Sean McManus, TelecomTV (00:07):
Welcome to the Future is Autonomous Nokia's Zero-Touch Network Revolution. Hello, I'm Sean McManus. To address the rising complexity of their networks and services, CSPs are increasingly adopting the autonomous networks model. It has the potential to not only enhance network performance and fault management, but also operational efficiency and customer experience. In our discussion today, we'll explore the drivers for autonomous networks, the challenges in implementing them and the role of ai. So let's meet our panel. Joining me today are Daniel Askeroth, chief Technology Officer, network and Infrastructure at Norlys. Max Gasparroni, chief Technology and security officer from Fastweb/Vodafone and Erez Sverdlov, VP cloud and network services and market leader for Europe at Nokia. Hello everyone. Good to see you all and thank you for taking part today. To start us off, what do you see as the main business and technical drivers accelerating the adoption of automation in networks? Erez, can I start with you for Nokia's perspective?

Erez Sverdlov, Nokia (01:15):
Thank you for the question, Sean. Consumer enterprise needs are rapidly changing with expectations for faster, more seamless services. To meet the dynamic requirement, CSPs must deliver what we call zero X experience, zero weight meaning instant order fulfillment, zero touch assigned to automated delivery, zero travel, ensuring flow of service and zero trust bid on strong foundation of security. That's where autonomous network comes in using TM forum autonomous network maturity model. As a foundation, Nokia offers our customer a multi-domain strategy. IT integrate powerful combination of security orchestration, assurance, analytic and AI to simplify complex network operation and deliver the fully autonomous zero experience. Our goal goes beyond automating simple tasks. We are working towards building network that can self-manage, optimize and repair themselves across all domain, radio, transport, any cloud while focusing on security and automation from a physical infrastructure to a cloud-based application. In short, it's about creating simplest, efficient networks that can also open new revenue opportunity for the CSP. Achieving true network autonomy means simplifying the complexity CSPs face, improving network efficiency and resiliency and unlocking new revenue strengths.

Sean McManus, TelecomTV (02:53):
Thank you Erez, interesting that this is in part a response to market demand and the way of meeting those customer expectations. Max, let's come across to you next. What drivers are you seeing that are accelerating the adoption of automation?

Max Gasparroni, FastWeb/Vodafone (03:06):
Hi Sean. For FastWeb/Vodafone, we see on the business drivers two categories. First is to dramatically improve the services that we can offer to our customers, like improve network quality and also the understanding the quality delivered to each specific customer, improving resilience, fall detection, fall prevention. And the second category is on the speed and volume of which will be able to introduce new services with open APIs to quickly integrate telco infrastructures with our B2B and B two B2C customers. On the technical drivers is automation AI is fundamental across all end to end cycle starting from planning, design, implementation, then performance monitoring optimization operations. Without automation AI simply we cannot cope with the increased complexity of the networks and the services, the cyber threats and most importantly in achieving the advances in network performance. Just like in healthcare for example, where AI is transforming the diagnostics treatments in telecom, it can unlock a new level of efficiency and service quality if implemented with purpose and precision.

Sean McManus, TelecomTV (04:32):
Excellent, thank you. So it is quality, agility and just a sheer need to use automation to meet today's needs. I guess tell us Daniel, what are your thoughts on the drivers for autonomous networks? Well,

Daniel Askeroth, Norlys (04:46):
I think our drivers are very similar to Max is because what we are looking at is very much cost quality and security in the end. So I mean from a cost perspective, running big operations and so on in Denmark will give you a cost issue. So the more we can actually automate that will simplify our operations and actually make us capable of well earning money in the end and also becoming national autonomous because I think that will be a more and more requirement going forward quality. I think then that will also be a big part of it because the AI can learn a lot of things and can also prevent things from happening towards our customers. In Denmark you have very high expectations from the customers on what they expect from us. So we are constantly trying to improve our network performance and a bit contradictive maybe. I think the customers want cheaper, simpler solutions, but at the same time they want a tailored experience and I think this is something that the AI will be able to bring us or at least that's of our hopes for it. And then in the security part, the fewer people we need in the more secure we can make this.

