious about the change, that it's inevitable and happening and that people need to get on the bus or leave voluntarily if they don't like it.

(32:24):
And that's so important in change management, and I've done this several times as an HR executive helping cultures change that if you don't have those visible signs that were serious about changing, everyone just goes back to their cubicles, their offices and they working the same old way and then a year later you're like, oh, we did autonomous network, blah, blah, blah, AI use case. And everyone's like, what was this again? They've completely forgotten the narrative and it's not connected to the story. And so when you're going out with your executive team and talking about your big new bold ideas, you need to have some quick wins pretty much already lined up. So you can just say, we're doing it, it's real. And people start to feel that change. They'll get on board real fast.

Guy Daniels, TelecomTV (33:08):
Thanks Eric.

Eric Van Vliet, Dell Technologies (33:09):
Yeah, I totally agree. I think using the words change and stuff is less in my mind it's always how can I and how can I make my team better by leveraging ai? And we in Dell for the last couple of years, we've seen an influx of tools that's come up and even though we are humans, we don't like change. We are creatures. We like our own habits. When you see your colleagues do something a lot faster than you did, and especially when it's a task that is not fun and you wonder, how did you do this in an hour while it takes me a day, then that's a small win when you able to not push it to them, but they see others. And I always say, it's not that you say if you don't adopt it, you leave, but you're going to be left behind.

(33:59):
So you want to get on that train and looking at your father. If your father now after 30 minutes can use chat g pt, imagine what a new employee can do that comes on board when rather than having to talk to 50 people for a year to be onboarded, it takes him a week. How about that now a week and he's at the same level. Now that's scary because that could mean all other sort of business benefits for companies that you can imagine. But on the other side, I want to hire somebody within a week. They're onboarded, they can use the tools, they know the company, they can deliver what I need them to deliver. If AI can do that for me as a company, then that's a huge win for everyone because I also don't have to then train someone as I can get on with my job rather than actually, so there's lot of benefits that, but I always think when you talk about adoptions, talk about how are we getting better? How can we more effective doesn't mean getting rid of people means what? Can I do more with the people I have? What other special projects or features can I deliver? Can I leverage? And now I don't have to spend a year making a BSS. That means I have a whole engineering team that can do a year something else that doesn not even have to bring revenue because I have the outcome already. So what can I do? How can I get better?

Guy Daniels, TelecomTV (35:17):
Thanks Eric. Nabil will come in.

Nabil Lahyani, Nokia (35:19):
Yeah, I think it's a very good reflection. I would like to answer the question short term as you said, quick wins with the long term short term with what Vivek and Eric said if I pick up a board from you, definitely in the short term, operators are looking for use cases that can show to the board there is a benefit of AI and from Nokia, from analytics and autonomous network. I'm proud always to talk about the use case, which is in sustainability, in energy efficiency, there are real use case where the network can save energy, you can reduce the footprint, sorry, without having human being interaction. Definitely before, to your example, we had engineers in optimization behind the power saving features. It requires effort, it requires time. There are expected human mistake in the process. Now you can predict the traffic, you can see the behavior of people around the site.

(36:11):
You can shut down the radio module through AI ml and then you save energy in your electricity bill. This is fact, this is real. Other use cases, quality of experience, there are many other use cases also in predicting the capacity on how to reduce the response time when there is an incident like anomaly detection, the famous use case now to move from the quick win short term to the long term. Definitely as you said both of you, Daniel and Andrew, there is a transformation, I call it transformation beyond the change management. And we need to accept it. We need one voice from the executive to the bottom. People have to believe in out of comfort zone, which there are many unknowns in ai. There are a lot of stuff that we don't know. We need new people. And I always use an example of something that we didn't touch here.

