The Future of Network Operations: Navigating the Shift Toward Agentic AI and Autonomous Systems

The landscape of network operations is currently undergoing one of the most significant transformations since the transition from hardware-centric environments to software-defined networking. As enterprises grapple with increasing complexity, the sheer scale of modern infrastructure is beginning to outpace the capacity of traditional manual management. In a recent industry discussion, Rekha Shenoy and Irfahn Khimji of BackBox sat down with Scott Robohn to explore the trajectory of network operations (NetOps) as the industry moves toward the year 2030. The conversation centered on the emergence of "agentic NetOps"—a paradigm shift where artificial intelligence does not merely suggest actions but possesses the agency to execute them—and the critical necessity of maintaining operational control through robust backup and rollback strategies.

The evolution of NetOps has historically moved in waves. For decades, the primary interface for network engineers was the Command Line Interface (CLI), requiring deep specialized knowledge and manual entry for every configuration change. While the introduction of basic automation and scripting helped alleviate some of the repetitive burdens, the industry is now standing on the precipice of a much more profound change. The move toward agentic AI represents a departure from simple "if-then" automation toward systems that can understand intent, analyze real-time conditions, and take proactive steps to maintain network health and security.

The Rise of Agentic NetOps

At the heart of the discussion was the concept of "agentic" systems. Unlike traditional AI, which might provide a dashboard of alerts or a set of recommendations that a human must then approve and implement, agentic AI is designed to operate with a level of autonomy. In a NetOps context, this means an AI agent could identify a performance bottleneck, determine the necessary configuration changes across multiple vendors’ equipment, and execute those changes without direct human intervention at every step.

However, the shift to agentic NetOps is not merely a technical upgrade; it is a fundamental change in how network teams interact with their infrastructure. Shenoy and Khimji highlighted that the transition to 2030 will require teams to move from being "doers" to being "orchestrators." In this new model, the role of the network engineer involves setting the guardrails, defining the desired state of the network, and supervising the AI agents that carry out the day-to-day operations. This shift promises to solve the "talent gap" by allowing senior engineers to focus on high-level architecture and strategy rather than losing hours to routine maintenance or firefighting.

The promise of agentic AI is particularly relevant as networks become more distributed. With the rise of edge computing, multi-cloud environments, and the Internet of Things (IoT), the number of endpoints and nodes that require management is exploding. Humans simply cannot scale at the same rate as the digital infrastructure they are tasked with protecting. By 2030, the consensus suggests that autonomous or semi-autonomous agents will be a requirement for operational viability, rather than a luxury for the most advanced enterprises.

Balancing Automation with Operational Control

One of the most pressing concerns for NetOps teams is the loss of control. When an automated system—especially one driven by complex AI models—makes a change, there is an inherent risk that the change could lead to unforeseen consequences or network downtime. The discussion with Robohn emphasized that the fear of the "automated mistake" is one of the primary barriers to the adoption of advanced NetOps tools.