Liberty Global is a leading multi-national broadband, video, and communications provider operating major European brands like Virgin Media O2 (VMO2) in the UK and VodafoneZiggo in the Netherland, and managing roughly 80 million connections across the bloc. For CEO Michael Fries, there’s a central maxim to be followed:

The telco sector, in my view, is ready-made to realize AI benefits, which over time should be transformational for us and our peers.

He backs up this claim by arguing that telco have ownership of the assets that AI needs most to succeed namely:

Large amounts of data that can't be replicated, massive cost structures, like call centers, field ops and networks that are built for automation, millions of daily touch points with consumers and the infrastructure like connectivity, and data centers that support the distribution layer for AI.

And, of course, telcos ought to be ‘eating their own dog food’ here, he states:

We are looking to benefit from the very same opportunities that our peers are attacking, namely driving margins through cost efficiencies, driving customer revenue growth through hyper-personalization, driving demand for our infrastructure, including power, space and cooling, and driving interest in our stock as investors rotate into sectors that are net beneficiaries of AI and not candidates for disruption.

A lot of lessons have been learned already, he attests, not least the importance of striking the right balance between build vs buy when it comes to AI solutions:

The results we're generating [include] things like reaching 65% of our VMO2 customer base with our personalization engine, generating 75% cost containment rates through our agentic AI pilots in the Netherlands, reducing fraud, optimizing CapEx and lowering truck rolls and technician costs.

But all of this, impressive though it sounds, is really just “table stakes for every telco”, he suggests:

Don't get me wrong. I'm proud of it. We're proud of it. On balance, we're realizing strong, marginal improvements to our economics, our customer interactions and our network quality, and as we've said publicly here, we expect to generate annual savings in the hundreds of millions.

But I know we are just scratching the surface here. Based on some work we did with McKinsey and Google, we analyzed some of our core operating expenses across the group to assess both the proportion of that cost which could be addressed by AI over time, and what some more ambitious savings targets might look like...[such as] savings of between 20% and 40%, even as high as 70% in things like customer care.  These are not fanciful numbers in my view. They look more realistic to me every day.

How then can these ambitions be realized? Fries sees a lot things working in its favor here:

On one hand, of course, we're implementing our own AI solutions with sophisticated and scaled partners to drive benefits. But equally important, on the other hand, we're seeing our largest suppliers, typically software and outsourcing partners, looking for early renewals in exchange for passing along to us the significant AI savings they themselves are realizing and must realize to stay relevant.

The cost benefits from AI are coming from two directions, he adds:

Self-induced organically-driven efficiencies, improvements, all the things that we know AI can do, that you're reading about it every day, we're on that....Every company in the group, both in the growth and the telecom portfolio, is implementing today solutions that are making them more efficient, faster, better, more profitable. That's happening organically as we speak.

On the AI front line at Liberty is CTO Enrique Rodriguez, someone who has an understanding of a major AI topic of debate - the impact of tokenomics, something that appears to have taken so many enterprises by surprises of late:

Like anybody else in the industry, we're watching the evolution of both token costs and the resulting benefits pretty closely. I can say categorically, we don't see a major issue with the increase in some cases of token costs because we've been, I think, pretty disciplined in making sure that we're applying those tokens against business cases that do bring those net benefits. This will be a continuing story, but I see a significant net benefit, even though like anybody else, we do see an increase in the usage of tokens and the related costs.

In terms of AI models, there’s a clear direction of travel, argues Fries:

I'm really thinking through and addressing the longer-term impact because the trend is only going one way, right? Models are getting smarter. More and more companies are arriving on the scene, taking advantage of that intelligence, driving solutions at scale for companies like ours and others. We don't see anything on the horizon that would change that trajectory. If you just extrapolate from where intelligence is moving and how costs are evolving in that space for beneficiaries like us, it's just going to get faster and cheaper.

As we apply that logic to more and more of our business, we just see nothing but upside. I mean we're only 20%, 25% in the cloud, 75% to 80% of our business is still on-prem. There are so many things our industry - and we're not different than any other telco - has yet to implement and take advantage of that.

It’s essential to think big, he urges:

I think it's almost irresponsible not to be that ambitious. I'm pounding the table every day with my team to tell me why we can't be that ambitious...you have to be thinking that broadly and, I think, aggressively over the next, let's say, two to three years, it's moving that fast. And that's how we're approaching it.

It's great to do the things we're doing. I'm proud of our industry, and I'm proud of my team, but it's just the start. There has to be a re-think of our operating models, how we're managing our businesses, talent and all the technology and software required to drive these kinds of step-change improvements. So I think it's real. I think it's sustainable, and we're anxiously working to deliver it.

We're taking a 360-degree view of the AI opportunity, which we believe is the best way to innovate and transform our business over the long term. I think it's going to be one hell of a ride.

Sounds like it might be.