(37:03):
That to me is fundamental, which is the holy grail data. You can have the best algorithm in the world, you can have the best Ferrari car in the world if you don't have the right driver, if you don't understand the data, if you don't trust the data, forget about ai, rubbish in, rubbish out. And that's why I always, I'm proud of the partnership we're having with Deutsche Telecom because they're really taking very, very seriously the data part observability, Hey, do we trust the data? So we start talking about concept like data quality, data lineage, is it data complete about the latency? Then we can power up the AI and to speak this new language, we need new people, we need the transformation in the process, we need the belief. So the transformation is a required is a masterful long term. However, we see some positive sign in the short term in some use cases as I mentioned.

Guy Daniels, TelecomTV (37:49):
So we know that short-term gains are important, quick wins are important. We establish that as well as the longer term planning. How do you go about that? Do you split your teams, one team focuses on the today the quick wins one team is more longer term planning or do you have to sort of parallel work together?

Vivek Chadha, Rakuten Symphony (38:08):
I can share what in certain parts we're doing within Rakuten Symphony, but also other parts of Rakuten. But let me just take Symphony specifically. We do have a notion of run the business and change the business. Most of the quick wins essentially target either one or all of CQT cost, quality and time. So the low hanging fruit is easier. So for run the business, there's a set of processes you're doing today whether you're in engineering or marketing or sales or legal or product. And with the use of ai, we are able to impact either one or more of cost quality on time. And some of this is already underway in Symphony to Nabeel's point, the voice of the executive, the exec review for the AI initiatives in Rakuten is done by hi Hir himself.

(39:00):
So yeah, we don't get to hide. It's down to every team, every function, et cetera. So it is full on and has been for about a year now, maybe a bit more. But the longer term which has changed the business is where elements of transformation creep in. Again, I would say that this, and this probably doesn't come as a surprise to anybody in the audience. If you're a brand new startup today, you can probably leapfrog what would otherwise be friction of gravity compared to most enterprises is you can't just throw out what's running and start afresh on an existing brownfield estate. I'm not just talking about a network, any business because you also have people, you have cultures, you have processes, you have skills. So I think recognizing that in change the business, the way that we are scaling is whatever is proven in run the business and if it has a bearing and an applicability into change the business, we are starting to redo the process for that in a phase banner.

(40:02):
But then ensuring top down that the change is pushed through to examples I can take. One is around new product introduction. I've mentioned this in the past. We are actually using elements of AI now to qualify which products are actually getting investment versus others. And the other is around contractual clarity and reviews. For some of our commercial engagements, none of these is related technology per se, but they do intersect all fields of commercial product strategy, technology, et cetera. So I think that's how we are trying to create a two faceted engine of saying can we start, there's actually a third benefit which is actually more critical, at least in our opinion, sorry, it's very hard to change minds of people and you've had some experience in that compared to most of us in the room.

(41:00):
And change begins from within. So if I have three and a half thousand people in Symphony and if 80% of them are going to question, what's this got to do with me, I have a problem. So one of the things we've done as part of run the business is give the freedom and flexibility and we run an AI hackathon, an ideation hackathon every quarter, and now we are publicly celebrating people who are coming out with ideas. We recently had something very clever come out on cost control and this is now imbibing a ceiling of ownership in our employees where AI becomes relevant to their daily life and they feel part of the journey rather than being preached to or being told that you have to change or learn something. So I think that is important.

Eric Van Vliet, Dell Technologies (41:47):
Yeah, you're feeding the entrepreneurial desire that everybody has. We all want to do more and create things. And I agree with that and I think we've seen that in Dallas as well where we're asking our people to think outside the box. But what you get then is 500 different use cases and then you have as a business challenge, how do I pick the first five or four to pay for the next four or five? And I think that is the other challenge. It's not for free, even though it doesn't maybe include a lot of human costs, there's a lot of other costs and in a world for telcos where they almost becoming a utility now, commoditized is almost no longer the case. Their connectivity is a utility, how do we pay for this? And cost saving on the human side or another side is not the answer because it doesn't scale that way. We have to find revenue streams for ai. And again, I said, how many toss does this cost for my network? Well, if I have an enterprise and I get a fixed line, do I get free tokens with that? Is it something that Elco could monetize to offer AI GPU as a service coming with the connectivity Today you don't see these things, but those are new business models that could generate revenue to pay for the investments to drive AI in the organization.

