6G Beyond The Pipe

6G is closer than most people think, and the biggest surprise is that the headline might not be “faster.” We sit down with Mats Karlsson from Ericsson to talk about the move toward physical AI, where networks help systems sense and act in the real world. When robots, vehicles, and digital twins depend on connectivity for safety and performance, “more bandwidth” stops being the product and guaranteed outcomes become the real promise.

We unpack what outcome based services actually mean in practice: collision avoidance, factory uptime, immersive experience quality, and other measurable KPIs that enterprises can justify paying for. That naturally leads to the toughest question for telecom operators and service providers: monetization. Matt explains why the business model has to evolve along with the network, translating intent into offerings, pricing, and even revenue sharing in real time, while still being able to prove the network can fulfill what it sells.

From there, we get practical about AI in telecom, OSS/BSS transformation, and where ROI shows up today. The message is blunt: don’t start with AI, start with trusted data. We talk about common OSS and BSS pain points like siloed datasets, uneven data quality, and limited end to end visibility, plus real examples of value like revenue assurance, billing anomaly detection, predictive operations, and faster root cause analysis. We also dig into agentic AI and why industry collaboration through TM Forum and open standards is key to making autonomous networks work at scale and unlock new revenue streams, not just cost savings.

If you care about 6G, autonomous networking, AI in telecom, and the future of outcome based connectivity, hit subscribe, share this with a colleague, and leave a review. What outcome would you pay for first: uptime, safety, or experience quality?

Transcript

Welcome And Guest Introduction

SPEAKER_01 0:01

Hey everybody. Fascinating topic today with one of the true innovators in the telecom and wireless space with Erickson, talking about how 6G is on the horizon and uh super exciting to chat, Matt's Carlson. How are you? Good. Thank you a lot, Evan. Well, thanks for being here. Always great to catch up, particularly on this topic. Uh, before that, perhaps introduce uh yourself and your team within Ericsson. Of course, Ericsson needs no introduction, but what are you up to these days?

SPEAKER_00 0:31

So uh I am responsible for uh BSS and OSS within Ericsson, within the Ericsson Group.

What Makes 6G Different

SPEAKER_01 0:38

Fantastic. You have many responsibilities, but um, one of them is looking at what's ahead for 6G, which is only a few years away. Um so why are we talking about it now and what what makes it fundamentally different from the 5G approach?

SPEAKER_00 0:55

I mean, one fundamental difference is uh the movement into uh physical AI, where you're not any longer just processing data or sending pipes through a pipe, but where you're actually having use cases where you have sense and acting of objects in in your physical network. That could be robots, vehicles, digital twins. Uh and I think this is really a fundamental shift, so say, that we are not the networks are not longer a pipe. It becomes like an essence in a cyber physical world, so say.

Outcome Based Monetization For 6G

SPEAKER_01 1:29

Yeah, and physical AI is such an exciting space. Every day I'm blown away by new use cases and videos of physical AI in action, so many different use cases, but it but a challenge. There's always been a challenge monetization new generations of connectivity. We've seen it with every generation, but how does 6G perhaps change that equation?

SPEAKER_00 1:51

Uh I think when we are transitioned to physical AI, we also need to transition to being able to have measurable outcomes. I mean, it can be collision avoidance detection, industrial uptime, immersive experience quality, or safety guarantees or uptime requirements in a factory. And of course, uh, enterprises and developers will pay for the outcomes, not just for more bandwidth or the latency. And that of course changed in the whole conversation around monetization. And monetization must follow the same pattern as the network. It needs to understand the intent and the outcomes that are requested and autonomously translate them into products, offerings, prices, and revenue sharing all the time in a real-time perspective. And I think the new value appears if we can, of course, uh translate the like the capabilities and the experience of a CG network into an outcome and then monetize it as an outcome. And of course, if you're going to monetize the outcome, you actually need to make sure that you you can actually fulfill the outcome. And then we are coming back into also like the autonomous network perspective, so to say.

SPEAKER_01 3:06

Very exciting. And of course, operators uh continue to invest in their 5G networks and in new capabilities, standalone and beyond. But how can they start preparing for 6G? What's a uh uh you know best practice there for getting ready?

SPEAKER_00 3:23

I think operators that have already deployed a 5D standalone network and are executing or next generation sharding and building architecture, they have, of course, a real head start. And I think the path for us is really clear, so say, from where we are today, the 5G sharding through early 6G preparation or 5G advanced, or what name we use for it, and to a full 6D deployment. I think what we are now launching is really to have something in place that can take you the full journey from today into the 60 era.

Pragmatic AI Starts With Data

SPEAKER_01 3:54

Really exciting. You can't talk physical AI without talking about data and AI in general, and of course, a lot of excitement about AI and telecom. Uh, some of its marketing, some of it's very real and practical. What is a practical, pragmatic approach to AI look like in the customers and partners you're working with today?

SPEAKER_00 4:15

As you said, I think there is a lot of hype, and and I think the real value comes when you can actually take a more pragmatic approach and making sure that you're using AI to fix the real problems of the world and not only trying to innovate on the most fancy stuff, but you're really to make sure that it's of uh daily use. And I think one of the things that I at least meet a lot of people that you don't really start with AI, you start with data, because I think with high-quality data, well-governed data, curated data, I mean, without it, AI doesn't really work, so say, because you you're kind of you need to have this trusted foundation based on data, so say.

SPEAKER_01 5:00

Indeed, yeah, it all starts with data, data quality, data preparedness. Uh so true. Um, now you you're really leading a team developing really novel approaches to AI for OSS and BSS. And I've seen some of your demos and real-world deployments, even. Um where does it start when it comes to AI and OSS, BSS? And what's what's your vision uh when it comes to you know the next few years?

