As AI becomes an increasingly dominant theme in telecoms, a recent TM Forum webinar explores how the technology could be used to create new services and ultimately, it’s hoped, generate revenue.
(Source: Skorzewiak/Alamy Stock Photo)
Mobile World Congress is only a few weeks away and it’s already clear that artificial intelligence (AI) in telecom, from predictive through to generative and agentic, will be a prominent theme, as it has been since ChatGPT first made its presence felt. A quick glance at the conference program alone makes it clear how much AI will dominate the agenda, with sessions on AI and diversity, on-device AI, AI and governance and more besides.
More nuanced operator approaches to AI in telecom are beginning to crystallize, meanwhile. For instance, Laurent Leboucher, group chief technology officer and executive vice president of innovation networks at Orange, distinguishes between “AI for network”and “network for AI” to explain how operators handle a dual focus on AI, whereby the technology is used to optimize networks on one hand, and support revenue-generating services on the other.
A key question that many will be asking at MWC is: how will operators generate value from their deployments of AI, and what should the next steps be now that the industry is, perhaps, starting to move a little beyond the hype?
Driving value
Recent developments in AI mean that operators have the opportunity to use the technology for new product and service creation, potentially creating new revenue streams as well as optimizing their own networks and operations. But where do the opportunities lie, and where do they fit into the value chain?
This was the topic of a TM Forum webinar on Tuesday, titled “Leveraging AI for service and value creation.”
Mark Newman, chief analyst at TM Forum, remarked that much of the industry focus until now has been on driving operational efficiencies for greater productivity.
“That can be one part of value creation. The other part of value creation can be creating new products and services, or enhanced products and services,” Newman said.
At the same time, he added, the idea of developing new products is “newer to operators, and it’s lagging behind. However … there’s a recognition in the industry that over a period of time, the focus on value creation will become as important as the focus on productivity and efficiency gain.”
To be sure, operators “have had a tough time building new value-added products and services. They’ve been endeavoring to expand beyond voice, messaging and connectivity for the last ten, 15, 20 years, without too much success,” Newman said.
For example, he made a comparison with operator attempts in the past to become public cloud providers. Ultimately these strategies largely failed, and operators became consumers of cloud computing services instead.
“Is the industry in a different position for leveraging AI for value creation than it was for cloud computing? … There is generally a view that operators might have always been behind the curve when it came to cloud computing, but when it comes to AI, maybe some of them are not so far behind. So they’re at an earlier stage in the exploitation of AI, to play an important role here,” Newman said.
Getting an edge
One topic that also emerged was the potential for edge computing with AI. As commented by Newman, “many enterprises will want to put their AI workloads both close to their own business and close to the network to guarantee performance, security, latency. So if that is the case, if that compute needs to be right next to network, does that therefore give telecom operators the opportunity to play in the edge computing space? Is it time to dust off those edge computing investment plans?”
Volker Tegtmeyer, principal product marketing manager at Red Hat, said edge will be key to avoid having to move all of the data to a central location.
“I think from a deployment point of view, the flexibility is key. It’s either edge or it’s a private cloud, or could even be a public cloud, if you want to start something … at a small scale,” he said.
Tegtmeyer added: “What we expect to see is that service providers will have hundreds of AI models. They will leverage agentic AI, predictive AI, GenAI, and they might actually combine it, depending on what they need, to at the end, build a service. Not to make things more complicated, but yes, edge is important, and it will be a part of a much bigger puzzle.”
Getting down to business
According to Newman, “we can be pretty sure that there will be an explosion in new AI-infused products and services.”
He cited examples such as enhanced voice services, where operators try to modernize legacy voice and integrate them with services such as Microsoft Teams. Or there are concierge services, such as GenAI-based tools to help manage everyday life, as well as the concept of “artificial intelligence of things,” such as the combination of AI with video capabilities.
Ryan Walton-King, global industry market leader, communications, media, and consumer services at Pegasystems, also pointed out that predictive and adaptive AI can be used to gain insights about customers and then determine the best way to engage them, “whether that’s via digital, whether that’s via a call center rep, and making sure that you’re doing that with empathy and with context.
Meanwhile, Richard Doughty, business development director at Cerillion Technologies, provided a couple of specific case studies of current AI implementations at telcos involving the deployment of AI-enabled catalogs and workflow. Here, he cited work with Paratus in South Africa as well as Ucom in Armenia.
Cerillion is “putting AI into the products to make them increasingly frictionless to use, and there’s a big focus on that ease of use, breaking down the interface … and then using that ease of interface with things like catalog to then help create products far faster, so reduce that time to market from weeks, maybe where it was 20 years ago, to days, really down to minutes now,” Doughty said.
Walton-King also made reference to work with a US carrier on what he called outage deflection using AI. “We’ve been able to deflect over 200,000 calls in the call center and identify outages 15 minutes earlier than when they could before,” he said. “A struggle for a lot of carriers is, how do we get ahead of these outages so that we don’t have people calling in? We can deflect calls, but even better, we can fix it before it becomes an outage.”
Source : Anne Morris (Contributing Editor, Light Reading)