Guy Daniels, TelecomTV (43:10):
And the token is the new currency, isn't it? Andrew?

Andrew Collinson, Connective insight (43:12):
I just want to say people, I love what you're all saying, but I want to say something cultural about us, about Europeans, and I'm from the uk, apologies that everyone else in Europe, but we do think of those as Europeans.

(43:25):
We have a great benefit in the world in that we have Americans like Daniela who are going to forge ahead with this stuff because they're just going to do it. They're just going to do it. We also have the Chinese and they've already done it. I was talking to George Glass briefly earlier about, because if you look at the Chinese operators, they've made a lot of progress with a lot of this stuff. They've just decided they're going to do it. Now the Chinese don't want to put loads of people out of work. So my point being to US Europeans is this is happening, it's going to happen. What I'm saying to management is don't cover it up, be honest, talk about it because it's really interesting when you start talking about the reality of it, it's not quite so scary what fear is. Fear is a hidden emotion and particularly the most obstructive fear is a hidden emotion that people are not telling you about and are just not doing things as a consequence of.

(44:19):
So in my mind it's not about is this going to put people out of work or not? I don't know. Spreadsheets didn't particularly did they? So why should AI things will change, but the hidden nature of it, the fact it's not being talked about and grasped in the way that some of you're now talking about it, which is wonderful, is so important. So I'm saying add to your technology, important things, talk about it, make it real because what's going to happen is if you don't change, Daniela's company is going to come along or one like it is going to come along and put you out of business anyway. So there's no point in pretending it's not going to happen. There's no point hiding from it and not dealing with it. You've got to face the change.

Guy Daniels, TelecomTV (45:03):
We've kind of covered the last point I wanted to raise, which is quite handy. We've run out of time, but I wanted to look at the challenges from moving from concept reality. I mean it's kind of coming around here. There's a lot of cultural challenges and people and process challenges to get there and these will be discussed rest of today and tomorrow. But I'm just going to go to the audience some questions in a second, but I just want any last comments from the panel about any other challenges that we face that prevent us from moving to where we need to be with the reality of the AI native.

Vivek Chadha, Rakuten Symphony (45:33):
I mean I can share what we generally discover when we talk to partners and customers. Some of this has been in parts true internally as well, but I would rather refer to the larger sample data set talking externally outside of Rakuten, three things typically general, whether people admit or not, general lack of clarity on what do they mean when they say they want to be AI native or AI enabled and they are now using ai. Is it fomo? You're just making sure the boss is happy. We do get elements of that and there's no shame in that because there isn't a silver or a gold standard today available for anyone. A lot of people are enterprises and the leaders are discovering this as they go along. The second is around skill sets because because some people who had taken an early plunge into this are now discovering that scale does require at least in the foundational years, fairly complex skills depending on what part of the AI value system you on a play in because nobody does it all either your infrastructure or the application layer, et cetera. And the third point, which has probably been not quite beaten to that but highlighted significantly is data soon becomes the oxygen that starves this initiative. And it is not lack of data. It is sometimes too much data in multiple places. You really don't know how to grapple with it and data scientists don't like playing it. They like doing data science,

(47:04):
But more about that in the second session.

Guy Daniels, TelecomTV (47:06):
Perfect. Thanks very much. Any more comments? Can we take a Nabil? Do you want to come in?

Nabil Lahyani, Nokia (47:10):
Yeah, I think you highlight very well Vivek and just a reality check. I talk a lot to customers and sometimes they come and they say, Hey Nabil, what's your own catalog of ai? Share with me please. Why? Because as you said, the boss wants to implement AI use cases to monetize 5G investment and they say, yeah, but what's your problem? Tell me three problems. Just share with me the list of AI use. I think it has to do with leadership. It has to do with the strategy. I think IT department has different objective from network department, from operation, from marketing. I think that's top down understanding of where is the strategy of the company and what are the problem we're trying to solve. It'll help us to define and work more on specific use cases, not just, Hey, I want to claim I'm ai.

