Autonomous Networks Now

AI hype is everywhere at Mobile World Congress, so we went looking for something rarer: real operational proof. From the Ericsson booth in Barcelona, we sit down with Ibrahim Eldeftar, who leads Cognitive Network Solutions product and portfolio for telco AI software, and Claudia Muñiz Garcia, Global Head of Sales for the same unit, to talk about what it actually takes to run autonomous networks.

We break down the autonomy journey in plain language: moving from manual ways of working to systems that can sense, analyze, decide, and execute, with intent-based networking as the layer that helps unlock Level 4 autonomy. Ibrahim shares where carriers really are today (around Level 2 on average) and why many are publicly targeting Level 4 by 2028 to 2030. The driver is network complexity plus rising expectations for “networks for AI” that can support new AI workloads and millions of connected devices without brittle operations.

Claudia explains why Ericsson is being recognized for 5G RAN automation platforms, from standards compliance and security to commercial deployments and scale, and how an open rApp ecosystem accelerates innovation. Then we get into the part operators care about most: outcomes. You’ll hear concrete results from trials and deployments, including major OPEX efficiency gains, fewer issues through anomaly detection and root-cause approaches, spectral efficiency improvements, and a standout uplink story where AI optimization drives meaningful uplink quality and throughput gains.

If you care about network automation, 5G RAN, telco AI, rApps, and the practical road to Level 4 autonomous networks, hit play. Subscribe, share this with a network leader on your team, and leave a review with the one autonomy question you want answered next.

Transcript

SPEAKER_00 0:01

And I'm here at Mobile World Congress with Ibrahim and Claudia from Ericsson. Guys, how are you? We're doing fine, fantastic. Great to see you here. Really exciting time to be in the industry, exciting time to be at Barcelona. You guys have so much going on. Before that, let's start with introductions. Ibrahim, your role and your team within Ericsson.

SPEAKER_02 0:20

So I'm heading the solution line for Cognitive Network Solutions. So that's basically product and portfolio management for telco AI software, our apps, as well as for services.

SPEAKER_00 0:31

Fantastic. And Claudia, we've spoken before. Introduce your role and your mission within Ericsson.

SPEAKER_01 0:36

Yes, I'm Claudia Muniz and I'm Global Head of Sales and Cognitive Network Solutions. And essentially I help commercialize all the services and software that IRIM develops.

Where Carriers Are On Autonomy

SPEAKER_00 0:45

Fantastic. So Ibrahim, you know, lots of conversations and demonstrations around autonomous networking and the road, no pun intended, to fully autonomous level four, level five, I guess you'd call it autonomous networks. You have a unique bird's eye view of the industry at Ericsson. Tell us where are the carriers on this journey?

SPEAKER_02 1:05

So uh I would say like the autonomy journey is a journey that takes uh the carriers through the different phases of evolution of network automation. So moving from the full manual ways of working and through in their operation to a fully autonomous operation that it like you know spans across the awareness, the analysis, the decision making and execution phase, and of course intent that comes on top. Uh so far in uh basically what we see, CSPs right now are early in their stage. So the average industry is between 1.8 to 2. So very early actually to reach to level 4 and later on to level 5. But we see a lot of interest and ambition and public, you know, uh commitment that CSPs move to level 4. For example, we see some of them saying like we would like to move by 2028, some of them are saying like we would like to move by 2030. And uh, we are very excited about this journey because we have developed the capabilities that can help them uh get through that. Uh, very quick reflection from my side, why I see actually this happening is networks are getting complex. So CSPs are realizing that that's not something that humans can manage anymore. Second thing is that there are more demands on CSPs now and the connectivity, making the connectivity viable and critical for AI workloads. How we call it like, you know, networks for AI is top of mind. And in order to run this network for AIs, we need to make sure that networks are ready and autonomously actually running for such kind of workloads of the future. So I'm glad actually to see the excitement. All my discussions with customers uh like for the past days have been around like how we achieve network autonomy, how we can like AI help us with this journey.

Why Ericsson Leads 5G Automation

SPEAKER_00 2:50

Fantastic. Well, it's definitely the most exciting time in the industry, and this seems to be the topic de jour here at MWC. Everyone's talking about autonomous networking. And Claudia, I I noticed your ABI research has named Erickson the number one provider in their 5G RAN automations platform. First of all, congratulations for that. I know you dedicated your career onto this mission. You know, what are some of the key findings that um helped deliver this award, this recognition?

