- EchoStar's Boost network already uses AI and automation for root-cause analysis, remediation and correlating data across different parts of the business
- EchoStar EVP Jeff McSchooler said its AI challenge is getting decades of data cleaned up, inspected and structured so AI systems can use it
- The Boost network is already heavily automated and increasingly incorporating AI into network operations
FUTURENET NORTH AMERICA, NEW YORK CITY – The telecom industry is shifting toward autonomous networks – or trying to, given the myriad of challenges in the way, including legacy networks. And like most operators, EchoStar isn't pretending it has already crossed the chasm.
Speaking on a panel focused on AI-driven network operations and closed-loop autonomy at FutureNet North America, Jeff McSchooler, EVP of engineering and operations at EchoStar, said the company's Boost Mobile network is already heavily automated and increasingly incorporating AI into network operations but the scale of that autonomy is limited.
"Scale is tough because it depends on what you want to look at, what that word really means," McSchooler said. Much of the Boost network operations center runs in an automated fashion, he said.
The company started with AI and machine learning and is now moving toward agentic AI. It is applying the technology to root-cause analysis, automated correction and remediation, he said, adding that EchoStar is also combining data from different divisions to determine things like whether an apparent outage is actually a network outage.
"Sometimes we think it is, but it isn't," McSchooler said.
The goal is to pull those different sources of information together so operations teams have a better understanding of what is really happening in the network, he noted.
As Fierce Network has heard, AI hasn’t quite moved from diagnosing a problem to making changes to the network. However, McSchooler said Boost is likely the furthest along among EchoStar's businesses when it comes to network-operations automation. In some circumstances, systems are allowed to autonomously remediate problems. In others, they aren't, he said.
Over the next 12 months, McSchooler expects EchoStar to expand what its systems are allowed to do autonomously. How quickly that happens will depend on whether the company has adequate safeguards to validate the actions its AI agents take.
"If you ask the same prompt to three different AIs, you'll get three different answers," he said. "They're somewhat the same, but they're not the same, and so you've got to really be careful."
The hard part is always the data
One big obstacle is data — Fierce has heard this same story again and again and there doesn’t seem to be a solution as of yet.
Different parts of EchoStar are starting from very different places, McSchooler said. The Boost team had the advantage of building systems from the ground up around data, while other parts of the company have decades of legacy information to sort through.
"There’s a huge selection of data that's done, and then a huge inspection and cleanup of that data," McSchooler said. The goal is to retain the information that can help the company operate its networks going forward, he said.
When that data can be combined effectively, an AI system can pull together information from cells, network infrastructure and other systems and help determine what caused an outage, he said.
His advice for other communications service providers chasing autonomy was consequently straightforward: get the data right and don't attempt to automate everything at once.
"Make sure your data is good. Bite off what you can chew," he said. "Go after that, get really good at it and then bite off the next jump."
Read more about EchoStar and Boost Mobile on Fierce Network:
EchoStar loses 182K wireless customers in Q3
EchoStar gains 3% on report SpaceX eyes Boost Mobile
EchoStar’s Dish DBS emerges from bankruptcy – wireless still in limbo