Danielle Rios, Totogi (47:53):
I mean that's what we do with Totogi. You have to have a business case of what we're we write it down, what are we trying to do? Is it a cost thing? Is it a revenue thing? And then we measure it within four weeks and do it, go do it, deploy it and it works. And I think telcos that we meet with are surprised that our stuff really works. They're usually skeptical. They're like, there's a lot of people talking about it and they can't get the pilot off the ground. They can't get it done. We are shipping working pilots and I think the second biggest barrier is the way telcos buy today, right? They want RFP it. They want three vendors and I have to write now at the top of the RFPs, literally I'm like, I'm going to write answers and they may be totally different next week there could be MCP came out, changes everything.

(48:42):
Agent kit comes out from OpenAI changes everything and we're pumping everything back into BSS magic. I'm not reinventing agent kit or MCP, I'm leveraging that. I mean we're talking about hundreds of billions of dollars that are going into these models. I'm writing that way and I disagree with you that worry about costs, cost is going down on the models. It's a massive commodity. So you got to figure out where that value layer is and ship it and get it installed and build from there. And the hardest thing is going from zero to one, but once you get to one, it fricking explodes. It's pretty amazing.

Guy Daniels, TelecomTV (49:17):
Thanks, DR. Eric.

Eric Van Vliet, Dell Technologies (49:18):
Shockingly, culture, people and data are the challenge for the telcos to implement.

Guy Daniels, TelecomTV (49:23):
We've never heard that before, have we?

Eric Van Vliet, Dell Technologies (49:25):
That's the first unknown. Unknowns, right? And the unknowns, right? That's the future. Look, I didn't mean cost is high and cost is a factor. So going to AI is not free. That's what I mean.

Danielle Rios, Totogi (49:38):
Yeah, but I mean the cost keeps going down. Correct. I have the stat. We think it's down like 90% from when it first started, so it's dropping.

Eric Van Vliet, Dell Technologies (49:45):
Totally agree with that. Look, there's a lot we can say more. It's a technology that we all have to get used to. I think culture and people is by far the biggest blockers you said or somebody said they need the skills, how to use ai. Now I've used one element of AI to tell me how to use another element of ai. We have copilot now helping me learn how to use another tool, how to fix a piece of code or how to do something where before my behavior was as an individual, I go to a search engine and I read and I digest and I use my brain to come up with,

Guy Daniels, TelecomTV (50:26):
Read the manual, just read the manual, but not anymore. Read the manual.

Eric Van Vliet, Dell Technologies (50:28):
And I think that is the culture and the people aspect. If we all slowly get into this, we can go as fast as the people go in the end. And I think that's in the end how this is going to be a success. It's here to stay. It's a business benefit. Let's go. That's basically the message that everybody should have in their mind.

Guy Daniels, TelecomTV (50:49):
Fantastic.

Danielle Rios, Totogi (50:50):
Zero to one.

Guy Daniels, TelecomTV (50:51):
Yes.

Danielle Rios, Totogi (50:52):
Get to one.

Guy Daniels, TelecomTV (50:53):
Get to one. I'm going to give you a real horrible choice. A question or lunch.

Tony Poulos, TelecomTV (50:58):
I've got questions.

(51:00):
If that I ask the question, I'm going to ask one,

Guy Daniels, TelecomTV (51:02):
Right? You can have one. No questions. No questions, no questions. Great. We're going to get a question from Tony. This is going to be a cracking question. Come on, Tony.

Tony Poulos, TelecomTV (51:08):
If I was a telco and I had to decide what use cases to go with, what do I aim for with ai? Do I look for problem areas? Do I look for cost cutting first? Do I look for profit making areas? What's the driver for finding use cases for ai? Because we're hearing about a lot of use cases, but I don't know where the synergy is. Some companies have got 32 use cases going and they're not connected with each other. So how do you know what to aim for? What's the advice for a telco,

Guy Daniels, TelecomTV (51:39):
Right? Tony's going to be selling SIM cards later, but for now, help him out. What do we aim for?

