Guy Daniels, TelecomTV (00:09):
Hello and welcome to our special program on the Future is Autonomous Nokia's Zero-Touch Network Revolution, I'm Guy Daniels and in the rapidly evolving world of telecommunications, the demand for seamless connectivity and superior user experience is high. Autonomous networks powered by artificial intelligence are at the forefront of this evolution. In our discussion today, we will explore how AI is revolutionizing telecom boosting resilience, cutting costs, and unlocking new business opportunities. We'll delve into the opportunities and challenges of integrating AI and data and discuss how to identify and prioritize AI strategies that deliver the highest ROI in an AI driven world. And joining me to discuss these issues are Didier Clavero, Group Network operations director at Vodafone and Jakub Votava, Director Network for O2 Czech Republic. Hello everyone. Thank you so much for taking part in our program today. Now CSPs are increasingly leveraging autonomous networks models to enhance network performance, fault management, operational efficiency and customer experience amongst other functions. However, leveling up autonomous network capability is complex. How would you both describe an AI ready telco? What are the key characteristics that make a telco AI ready and what stage is the telco industry at today with its AI adoption? Didier, let me come and ask you first.

Didier Clavero, Vodafone (01:58):
Okay, thank you guy. I think that, let me probably try to define what is the future CSP, I mean what we want to have in a future CSP and what are going to be the characteristics. I think one of them is data availability. I think that without data all the rest probably is not going to happen and we need to have a very clear data strategy and a very clear, I mean, data repository where we can base all the rest. Then of course it's about autonomous network. I mean we are hearing a lot about the different level of automation. We are hearing a lot about level three, level four, level five. We need to be at level four in order to provide the services that we need to provide and now we are, I mean, just building the basis in order to have this evolution and to ensure that we are there.

(02:50):
Second topic is cloudification of our networks. I think that in we started many years ago, more than 10 years ago with the NFV transformation, now we are in a different phase and it's in the clarification of our networks and this will give us the possibilities to offer new services and network slicing or differentiated services per customer and this is key also in the future telco. Then probably the third component that we have is how we are opening our networks, how we can also ensure that different APIs and different services can be connected to our network. Seamless, and this is a very important topic that we are working and it's also a very, very important topic for the telco of the future. And of course finally is also ai. I mean we are hearing a lot about this. We need to be ready and prepared to the new ai concept. I mean Agentic AI is going to be more and more happening. We'll discuss later about different use cases, but these are from my point of view, the four different components that I see that are needed for the telco, the future.

Guy Daniels, TelecomTV (03:55):
That's great, Didier, thank you very much for that. And Jakub, what's your thoughts? What constitutes the AI native telco here?

Jakub Votava, O2 Czech Republic (04:05):
Thanks guy for the question. So the data and the data structure is really one of the cornerstones. How to build the capabilities of the AI that's not just having the operational data in a good shape, but as well about the building of the knowledge basis and the knowledge data in parallel with the operational data to be able to feed the ai. But I see as well two other important aspects to that. One are the processes and really being able to handle the processes in the company in a digital way because digital digital process is for me the prerequisite to be automated and to plug in the AI capabilities into such a process. But I would say for me the most important one are the people. So people needs to be ready for the AI challenges which are coming and if we talk about the AI readiness in a telco space, then the people are the most important and for me there are two aspects of that.

(05:22):
One is the scale so to people has to get used to the AI skills and have to shift their capabilities from let's say being the engineers using the CLI to let's say being partially data analysts which have deep knowledge of the networks to help to build the right AI frameworks and they have to have a will, which is by my opinion, even more important because sometimes you can see that the people sees AI as something that is going to replace them and we need to change the people mindset to see AI as a tool that can in a great way help them to build a better network and to serve our customer better.

Guy Daniels, TelecomTV (06:17):
Can you share with us the biggest challenges that you currently face with integrating AI into your network operations and how you are going about addressing them? Didier, can we start with you?

Didier Clavero, Vodafone (06:38):
I think that data strategy, I would say that in all CSPs we have plenty of data. I mean the amount is incredible, is huge. The numbers that we need to manage and I think that we have a very clear data strategy and to really ensure that this is integrated by the way in all the different processes in the company is key. Then I firmly believe that the other important challenge that we are all facing is we need to adjust and we need to change our processes. I mean you can have the best automation in place that if processes still are not updated and are not in line with the strategy that you want to follow, it's going to be impossible to achieve a good result with this.