Sean McManus, TelecomTV (06:05):
That's great, thank you. So even as customer expectations become more demanding, automation gives you a way to satisfy those needs. I guess a lot of consensus in those answers there. Now that we know what's driving autonomous networks, let's talk about the destination and the journey to get there. Now many CSPs have high expectations for autonomous networks. What's the biggest gap between their aspirations and the current reality and what factors contribute to this difference specifically in Europe? Erez

Erez Sverdlov, Nokia (06:36):
Thank you. I think this is a great question, Sean. Right now the industry is sitting at level 1.5 to two when it comes to autonomous network maturity. When we talk to CSP, one of the things we are focused is helping them benchmark the network's automation and high maturity level. And when I look at Europe, European CSP are taking deliberate and I would say very smart and very wide approach of building fully autonomous network in stages. A recent TIM informed benchmark study gives a good picture of Europe's position in autonomous network landscape. When we ask about their strategy, 32% 72 of the European and CFP said they are following a gradual continuous improvement model. Only 28% described the plan as ambitious and for a full transformation compared with that. For example, if you go to Middle East where over 50% responded to pushing aggressive autonomous strategy, repeat operators are balancing innovation with operational realities.

(07:45):
They face existing system complex architecture and a lot of regulatory pressure. Rather than chase a hype, they're prioritizing use cases that show clear RI especially around operation and customer experience. What helps close the gap, it's giving the operator tools that fit into their current reality with that overwhelming them. That's how we close the gap for our customers. CSP are starting from a different point of the autonomous network journey with varied goals and speed in our middle solution we meet CSPs when they are in and help them move forward without adding to the complexity of hybrid networking. Most recently with a recent introduction of network fabric, it's a connectivity tissue that links data AI and automation across our domain For this gradual approach is the way currently we see the world is.

Sean McManus, TelecomTV (08:46):
Thank you Erez. Daniel and Max, let's talk about where you are on your journey towards network autonomy. What are the biggest challenges you are currently facing with network automation and how are you addressing them? Daniel, let's start with you.

Daniel Askeroth, Norlys (08:59):
Yeah, I think the biggest obstacles so far is probably the data quality and many of the legacy systems and also the process maturity across the ESP world. I mean we are coming from a world where there's a lot of m and a activities and so on, so we are the result of many purchases across the Danish market. On our side we are reorganizing as part of becoming new lease the first Danish pure Danish company carving out hotel companies. So we are actually investing a lot in new systems but also looking into processes and also looking into the data quality and I think actually data quality and the process would be the biggest obstacle for us.

Sean McManus, TelecomTV (09:44):
Thank you. Daniel. Clearly legacy is a huge challenge for operators. Very few of them have the luxury starting from scratch. Max, let me come to you now. What are some of fast web vodafone's challenges in implementing automation in your network or what are you seeing more generally?

Max Gasparroni, FastWeb/Vodafone (10:01):
Thank you Sean. In FastWeb/Vodafone, we are at the beginning of the automation journey as we only recently acquired Vodafone, Italy and we are undergoing a radical modernization and automation program with reference to the TM forum excellent autonomous network level framework. I would say that we are between level one and level two roughly much closer to level two actually. And for some aspects I would say we are close to level three. What are the biggest challenges? First is people's mindset. As the advances of AI have been wrongly touted as a mean to achieve FDA efficiencies which creates resistance. Instead we frame internally automation as a tool to improve quality, speed, scalability and relieve our people from the most tedious and boring jobs. Second is the complexity of the legacy systems platforms processes where it will certainly take time to unravel and modernize. What are we doing internally we have empowered our teams to challenge the status quo.

(11:15):
They identify inefficiencies and the pain points bottom up and then we discuss in order to prioritize the use cases to implement based on cost, benefits, outcomes. The execution is led by internal teams where they are supported by either by our newly launched AI factory fast web or by selected third parties. But in following this path we have three key principles. First is to have open APIs to ensure interoperability. The second is that we want to own the intellectual property so we can retain control over strategic assets. And last but not least is to ensure to have a shared data infrastructure working closely with a T on unified inventories and data fabrics. And in order to do that we are heavily investing in people re-skilling.

Sean McManus, TelecomTV (12:13):
Thank you Max. Interesting to see the mindset dimension come in there. There's always a huge amount of organizational change with transformations like this and I can see how that can be a significant challenge. Let's come across to Erez now. Tell us how does Nokia address the integration challenges inherent in multi-domain, multi-vendor network environments to achieve seamless network operation? There's a lot of complexity to deal with there.

