Appledore Research talks APIs, BSS/OSS, network automation and monetization
Operators are busy at the moment—they’re navigating a tough macro environment while working to muster the capital needed to modernize network hardware and software, including the slow transition to cloud-native networking. Simultaneously, the commercialization of artificial intelligence (AI) is shining a light on the latent value of massive data stores largely locked in vertical silos. To understand the dynamics shaping investments and strategies, RCR Wireless News spoke to Appledore Research Founder, Partner and Principal Analyst Patrick Kelly.
“There’s a lot going on right now but not necessarily a lot of revenue being generated,” Kelly said. “Let’s be real,” he continued. “There’s been hundreds of billions of dollars invested in the 5G radio access network and in the core network, and yet there hasn’t been material improvement in top line revenue.”
Whether the focus is in opening up the radio access network (RAN), deploying a cloud-native 5G core, standing up AI-enabled customer service processes, figuring out go-to-market for network application programming interfaces (APIs), or something else, the general direction of travel is toward increased levels of network automation, which is Appledore’s primary focus. Citing his firm’s research, Kelly reckons the network automation software market will grow at around a 14% CAGR to approximately $15 billion in 2028.
“When we talk to most of the operators,” he said, “the area that they’re really focused on…is modernization of the software. There’s [also] obviously a lot going on around data and AI advancements, and I’ll include generative AI in there, quite a bit around operational efficiency.”
An area of opportunity for generative AI (gen AI) that was foregrounded during the TM Forum’s Digital Transformation World—Ignite show hosted in Copenhagen, Denmark, in June, is abstracting legacy BSS/OSS systems and adding in a gen AI interface to help accelerate time-to-value of existing customer and network data. Kelly acknowledged this approach “is a bit of a wrapper…[for] legacy-based systems,” but the reality is, “They’re not going to rip-and-replace the legacy systems they have in place…I think gen AI can be used to establish intent of the service with the right data, and then still leverage the legacy systems that have been in place to execute the instructions. I think that’s more of a practical method that we see sort of evolving as you see more of these implementations reach commercialization.”
In terms of specific use cases gaining traction, Kelly called out “bill shock,” where a customer receives a monthly bill not aligned with their expectations, and then calls into an operator’s customer service center. He said about 30% of customer request are around billing questions. “I think what gen AI can do is you can deflect some of these basic requests by first understanding that data that’s within the operator’s environment, but also looking at behavioral patterns.” Essentially, gen AI could help cut down on call center resources, while simultaneously delivering an improved customer experience which potentially maps to improved account retention and even growth.
As to the 5G monetization piece, Kelly suggested operators look at adjacent markets they may be able to participate in; enterprise WAN, for instance. “We’ve sized that opportunity at about $100 billion on a global basis. But the way to actually participate and grow into these new markets is, I think…you need a new architecture based on new assumptions. At the same time, you have to risk mitigate. It can’t be a big bang, so you have to do sort of incremental, phased…investment in some of these systems. We believe that network technologies are becoming more software-based. They’ll be highly-programmable and flexible. The whole principle of intent and abstraction is a whole other area. If you can abstract away some of the complexity that’s still used—even some of these legacy systems—to configure or heal or scale this infrastructure, that’s going to be a big win for a lot of the operators that are looking to make this transition.”
Appledore Research has a multi-part report series focused on principles of autonomous networks that goes into great detail about the organizational and technological considerations that should inform network automation investments and strategies. Kelly said, “Effectively what we looked at is the principles, or best practices, in moving or migrating your systems to a more modern architecture.”
He continued: “One of the areas in these best practices that we look at is the orchestration process…how you orchestrate across many different technology domains or even organizational domains.” Circling back to the idea of intent, Kelly described it, in the context of a network-enabled service, as “the ability to actually describe a service more on business process, but allow the systems to code and instruct the infrastructure on how to deliver that service. The other thing that we looked at is there’s this whole thing in our industry around closed-loop automation. We kind of believe there will be multiple nested loops, so multiple control loops, in each domain. And then those will be nested in other control loops to bring it to that higher service layer that sort of satisfies that intent-based networking.”
