3
As Boost Mobile has learned, ervice providers have to understand, “Where does AI bring value to the business and where does it not?”
The telecom industry has spent the past decade talking about virtualization, cloud-native networks, DevOps methodologies and now, agentic AI. The sequence is telling. Before an operator can delegate meaningful work to AI agents, it has to know what those agents are reasoning over. That puts data, not algorithms, at the beginning of the story.
For Boost Mobile, formerly Dish Network, that starting point was unusually intentional. Speaking at DTW Ignite in Copenhagen, Boost executives Jeff McSchooler and Dawood Shahdad framed the company’s AI strategy around a foundational decision made when the network was first designed: own the data.
“As we were designing the network originally…one of the precipice things that I put in place is that we had to have ownership of all data,” McSchooler said. “We couldn’t let that data be held by another vendor or by anybody else. We got to have it.”
That decision now looks prescient. Boost’s cloud-native 5G network was built with data acquisition from the RAN, core and other network domains in mind. The result is that the data is already in one place and can be correlated across domains. “Now for us to do anything…the data is already lined up,” McSchooler said. “It’s in one place. We can correlate it or do anything we want with it.”
But the more interesting point is that Boost’s executives are not arguing that more data is always better. Shahdad made the more pragmatic case. The company built the “piping” to collect data broadly, but quickly realized it did not need to move everything into a centralized environment. The emerging AI problem is not just data readiness but data selection.
“We’re now deciding…at the function level and domain level itself, what is the data we want to ship out to the centralized location,” Shahdad said. “At the end of the day, we want to make sure it’s efficient.”
That is where agentic AI becomes useful rather than fashionable. The objective is to understand which data matters for which operational decision, then use agents to act on that data in a way that improves the network or the business. Shahdad said Boost is working with partners including Nokia, Google, Oracle and Mavenir to determine “just the right amount of data” needed to make meaningful decisions.
This is an important distinction. AI-native networking can easily become a slogan. McSchooler pushed instead toward the language of value. “Where does AI bring value to the business and where does it not?” he said. “There’s some amazing things you can do, but they don’t bring any money to the company. They don’t bring any value. They don’t bring another customer.”
Shahdad’s north star is more operational: “Three years from now…the network that runs itself.” His example was software management. Today, network upgrades still require significant manual effort. In an AI-native world, he said, network software should update more like a smartphone with the latest features pushed automatically, safely and without human choreography during every step.
The payoff goes beyond efficiency. Once service providers can reduce the operational burden of maintaining the network, they can redirect scarce technical talent toward services, monetization and customer experience.
That, however, requires organizational change as much as technical change. McSchooler said the first issue is trust inside the company. People have to believe AI is “the right thing for the future of them and their jobs.” Shahdad was blunter. Boost built its operating model around in-office collaboration, software development-style ownership and reduced silos. The same teams that design, test and deploy now have to learn how to apply AI to that operating model.
“There’s going to be meaningful jobs,” Shahdad said. “We as humans need to provide oversight of AI during this initial journey of AI taking over the networks.”
The agentic network, then, is not a technology overlay. It is a data architecture, an operating model and a cultural transition. Boost’s position is that its structural choices — cloud-native from the start, data ownership by design and cross-domain operational thinking — give it a cleaner path than most. But the broader lesson applies across telecom. Agentic AI begins with knowing what data matters, who controls it and what business outcome it is supposed to serve.