Erez Sverdlov, Nokia (12:37):
Yeah, so I think the key is to make sure the different networking domain, radio, transport core and cloud work together seamlessly. To do this through AI powered observability, which give a deep visibility into the network performance and closed loop automation that automatically adjusts configuration. These technologies ensure that the inevitable sims between systems are invisible along for smooth and efficient operation without constant human intervention for the matter to this is of course security embedded into all network layer. The networks that Nokia built are secure by design, integrating orchestration, assurance, analytic and security to deliver AI powered autonomous network. As we understand the network and software and we have the right people with the right skillset to help our customers making us a trusted partner. In navigating this complexity of software integration, we have built a proven use cases to help CSP worldwide addresses the challenges and complexity of a multi-domain, multi networks.

Sean McManus, TelecomTV (13:51):
Thank you Erez, I think it is important to take that multi-domain view. If the goal is ultimately to achieve that multi-domain integration, it's good to start out. With that in mind, let's talk about security now, which we did just touch on. The telecom sector has grappled with a diverse range of cyber threats. As we know networks are growing more complex, threats are becoming more targeted, and yet at the same time regulations are demanding ever faster action. Why is it important to use telecom grade security tools rather than relying solely on traditional IT solutions, especially as operations become more automated? Erez, let me come to you for your thoughts first.

Erez Sverdlov, Nokia (14:28):
Thank you. So we believe security is a cornerstone of autonomous network security is non-negotiable in the autonomous network strategy. Instead of being an afterthought, we built it into every layer of the network. This includes advanced threat detection, real time monitoring and predictive security measures embedded within the network foundation by ensuring security is backed in from the start. Nokia aims for the network that can self monitor, detect vulnerabilities early and self feed itself without human intervention, keep the operation secure a critical requirement As our industry involved, the extended detection XDR security market is projected to grow by 150% from 955 million in 2022 to 2.4 billion by 2027. According to industry estimate, our net guard cybersecurity dom is in remote winning security software. So prebuilt 5G use cases for CSP and critical infrastructure enterprises. It was a real time visibility across networks, cloud infrastructure and even endpoints. IT unifies security control points, also security telemetry and provide analytics and operation in a single view all to drive the resiliency of the network.

Sean McManus, TelecomTV (15:57):
Thank you, Erez. Nokia has clearly identified the need for security to be built into those foundational technologies. Max, let's come across to you. Tell us a bit about security tools that you are implementing in your network.

Max Gasparroni, FastWeb/Vodafone (16:09):
Yeah, Sean Telco for us is a critical sector because a cyber attack on a telco company may have very big impact ramifications on many other sectors like energy, healthcare, transportation, et cetera. And regulation is rightly so increasing. So at European level for example, we are seeing needs two CR dora and the same and there are a couple of frameworks for this specific for the Italian markets. And securing a telco is much more complex endeavor than securing a standard IT environment. For example, we have loads of components, so not just servers, but we have run nodes, routers, core connect or castex as was saying. And then in protocols we have loads of legacy protocols like SS seven that have not been designed to be secure, but unfortunately they are still in news. We have complex supply chain and then we also have the need for specialized solutions like specific firewalls, specific hardening and specific performance requirements like minimum latency, high throughput in order to identify threats in real time and mitigate them.

(17:24):
How can really handle this complexity. But we are encompassing a multiplicity of initiatives. For example, we are expanding the perimeter of the assets under supervision and the procedures that we have under control. We perform sector specific risk assessments, rigorous implementation, cybersecurity controls, and we have organized ourselves internally into a three level with clear separation to ensure that we have a strict revision processes and corrective measures are implemented and validated by the audit, the internal audit teams. And so in terms of guidelines, we have security data integration and correlation, real time monitoring and simplification. Last but not least is the training of internal employees is for us a paramount importance because as unfortunately history teaches us cyber incidents, most of the time were happened because they exploited the human factor and what kind of tools. So the deployment of advance and evolving tools is critical as the attacks are becoming more and more sophisticated and they're using ai. So our responses should be to monitor automatically the amount of type of traffic quickly spot anomalies versus what is the norm. So how are we doing that? Well, we starting with the integration of telco specific data AI training, so the signaling traffic run 5G logs, network device telemetry data behavior and then integrate it into AI models that are tailored for telecom threats.

Sean McManus, TelecomTV (19:18):
That's great, thank you Max. It's interesting, obviously the stakes are very high in telco as you identified, but there's also so much bespoke telco technology to deal with there, so it's interesting to hear how you are using AI to address that and attacks on it. Let's come over to you now Daniel, for Norlys' perspective on using telecom grade security tools over to traditional IT solutions.