(07:22):
Of course, there are other challenges that are probably more linked to the network transformation that we are having and if we talk about network automation for example, if we talk about cloudification of our core networks, of course we are also now in a moment where we are moving for example, in terms of so rollout in terms of how often we need to upgrade our core system, et cetera. I think that we are moving to steps that are happening every probably two, three months and we are working of course with Nokia and with other vendors in order to see how we can do this in a much more transparent way. I think that if we continue managing the network as we were doing two, three years ago, I wouldn't talk even about 10 years. It could be completely impossible that we can follow the page that we need to follow now in order to be able to have a cloud network open to APIs and ready for offering differentiate customer services. Then from my point of view, also cloudification on how to change the processes and be ready is something that is not only about C-S-P-C-S-P and vendor activity that we need to work together in order to achieve the goals that we want to achieve.

Guy Daniels, TelecomTV (08:37):
Great. Thanks so much Didier and Jakub. Are there any additional challenges that you face that you can share with us?

Jakub Votava, O2 Czech Republic (08:45):
For sure, and we've been talking a lot about the data today. For me, there is one more aspect. Let's say the data quality which needs to be covered. If we would like to use the data for the AI and especially in the network operations where basically times matters and that's the real timeness of the data because currently a lot of the systems is providing us, let's say the tons of data, but they are most of the time delayed and we need to as well talk about the data granularity in the time domain and as well in the subscriber domain because if we would like using the AI tools and using the automation to get to brisk and let's say focus answers, we need the data in the granularity which does support that, and this is something I currently see a lot of the systems and a lot of the legacy systems is lacking is to provide the data in the cadence and in the granularity which is needed for all that modern use cases.

Guy Daniels, TelecomTV (09:56):
Let's move on now and talk about security. Can we, because security is a cornerstone of autonomous networks, so let me ask all of you about this. In the context of a rapidly evolving threat landscape and threats becoming more targeted, what's the status of AI network security? Didier can I come to you first?

Didier Clavero, Vodafone (10:19):
I think that the securities, well a very, very key aspect that we need to cover when we talk about AI or also when we talk about network automation. I think that, let me take an example. I mean the first phase of AI or network automation, I think that in all the different CSPs was focused more on the, I could say how we are detecting fault, how we are opening tickets in an automatic way. Okay, that is true that it's something that should be protected. It's something that should have also security on top of this, but it's true that the implication in case of a problem are not so big as the ones that we can discuss later. That is when you are moving to the next phase of network automation, that is you are closing loop. You are really, I mean taking decisions and there are agents taking decisions and non-human humans and then the risk that you have from a security bridge from security threat is much bigger because at the end, I mean you are opening access to your network, you are opening access to your element managers, to your systems, and this is a key aspect that we need to protect from the beginning.

(11:30):
Then I think that from my point of view, when we are talking about the AI evolution and when we are talking about agents especially, I think that security by design and ensure that what we are doing, what we are deploying and coding is taking from the beginning security as the most important aspect is a key factor that we need to take into account. We are doing this in Vodafone. Of course, the other important problem that you have in all these cases is that this is an evolving environment. I mean you are never protected, you are never 100% protected and we are learning also how even the attackers are using are using AI in order to be even more targeted and sophisticated on their attacks. Then it's an evolution that we need to have. Of course, on a gents model, we need to be extremely careful and of course I think that we need also to evolve on our side because our attackers are much more sophisticated than before using also AI techniques.

Guy Daniels, TelecomTV (12:38):
Absolutely, Didier, absolutely. Thank you very much. And Jakub, what are your views on AI in network security?

Jakub Votava, O2 Czech Republic (12:44):
Basically, it's more and more clear that the use the AI in security in general nowadays is a must, so it's coming into the, let's say, bad guys portfolio and it needs to be used and as well in our tools to be able to protect ourself against, again, these threats. We need to understand that the latest development in the telecommunication architecture are bringing more threats into our perimeter where in the past more IT related with all that containerization and the virtualization of our telecommunication workloads, the parameters is widening and as the year said, we do securities as a part of the architecture and design, but to be able to cover all these new attack vectors, the AI tooling is necessity to be in place.

Guy Daniels, TelecomTV (13:48):
That's such a critical area there. Thanks everyone for your views on security and let's move on and take a look at AI use cases to solve business and customer challenges. What are the key factors to consider when identifying the most promising AI applications? Are there any real world examples that you can share of successful AI deployments? And let me come to all of you on this one Yaku, can we maybe start with you first this time?

