Photo By: David Arrowsmith
Artificial intelligence is moving deeper into the telecom network, taking on an expanding role in monitoring, troubleshooting and optimization. But according to Kyle Brueckner, VP of AI Service Delivery at NewRocket, the rise of AI will not make network engineers less important. Instead, it will change what the job demands.
NewRocket is the trusted AI partner helping enterprises build AI workflows customers trust. As a launch Select Partner in Anthropic’s Claude Partner Network, NewRocket helps organizations unlock the value of AI through strategy, governance, adoption, and delivery. Combined with ServiceNow’s workflow platform and more than two decades of enterprise transformation expertise, NewRocket designs, builds, and scales AI solutions that create outcomes that last.
As AI takes over more routine operational tasks, engineers will increasingly focus on architecture, resilience, service performance and preventing problems before they recur.
“I see network engineers taking greater responsibility for designing, governing, and improving how networks operate,” Brueckner says. “As AI handles more routine monitoring, correlates alarms, and recommends corrective actions, engineers can spend more time on resilience, architecture, service performance, and preventing recurring problems.”
That shift will require a broader technical skill set. Engineers will need to combine traditional networking expertise with automation, software, data and AI evaluation. Crucially, they will also need to judge when an AI-generated recommendation should, or should not, be acted upon.
A technically valid network change can still have unacceptable consequences for customers, security or operations, Brueckner points out.
That human oversight is particularly important as telecom operators attempt to move AI projects beyond experimentation and into production. The technology may work in a controlled pilot, but deploying it across complex, multi-vendor networks introduces a different set of challenges.
Brueckner identifies fragmented data, integration complexity, unclear ownership and a lack of trust under real-world operating conditions as some of the biggest obstacles.
“A production service needs reliable data access, defined permissions, performance evaluation, monitoring, recovery procedures, and an owner who can support it over time,” he says.
For operators, that means AI cannot be treated as an isolated technology project. The network signal needs to be connected to the service it affects, the customer impact it creates and the operational workflow required to respond.
At the same time, telecom operators are confronting another increasingly important issue: sovereignty. As networks become more software-driven and AI workloads expand, questions around where data is processed, who controls infrastructure and models, and how dependent an operator becomes on a supplier are becoming architectural decisions rather than purely legal or compliance considerations.
Brueckner argues that sovereignty needs to extend across the entire AI chain, from network telemetry and customer information to prompts, outputs, logs, compute, models and cloud infrastructure.
“Operators need to understand model provenance, licensing, update control, and their ability to evaluate, adapt, or replace the model,” he says. “For cloud infrastructure, the questions extend to encryption keys, management systems, support personnel, subcontractors, and recovery arrangements.”
The broader transformation of telecom infrastructure is also gathering pace. Cloud-native platforms, 5G standalone, network APIs and increasingly sophisticated automation are creating new opportunities to make networks more programmable and responsive.
But simply moving network functions onto cloud infrastructure will not deliver the full benefit, Brueckner argues. Operators also need to rethink the processes and organizational structures surrounding those technologies.
“The architecture and the way people work need to evolve together,” he says.
That is particularly true as software becomes a strategic asset alongside physical infrastructure. Spectrum, fiber, radio equipment and power remain fundamental, but software increasingly determines how efficiently those assets can be used and how quickly they can be turned into customer-facing services.
For operators, the challenge is therefore not simply becoming more automated or more cloud-native. It is building an operating model capable of taking advantage of those technologies throughout their lifecycle, from deployment and upgrades to security, scaling, failure recovery and ongoing optimization.
The ultimate measure, however, will be the customer experience.
Brueckner expects the transformation to deliver more dependable services, faster activation, fewer recurring problems and quicker resolution when issues occur. For enterprise customers, programmable networks could also enable more tailored connectivity, private networks and services supported by edge computing.
The next phase of telecom transformation, he argues, should begin with customer and operational outcomes rather than technology for its own sake.
Operators need trustworthy data, clear service inventories, controlled automation and defined rules around what AI can recommend or execute. They also need to invest in people, ensuring engineers receive the training and practical experience necessary to work confidently alongside increasingly autonomous systems.
Ultimately, the promise of AI in telecom is not simply to make networks more intelligent. It is to make them more resilient, responsive and useful.
The operators that succeed will be those that connect technology, people and operational accountability, turning AI and software-defined infrastructure into tangible improvements in the services customers experience every day. The competitive advantage will come not from simply deploying AI, but from integrating it into the network, workforce and operating model in a way that delivers measurable results. For telecom operators, that could make the difference between AI remaining an ongoing experiment and becoming a core driver of the next generation of network services.