Daniel Askeroth, Norlys (19:40):
Well from our perspective, I mean security is at the center of what we are doing. So everything we are looking at for new technologies, et cetera, we are doing the risk assessment as Max was also in on. And also we have a collaboration with Aris and the Nokia team as well to help us out with some of the security tools. Much of what you're doing is going to be impacted about the legacy because we need to understand how do we actually handle the legacy in these situations as well. So understanding that and also doing the training part because training of the people is going to be central in all of what we're doing. The reason for needing the telco grade tools is the fact that it's real time based. It is the anomalies we need to catch because it's not just an on and off, we need to be there 24 7 in everything that we do. So it's about how can we be proactively in what we're doing and so on. But to a very large extent it's about monitoring, but it's also about doing the risk assessment properly from the start. So we actually take in the measures when we design platforms.

Sean McManus, TelecomTV (20:49):
Thank you Daniel. I guess the need for real time anomaly detection is always going to drive a desire to use telco tools, but it is important to wrap training around it. As you said. We must move on to our next question. Now we've seen that AI and or automation go hand in hand when it comes to operational efficiency. How can CSPs apply AI ops to enable closed loop automation and drive down operational complexity and costs? Daniel, let's come to you for your thoughts on this one first.

Daniel Askeroth, Norlys (21:20):
Yeah, from our perspective in nor it's very much about understanding the inter domain part of the automation, how can we actually train the processes and how can we improve our data so that we actually have something end to end so that that's the way we can see that you can drive the AI ops going forward, but I think in the end it's going to be a mix between AI and machine learning for how we actually track anomalies, patterns, et cetera, across. And yeah, so I think it's about many of the standard things that we had also in one of the previous questions, it's about how can we train ourselves on processes and data. I think that will be the key for us.

Sean McManus, TelecomTV (22:03):
Great, thank you. Some great points there. Thank you. Max, what are your views on closed loop automation to drive down complexities and costs?

Max Gasparroni, FastWeb/Vodafone (22:12):
Well Sean, for us we see automation, machine learning and ai, especially generative AI as essential and transformative. We need to have the right team expertise and mindset in order to produce a compelling roadmap for each of our internal domains such as mobile fiber, converge operations, and then start implementation of the most compelling use cases first. The so-called low hanging fruit inside the company, it's critical to have common data storage and common data fabric as eventually the power of these use cases will be when we can get them interacting with each other. We have embarked on this journey. We are at the beginning, as I said earlier, we have implemented some automation with AI using in-house, for example, on the fiber rollout feasibility study in fast wave Vodafone, we're doing more than 50,000 of them a year and we switched from manual to automated feasibility and cost analysis and we are reduced by 60 to 70% the speed and the amount of resources involved.

(23:28):
And actually we have improved the accuracy of the results by more than 15% and we have set ourselves up for scaling to over a hundred thousand feasibility studies a year on mobile. Besides, we've been using what we call smart planning where when we have to plan network improvements and expansion, we take into account many dimensions, not just network congestion. We recently started deploying automation and AI for site upgrades. Acceptance, the results had been absolutely excellent, mind blowing over 90% back office reduction with 95% first time right and 99% accuracy. Then in-house we have deployed, we are using AI and automation in order to have ai, what we call it in-house assisted brain to continuously improve the wifi performance without disturbing the users when they are active. And actually we use it, the feedback to suggest the customer's corrective actions that they may need to take.

(24:34):
And also in the area of cyber, we started not just on tools like enhanced threat detection, automated incident responses, vulnerability management prioritization and et cetera. But also we are using solutions like real time adopt security policies and also to help producing, documenting and auditing the processes and the procedures that the teams needs to undertake. So this gradual and focused rollout of ai, as re said, is a gradual process will allow us to first implement this closed loop automation within the single domains. And once those domains are stable, we'll move through cross domain closed loop automation where the value compounds improving quality, reduce complexity, and driving down cost will then be materializing.

Sean McManus, TelecomTV (25:30):
Thank you, Max. It's interesting to hear how your vision is to have all of these use cases helping each other and working together and you're putting that foundation in place now even as these use cases are isolated for the time being. Let's come over to Erez now for some final comments on closed loop automation.