Jakub Votava, O2 Czech Republic (14:17):
Sure. I would say here o to Czech Republic, we've got quite a long journey with automation and AI based automation use cases which we use in the real world. Most of them are in the area of the customer care. We have started long ago with building the tools which are helping our internal staff to serve the customers better, but currently we are in a situation where we are having fully operational AI agents talking to our customers either using the web chat or using the voice conversation with our call centers and being able to answer the customer bill related questions, customer product related questions, really having the customer context and being able to answer the questions in those contexts. But back to the network, what we have achieved so far, and we know that there is the long journey ahead of us, but what we've been able to build so far as kind of the AI ML enabled detection mechanisms mainly of the network anomalies because you've mentioned guy that there are tons of data in the network space and frankly nobody's able to explore them all in the real time and this is something where AI can help, where it can spot on the stuff which is going out of let's say expected boundaries and bring that to the view of the people right now, which are then able to act on that.

(16:09):
So that's where we are. We know that there is a lot ahead, but we are building that step-by-step and that's how we have started.

Guy Daniels, TelecomTV (16:19):
That's very good to hear, Jakob, thank you very much. And Didier, what can you add for any real world successful AI deployments that you can share?

Didier Clavero, Vodafone (16:29):
Okay, yeah, I think that as Jakub said before, I think that many of the AI use cases that we were doing, I dunno one or two years ago, were focused more on the customer care and how to treat better customer faster and given them the info even faster than before, but I think that now we are moving to other aspects and we are moving to full automation on our network and for example, I mean there are different examples that I can say with you. I mean there is one very important that even was presented on the Mobile World Congress. That is field maintenance assistance. That is an AI concept that is helping our engineers outside. It could be engineers going to our network elements or it could be also engineers going to the customer house to repair the CPE to repair the fixed line and you have an AI assistance that if you are sending the photo about what is the situation of the modern, what is the situation of the cabling and this kind of things is giving you the solution remotely and just providing the engineer with the solution that okay, you need to change this cable, you need to change this fiber, you need to change the router because we think that this is the only problem that we have and what we have seen is a very, very positive result in terms of repetitive incidents and in terms of how fast our engineers are solving the incident that we have.

(18:01):
Then on top of that, and of course this is a very, very important savings for us, but it's also a very, very good customer experience because we are solving on the first interaction with the customer. There are other examples, energy management is key. I think that when we talk about sustainability in our world and when we talk about ESG, I think that energy management on how we can be efficient on our energy consumer is key for us, especially on the radio sites and this is something that we are using because we have seen extremely good results with more than 25, 30% improvement versus previous experiences that we have. And then finally, I would say that all related to incident avoidance. I think that it's no longer about detect an incident correlate and probably send to the right engineer on the second line or on the field maintenance.

(18:58):
Now it's more about how we can avoid incidents happening on our network or how we can act before the incident is happening or reaching the customer. In that particular case, we are working in two areas, incidents by itself, but also change management and how we can improve or how we can help our engineers when they are doing changes to avoid mistakes and this is something that we are working even with different CSPs because we think that in this particular area it's not about one CSP, I mean working alone, it's about the industry working in a common solution that we can apply everywhere.

Guy Daniels, TelecomTV (19:38):
Yeah, thanks Didier. Collaboration is so important. Good examples there. Thank you very much. Now we spoke briefly earlier about people and skills, so let's pick up on that because to successfully leverage AI in an automation journey, I'd like to know what initiatives your undertaking to upskill your internal teams and also to attract the best AI talent in, let's face it, is a very highly competitive market Now, Didier, can I come to you first and find out how you are going about this?

Didier Clavero, Vodafone (20:13):
Yeah, sure Guy. I think that on our side in Vodafone we are working in two different areas. One area is of course we have digital experts to have people that are software developers that they know how to automate, they know how to code, and then that they are giving us the base for what we need to do and how we need to do, but only with this kind of people. I think it is not enough. I think that what we need is to also have the people that are really doing their work involved on these automation activities involved on this digitalization of the network and then what we are also doing is to use the concept of cities and developers. We have people that are working on our first line teams, second line teams and people that are touching the network every day. Then on top of that, they are receiving and we are re-skilling them in order to have this kind of digital capabilities, this kind of software developers capabilities because what we realize is that is the best way to really develop use cases that are valid and are providing benefits because these are the people that they know the pain points that they have, that they know what activities are more painful, more long, I mean longer to do and this kind of things.

(21:25):
Then combining these two options, I think that we're achieving in all the different markets, very good results. I think that on the other side is giving us the possibility to also move people that due to automation, you need to probably to move from one area to the other because now this activity is no longer required, but then also the teams are engaged on this because they see this as a path of development from themselves. Then I think that is extremely important to have this path of development for the people because we need them in order to identify what is more relevant and then we need them in order to do other activities in the future when these activities no longer needed.

Guy Daniels, TelecomTV (22:06):
Absolutely. Thank you Didier. Jakub, let's come across to you. You spoke earlier about the importance of people here, so how do you go about training and retaining and attracting the right skill?