Danielle Rios, Totogi (51:43):
Yeah, I mean I wouldn't start on, I mean I know we've been talking about network use cases. I wouldn't start with network first. I mean this is your crown jewel and if you screw it up, it's noticeable by a lot of people. So I'd probably pick areas right far away from that. And so that's actually why we're looking at BSS first, because that's a good entry point, right? There's a lot of opportunity, there's a lot of cost. I think everyone should already be doing customer support. If you're not do that. Lots of structured data, lots of, I mean 60% of your support tickets are billing. So that's a great big pattern. That's a good place to start. So start with the easy step that zero to one, it doesn't have to be hard. You don't have to solve level five autonomy on networks. Whoa. Go back, do something easy, get that zero to one, get your people trained. Start to build a culture of acceptance around this and then go from there would be my advice.

Vivek Chadha, Rakuten Symphony (52:39):
More advice not to disagree, but most you can disagree. Most telcos have a BSS. So it would be coming back to Eric's point, it would be a displacement ROI calculation to move to a brand.

Danielle Rios, Totogi (52:51):
No, not with BSS magic. We don't displace it

Vivek Chadha, Rakuten Symphony (52:53):
With whoever. But on your point on the network on that, I will disagree because it's probably the easiest thing you can do without disrupting anything. We've been partnering with Intel, Dell, et cetera. Rakuten Cloud can transparently clock down C states and P states right from the core to the ran with zero human intervention and create anywhere between 4% to 18% energy saving across the entire network with no user impact on experience. Now everybody's running compute, everybody's got some sort of cloud, and this is industry standard stuff, and it doesn't require you to change a single line of code in any of your application. That's just one example. There could be a few others.

Danielle Rios, Totogi (53:35):
I mean, same with BSS magic, not changing any lines of code, right?

Vivek Chadha, Rakuten Symphony (53:39):
If I just summarize the question here.

Guy Daniels, TelecomTV (53:41):
Help Tony, the Tony the Telco, yeah, he needs help.

Vivek Chadha, Rakuten Symphony (53:44):
Find early wins to create headroom, to build the confidence to invest a little bit more for the longer term.

Eric Van Vliet, Dell Technologies (53:51):
So my comment would be revenue generating use cases, something you could sell. I think if we focus just on cost saving and then well one, the shareholders will be happy. But if you start saying, well, I'm saving money, but I have to invest it, that's a difficult conversation with the shareholders, but having a conversation, I need to invest and I'm going to generate more revenue because of that, that will accelerate everything. So from my point of view to revenue generating use cases,

Nabil Lahyani, Nokia (54:19):
Excuse me, your name, sorry. Tony. Tony. So really, thanks Tony. Really, really good question. And to link it to generator revenue, yes, new streams of revenues, but let not run away or share away from the issue that CSPs are telling us all day. There is a cost operational efficiency issue. Listen, 80%, 80% of the time spent by the data scientists and huge organizations is to understand the data. 80%. We have some very knowledgeable and respected analysts in the room and they can confirm that we're working together with them. 60% of the POCs, they don't work out because we don't understand the data to create the use case. So to correlate ran with core data to really understand how can I then generate new revenue streams or cut my cost? I think that is one of the biggest challenges they're facing today.

Guy Daniels, TelecomTV (55:08):
Fantastic. Brilliant. There you go. Now go make a success of that erritory of yours, Tony. Yeah. Let's know how you get on. And meanwhile, I'm going to click this slide because this means it's lunchtime. So I'm going to ask you just to stay seated for a second, but we will be back in just over an hour. It was an hour and a quarter, but it's now like an hour and five minutes because this panel was so interesting. We'll be back on stage at 2:00 PM show up also applies to our online audience who are watching the stream. So round of applause for all our guests. Thank you very much.