- New priorities for network engineers include systems thinking, data observability, agent governance and software engineering
- Verizon automated 70 million configuration changes last year, freeing engineers to focus on complex problems instead of routine tasks
- Find out more about AI and the automated network at our Fierce Network Virtual Summit June 10. And if you can't make it, the summit is available for playback on demand
The job of the network engineer is changing fast. AI is absorbing the routine work — configuration changes, ticket triage, pattern matching — and the role is shifting to something fundamentally different.
In preparation for our June 10 Fierce Network AI and the Automated Network Summit, we talked with executives from Verizon, Telstra and a leading industry analyst, who looked into the future of network engineering work. Join us for the summit, and read our accompanying report, The Telco Guide to AI-Driven Networks.
Human network engineers are still important in the world of automated networks. "Fully agentic does not mean fully unsupervised," said Sid Nag, president and chief research officer at Tekonyx, a research firm.
And Verizon, for example, is pushing from Level 3 automation toward Level 4, with human oversight throughout, said Anil Guntupalli, the operator's SVP of technology and product engineering.
From device configuration to defining intent
Increased automation requires a higher-level view for network engineers, said Regan Ireland, global head of pre-sales solution at Telstra. "It's not about defining a specific configuration on a device," he said. "It's about defining the intent and the guardrails and the constraints, and the parameters within which those systems will operate and have autonomy to make those decisions." That work — setting the rules, not executing them — is increasingly the mission for network engineering expertise.
From reactive to supervisory
The NOC as it exists today is essentially a ticket-routing operation. That model is ending. "The NOC becomes less of a ticket and troubleshooting processing center and more of an AI operations command and control center, supervising fleets of agents," Nag said. Traditional first-level triage roles shrink. New roles emerge around agent governance, policy design, exception handling and autonomous operations engineering, he said.
Solving more complex problems
Engineers who once spent hours on configuration work now focus on harder problems, Guntupalli said. Verizon automated 70 million network configuration changes last year. "It's not a productivity story, it's a capacity story," he said. Some CSPs operating at Level 4 autonomy estimate they can resolve 95% of trouble tickets without human intervention, according to a recent TM Forum report.
Systems thinking replaces domain expertise
The engineer of the future is someone "designing and monitoring an autonomous system" and "working at a level of abstraction and orchestration, where we've traditionally sat more at the device level," Ireland said. That requires understanding how the whole system behaves, not just one domain or device class. It's systems thinking.
Agent governance becomes a dedicated discipline
As agents take on operational standing inside the network, someone has to govern them. "Agents need role-based permissions, least-privilege access, traceable reasoning and approval gates," Nag said. "Accountability belongs to the organization that designed, approved and governed that agent." Managing that accountability is a new engineering function with no real precedent in the previous generation of network operations.
Data observability becomes a core skill
"You need that clean telemetry, you need the observability, you need that governed source of truth, and you have to have confidence in it to automate those tasks — because otherwise you won't get that predictable behavior under pressure. You're just moving risk around," Ireland said. Research backs this up: 38% of organizations that experienced AI setbacks cited poor data quality as a direct cause of failure, according to Gartner. Most operators remain stuck between Level 0 and Level 1 autonomy partly for this reason.
Software engineering joins the toolkit
"As we become more software-led with our networking as well, it is a different skill set," Ireland said. "We need to go on both a cultural and educational uplift of our workforce."
The AI and the Automated Network Summit is June 10, 11 a.m.–3:30 p.m. EDT. Speakers include executives from Verizon, Telstra, T-Mobile, AT&T, Lumen, Bell Canada, TELUS, MetTel, Zayo, Rakuten and others. Register here, and read the companion report, "The Telco Guide to AI-Driven Networks"