SPEAKER_00 5:28

We have chosen really to combine AI and data, and I think it starts with data. It may sound a bit simple, but I think that is all often where a lot of operators uh either had to build a strong foundation or they are in in a lot of trouble. I think in a lot of the OSS and BSS environments, they're pretty complex, having a lot of data. Uh, and I think we also see that they have a lot of silo data, they don't have an end-to-end visibility of the data, they have uh maybe uh concerns about the quality of the data. And and I think before AI can deliver, I think it's important that you get like access to all your data under one roof. That doesn't mean that you have to move all the data into a data lake, but that you can actually see and act on the full set of data across the network. And that is of course to making it more useful. I mean, to actually that the right people uh with the right application can ensure that they can access it in a timely and trustworthy way, so and AI, in my view, AI is only as good as the data feeding it. So uh the first question a service provider should ask is is why uh what AI should we deploy? This is actually where do I have my data? Where can I trust my data?

Real AI ROI In BSS And OSS

SPEAKER_01 6:56

Yeah, so well said. Trust is is paramount. And um there are so many opportunities, but telcos are looking for tangible ROI, tangible business value. Um, what are they getting in terms of real value from AI today, not just in the lab and nice demos?

Agentic AI And Multi Agent Operations

Why TM Forum Collaboration Matters

SPEAKER_00 7:17

As you said, I think the ones that are seeing real results are the ones that are moving away from using AI as a as a use showcase or or or I mean to really having specific solutions to specific problems, so I mean if you can take that step, you're you're uh one step ahead, so I'd say. On the BSS side, I think there is a lot of practical things or practical applications that we already see, like reducing revenue leakage in in a system, uh catching billing or sharding anomalies, uh, improving how you can serve, retain customers, how you quickly can create offers to the market, how you can have guided selling approaches. Uh, and I think these are things that are already happening today, and I think we see now operators moving ahead to making sure they have tangible applications, for instance, and supporting them in the BSS space. And of course, revenue assurance alone can have a significant impact on the bottom line. On the OSS sides, it's very much about moving from a reactive perspective to a uh predictive uh perspective. So say, I mean, finding faults in the network before they have actually happened, so and of course, that means also that you can reduce time to rest solution and uh, of course, finding the fault, doing the root cause, and uh resolving uh the problem, and of course, uh get away from a lot of the repetitive and and uh I would say tedious tasks to trying to understand what where the problem is. And I think Agentic AI is of course very rapidly advancing in this world, and and I think we can now putting a lot of multiple agents in place, and of course, uh having aids collaborating with other aliens, collaborating with products, and I think that is also in some way uh becomes uh I think as a very good platform for innovation, so say, to really uh uh making sure that both the BSS and the OSS system becomes more efficient and more practical and more easy to use. Wonderful.

SPEAKER_01 9:20

So, one of the most fascinating that you're leading is autonomous networking. You're you're part of the TM Forum Autonomous Networks Mission Leadership Team. TM Forum, amazing group, by the way. Uh, incredible get-togethers uh within the industry. And why does that kind of collaboration matter so much when it comes to topics like autonomous networking and AI in general?

SPEAKER_00 9:43

Yeah, I uh to start, I think autonomous networks are now at a stage where there will be like an achievable milestone in the coming years. I think we see different operators, of course, at different levels and have different ambitions. Uh I think the technology foundation is really here. Uh and the question is, of course, I mean, how can we move as fast as we need, so to say. So it's more as I said, an industry question: how to make this happen, how to make it happen at scale. Uh I think an autonomy will only kind of deliver the full value if it's something that you can have end-to-end across vendor, across service provider, and across platform. And of course, it's everyone is continuing, or continue, I wouldn't say, but if everyone builds their own platform with a different architecture and different, for instance, simple definition of intent, you will not get an autonomous industry, you will get a fragmented base. So I think the whole setup with Team Forum Autonomous Network mission is of course to solve creating a common language, standards, and open architecture that any service provider can build upon. And I think that is what will unlock uh the scale in the industry.

Autonomous Networks As Revenue Enabler

SPEAKER_01 11:01

Brilliant. Well, such an important mission. Uh, it's great to see Erickson leading that. Um, and you know, beyond the cost savings, obviously, CapEx, NetEx, OpEx, um what else? What's the real business case for autonomous networking?

Closing Thanks And Farewell

SPEAKER_00 11:16

I I think what people tend to see that this is only for uh cutting costs. And and I think when we I mean, this is tying it back to where I started, I think a bit with the 6G outcome-based. I think in order to manage these complex services, in order to set them up, uh, in order to making sure that they are delivering the guaranteed outcome, you actually need an autonomous network. You can't really do it in a manual way. I mean, things are getting real-time, things are getting dynamic. So, autonomous network will also be a prerequisite to actually unlock the new revenue scheme. And of course, making sure that the services are resilient, and of course, making them connectivities experience outcome-driven. And I think the operators today, because there is operators that are quite forward-leaning that are deploying these principles already today, they can of course see the real result and the tangible result in enterprise innovation, operational efficiency, and faster time to market. Uh, but of course, it's a journey now, and I think the good thing is that uh I think there is uh also a very good momentum around the TM Forum uh maturity model that we're seeing. And I think uh I have a big trust that we will both have more efficient network and network that can bring uh new revenue streams.

SPEAKER_01 12:37

Well, here's for that, and congratulations on this important mission and leading the charge uh at Ericsson and within the industry. Uh onwards and upwards, Matt. Thanks so much for this update. Yeah, and thanks a lot, Evan. Thank you. Thanks everyone for listening, watching, and sharing the episode. Until then, take care.