SPEAKER_01 3:17

Yes, I think it's just rounded on facts, right? We are the number one player in the industry in terms of uh contracts, in terms of commercial deployments, in terms of standard compliance, in terms of security features, in terms of scale. We have the largest Arab ecosystem in the industry. So we have 89 members, out of which 20 are CSPs, and they all get access to our developer portal where they have everything that somebody will need to develop Arabs. So there is a GNAI development assistant, there are Arab templates, there is a forum where they can help each other. And believe it or not, these are exactly the same tools that we in Cognitive Network Solutions use to develop our Arabs so that they are truly open. So, in addition to that, the other reason why we got this award is our vision around our own Arabs portfolio. So we have structured our portfolio as a true cognitive system where we have Arabs intended for awareness, analysis, decision, and recommendation, and we have a very solid vision toward to go towards intent so that we can truly materialize the autonomy level four. So it's a great award, it's a great recognition, but it's even greater to be hand in hand together with customers delivering value in the field. So I hope you saw the news. We are live with VISCOM, and we also recently assessed our Arab portfolio, something we're very proud of, using the TM Forum unlende methodology. So then we determined that in the Arab portfolio that we have today, with our Arabs, we can achieve between 3.5 and 3.9 autonomy. And then going on when we embrace intended is when we reach the magic level four.

Field Results On Cost And Quality

SPEAKER_00 5:05

Wonderful. Well, shout out to you and the team, especially here in Malaga and Spain. Tremendous work across Ericsson. And Ibrahim, you know, tell us about some of the real world impacts. We're already talking about reducing, optimizing some of the CapEx requirements, optimizing network expenses for the carriers. I mean, CapEx uh uh uh obviously can be optimized, but what are you seeing in the real world? Real deployments, real trials, real testing from your many, many telco partners.

Customer Excitement And Wrap Up

SPEAKER_02 5:34

Well, I mean, uh we are we are quite uh uh serious about this journey in a way that we would like our customers to understand that moving in autonomy levels is not just like scoring a higher number on a scale, it's actually crossing an economical threshold. So this is built on what we call the business levers, the value levers, business value levers like opex efficiency, capex efficiency, sustainability, running a greener operation, wow, uh uh user improved customer experience and user experience for subscribers, and having an opportunity to protect them as well as to to uh create new revenues, revenue opportunities. So, what we have seen in the field, for example, on the uh um opex efficiency side, we have seen that with the sales shapers, uh apps that we have actually like you know, tested with one of our lead customers in the US, we have seen up to 75% efficiency in the process of physical optimization. So that's like actually a great, a great uh uh value that we have seen in the field. We have also seen uh with different value with different R-App family that we call the anomaly detection and the root cause explainer, and also on the general optimization uh closed loop, we have seen up to 60% less issues with problems, 42% more efficiency for teams who are working on this optimization. So that's when it comes to the opex efficiency part. On the CapEx efficiency, we help it actually customers to see with cell shapers and also with the anomaly uh uh family that we just explained, uh uplink is spectral efficiency and downlink is spectral efficiency. We see number we saw numbers between 4% for downlink, up to 6% and 9% in some of the cases for uplink. And we understand uplink is extremely important for that. And uplink, uplink being a very important uh uh uh quality for the users and for workloads in the future as well, and many AI devices connected to the network. We have a dedicated R app for that, what we call AI uplink interference optimizer. And this one, we just actually finished its testing with one of the elite customers in Japan. We have seen 3x improvement in the uplink uh network quality. We have seen improvement, the spectral efficiency of 6%, we have seen improvement in uplink user throughput by 11%, and that all contributes to a better user experience, less churn for our customers. And so we can achieve this as you move with the different autonomy levels, you actually cross different economical thresholds. So it's real. I mean, like an AI is delivering value, and the technology has actually been proven in the field.

SPEAKER_00 8:07

Well, in a world so much marketing hype around AI, it's great to see so much practical, real industry value being delivered. Your customers must be thrilled, they must be so excited. Absolutely, yeah, absolutely. So much innovation. Yeah, yeah.

SPEAKER_02 8:20

We are both excited about that, but but you can see it actually in the eyes of the customers. I mean, like when they are here and we are engaging in the discussion, you cannot like immediately see how they react, how we would like to get action, how can I get my hands on this technology? It's a very, very active discussion in the past uh two days and also in the morning as well. Wonderful. Onwards and upwards, congratulations again. Thank you, thank you, thank you, Ivan.