Jakub Votava, O2 Czech Republic (22:24):
The initiatives has to be company-wide, so if we are talking about the AI and the AI mindset, nobody should stay aside. So we've launched several initiatives in the company which basically supports and encourage people to use the AI tools. We are having the internal competitions about the best use of AI in a daily work. Even for the, let's say the non-technical, non-technical positions. What we have created and we feel is very important for people's safety. We've created the AI sandbox where you can use the different models in a safe environment, not to be so much feared about using the internal data because the data are bounded to our environment and in parallel with that, we are creating the group of the, let's say, extremely skilled AI data analytic experts, which are coming from the universities to help the rest of the company wherever there is the needs to put some of those AI initiatives into the real world. So they are the helping health and extremely skillful, but as I said, the most important is that the old companies somehow involved and anyone can bring up the idea how to use the modern technologies to at the end of the day serve our customers better.

Guy Daniels, TelecomTV (24:11):
Let me just move on now to partnerships. You spoke about collaboration earlier, but leveraging AI capabilities does require partnerships. So what does a successful operator vendor partnership actually look like?

Didier Clavero, Vodafone (24:28):
Yes, for sure. I think that in terms of partnership, I think that there are different angles. I think that for sure, as I said before, I think that the partnership across different CSP would be great. I think that in this concept of how to automate, how to reduce incidents and these kind of things, we are all together in the same target with the same. Then if we can create all the CSPs together, something working together that is allowing us to be faster and to do this for our customers better, I think that is an industry common goal that we have and we can do this much more than what we are doing today. Then of course it's about partnership between CSP and vendors in that particular case, Nokia and the other vendors. I think that extreme, extremely important partnership because at the end, I mean for many of the topics that we are talking about, I mean for how to be able to classify our networks with more in service, upgrade with more transparent or faster ways of deploying new software, deploying new releases. I think that the work that we need to do with Nokia and other vendors is key because if not, it's going to be impossible that we can achieve the level of automation that we want to achieve and that we can be as fast as we need. Then from my point of view, these two, I mean partnership are key CS, P and vendor for sure, but also across CSP in order to create large language model common from all of this.

Guy Daniels, TelecomTV (25:57):
Great. Thanks very much Didier and Jakub, your views on successful operator vendor partnerships?

Jakub Votava, O2 Czech Republic (26:03):
Definitely for me, let's say the basis for the successful operator vendor partnerships lays in let's say the alignment of the strategies and visions. So really to be successful in the long term, we have to go the same journey. We have to have the same goal. That means same understanding how to achieve in that case the AI driven operations. So that's for me the alignment on those is the basics and I would say what I do feel that that talking with the Nokia guys, this is something that really, really resonates.

Guy Daniels, TelecomTV (26:42):
Well. I've got a final quick question. I'm going to ask each of you. If we look forward, say three years from now, what is the one major AI focus breakthrough that you or your company would like to achieve? Tricky question, this one, but Jakub, let's come to you first.

Jakub Votava, O2 Czech Republic (27:03):
Yeah, and it is guy, three years from now talking about the AI is a lot, so if we consider how quickly the pace is going there, so for me, I will turn to the basics and we've touched some of them already, so in three years from now I would like to have based on the AI tools, the zero service disruptions due to any kind of the configuration error and if there is a service disruption to be able to detect and do all the necessary troubleshooting resolution of the error using the AI tools within three minutes. So that's basically my goal.

Guy Daniels, TelecomTV (27:53):
Didier, I'm coming to you for final comment today. What AI breakthrough would you like to see?

Didier Clavero, Vodafone (27:59):
Probably something aspirational. That is zero downtime for our customers. Then all that we have been discussing about, okay, how to automate the network, how to prevent failures, how to react faster, how to avoid problems with changes on the network and avoid human errors mistakes. I think that the aspirational goal would be let's have zero downtime for our customers on our networks because you design in a secure way because you design in an automatic way. Having in mind and taking into account all the resilience that you need to have on the network because when you have problems you are anticipating and you are solving faster, but at the end, the result is that in a world that will be much more connected with all the new 5G services, with all the new APIs and the new things that we probably are not even dreaming today, how we can ensure that we have zero downtime for our customers. This could be my goal.

Guy Daniels, TelecomTV (28:58):
Yeah, that's another goal that is well worth the effort working to achieve. Well, thank you all very much indeed. We must end our discussion there, but thank you so much for taking part in our program today and if you would like further information on the topics covered in this discussion, then please follow the links in the text below this video. For now though, from all the team here, thank you very much for watching and goodbye.