Erez Sverdlov, Nokia (25:47):
I think the following party, Nokia AI Ops solution integrate AI into the network operation to automate task, predict issues and proactively manage networks. Nokia autonomous network fabric is a key enabler of this AIOps driven cloud group automation. It provides a unified platform for managing and orchestrating automation across multiple domain and vendors simplifying integration and reducing complexity. By implementing this AIOps driven approach, the CSP can achieve significant cost reduction through optimize ization. I think this is, as Max said, reduce manual intervention and the faster problem resolution, all of this. And in the same time, they can improve the customer experience to enhance network availability, faster service delivery and proactive issue mitigation.

Sean McManus, TelecomTV (26:43):
Thank you Erez, I think it is important to always keep in mind that ultimately it's about the customer experience and it's about the services that are delivered. Finally, let's come to perhaps the hardest question of them all. Will we ever see fully autonomous networks? Who wants to tackle this one? Max, let's hear your views first

Max Gasparroni, FastWeb/Vodafone (27:03):
And I would echo the words from Howard Watson who spoke to telecom TV the other week at DTW that I don't expect full autonomy in any short time horizon probably in the next 10 years. But what's really important is to be aggressive and committed on the journey as we will not move from current level, let's say 1.8 to level four through one giant step instead by hundreds of steady improvements. And to make this real leadership must set the tone we need to create a culture where automation is seen no as a threat, but is an opportunity to evolve. So upskilling becomes central and the teams must be empowered with the knowledge to implement these use cases effectively. We also need to stay focused in performance optimization. For example, if we don't define a minimum viable data set for our models, we risk of having inefficiencies being buried by tons and tons of data that are not effective and not relevant, that will slow us down rather than move us forward. So will we ever see full autonomous networks? Yes, I am very confident, but not tomorrow. What we'll see definitely is a smarter, faster, and more responsive network shaped by every step that we take on the road to autonomy.

Sean McManus, TelecomTV (28:33):
Thank you Max. Daniel, let me ask you the same question. Will we ever see fully autonomous networks?

Daniel Askeroth, Norlys (28:40):
It's one of the more difficult questions. I even asked my head of strategy and my head of operations to get their views on it and I got their diverging answers. So my head of strategy starts saying, no, we will not. And my head of operations says yes, which surprised me a lot. I must say, I think there are some perspectives into it. From a strategy perspective, you need to understand the accountability and because in the end there will always be someone accountable for the actions that we take, especially when we are going towards higher security classes and so on. So there will always be some kind of human intervention. My operations guy, on the other hand, he said very much similar to Max's story. It's going to be small steps, many of them and a long journey before we see the complete journey towards the automated. My take, I think it'll be a long journey to get there. I believe we're going to be there, but I think there will at least within the foreseeable future be a requirement that there is a human involved to intervene and take the responsibility. In the end, I think the capabilities will be there. I think we're probably blocked by other things such as ethics, regulations, et cetera.

Sean McManus, TelecomTV (29:59):
Thank you. And finally, Erez. Do you have any additional observations on whether we will ever see fully autonomous networks?

Erez Sverdlov, Nokia (30:07):
I agree with my colleagues. I think the autonomous network as a goal, but getting the as iation before with a step-by-step journey. Our vision or my future vision is to see optimize, sell, fill and deploy services with none to minimal human input, but we have to get them. The industry is definitely in the right steps together with a significant progress with ai, machine learning and automation. Every CSP as we spoke at a different stage of the cloud adaptation. So we are basically starting from a different starting point towards a common goal. Our integrated approach automates everything across radio to transport, to code, to cloud networks, and security of course is a top priority. We are currently running about 500 telco projects globally using AI driven security. If you just take account in Europe for example, we are partnering with a major CSP to train AI models for network efficiency and the sustainability in North America. We are working with a large CSP to automate flight operation using orchestration and assurance. I think Nokia, with our deep expertise in both software and network, we can meet our customer exactly where they are and move them towards the future of autonomous network. The journey going to be challenging and require integration of advanced technology, including various AI solution and modernization of the legacy system. I can summarize and say that in Nokia we are committed to help CSPs along their journey, both a fully autonomous network.

Sean McManus, TelecomTV (31:52):
Thank you for those last comments Erez. Clearly it's going to be a long journey to achieve fully autonomous networks, but there are wonderful opportunities to deliver business benefits along the way. Well, we must leave it there as we are out of time. It has been a very interesting and enlightening discussion. Thank you all for taking part and giving us a clearer insight into the challenges and drivers of autonomous networks. And if you'd like to find out more, please follow the links on this page to learn about Nokia's zero touch revolution. Thank you for watching and goodbye.