AI Advancing but Many Carriers Still Have Questions
Generative AI will see rapid growth in the U.S., including by telecom carriers, said Nelson Englert-Yang, industry analyst on strategic technologies at ABI Research. But many providers and other companies remain confused about how they will use AI, other experts said Tuesday during an RCR Wireless webinar.
Sign up for a free preview to unlock the rest of this article
Communications Daily is required reading for senior executives at top telecom corporations, law firms, lobbying organizations, associations and government agencies (including the FCC). Join them today!
Carriers are moving to live deployments of GAI and away from merely showing that the technology works, Englert-Yang said. ABI predicts $47 billion in annual investment in GAI by 2030, he said. “This will correspond with shifting priorities in how the technology is used.”
Early adopters prioritized areas that would lead to increased return on investment with lower risks, such as customer care, Englert-Yang said. Providers “are trying a broader set of use cases … and overall shifting beyond the traditional revenue drivers like network infrastructure,” he said. ABI is also seeing lots of “AI-driven network automation, including the use of AI agents.”
Providers remain concerned that one of the stumbling blocks to deploying AI is overcoming legacy data and infrastructure and removing operational silos, Englert-Yang said: “To address these challenges, providers have been shifting toward focused, small-scale deployments tailored to a narrow set of AI use cases.”
Some providers see AI as a “new shiny thing,” said Fatih Nar, Red Hat's chief architect-application platform solutions. Many have placed orders for “very expensive” accelerator hardware, he said: “We started receiving questions like, ‘We bought $10 million worth of [graphics processing units]. What can we do with them?’”
As a result, Red Hat sits down with customers, helping them "understand what they have bought and how we can accelerate their talent onboarding and skill-set growth, together with their real, down-to-earth use cases,” Nar said. The goal is to help providers understand how AI can “create new business channels” and go beyond lowering operational costs by doing their jobs faster and better, he added.
Business customers want more efficient networks as they modernize how their own operations run, said Steve Szabo, vice president-technology enablement at Verizon Business. “How can Verizon do better internally?” he asked: “That translates its way over to our customer.” Verizon wants to “marry the data” it has with the issues customers are seeing, he said. “The goal is always to leverage information and get smarter, faster … and more efficient” so that the experience of customers with the network keeps them happy.
A second goal of Verizon is using AI to personalize products and solutions the carrier offers “so that the customers can start benefiting from the differentiation that Verizon can bring to the table,” Szabo said.
Carriers want to keep their AI deployments as simple as possible, said Petri Hautakangas, CEO of AI-company Tupl. Companies want subject-matter experts, not data scientists, to be able to manage AI-based networks, he said. Using AI for customer care “has proven to create significant value for our operators,” with up to 90% of operations automated in some cases, he said. AI helps eliminate “manual, repetitive tasks.”
Carriers are also working with Tupl on network operations, Hautakangas said. “Our system is helping … transport-core engineers to automate their work tasks” and move “to the next level of expertise.” One of the growing uses of AI is managing networks to lower energy costs, he said. An AI system needs to be trustworthy and transparent, or it won’t be trusted by network engineers to automate operations, he said. “There always needs to be a clear business case attached to the use cases.” T-Mobile is among the company’s clients.