November 22, 2026

The Evolving Role of the Network Engineer: How AI Agents are Redefining the Mental Model of the CLI

In the traditional trajectory of a network engineer, the journey almost invariably begins with the command-line interface (CLI). For decades, the Cisco Certified Network Associate (CCNA) certification has served as the industry’s gatekeeper, requiring aspirants to spend hundreds of hours mastering the syntax of Cisco IOS, memorizing complex command strings, and perfecting the art of manual configuration. However, as artificial intelligence moves from the realm of theoretical experimentation into the core of enterprise operations, a fundamental question has emerged: In an era where AI agents can generate configurations and troubleshoot protocols in seconds, what is the value of the manual skill sets that have defined the profession for a generation?

This shift in the networking landscape was the central theme of a recent discussion on the Packet Pushers podcast, featuring Sergio Freitas, an Engineering Director at Cisco. In a conversation with hosts Ethan Banks and Holly, Freitas explored the profound impact that AI agents are having on network operations and, perhaps more importantly, how these technologies are forcing a total reimagining of what it means to be a network engineer in the modern age.

The Shift from Configuration to Intent

For those currently pursuing their CCNA, the curriculum remains heavily focused on the "how" of networking. Students learn how to configure VLANs, how to set up OSPF areas, and how to secure ports through specific CLI entries. While these skills remain foundational, Freitas suggests that the arrival of AI agents is shifting the engineer’s primary responsibility away from manual execution and toward high-level intent.

In the traditional mental model, a network engineer acts as a translator. They take a business requirement—such as "this department needs secure access to this server"—and translate it into a series of technical commands across multiple devices. This process is often slow, prone to human error, and requires a deep, granular knowledge of specific vendor syntax.

With the integration of AI agents, this translation layer is becoming increasingly automated. AI agents are not merely chatbots that provide information; they are functional entities capable of understanding context and executing tasks. Freitas points out that as these agents take over the "heavy lifting" of configuration, the engineer’s role evolves into that of an orchestrator. The mental model shifts from "how do I type this command?" to "what is the desired state of my network, and how do I validate that the AI has achieved it?"

Redefining the CCNA Foundation

The rise of AI has sparked a debate within the networking community regarding the relevance of entry-level certifications like the CCNA. If an AI agent can handle the CLI, does a new engineer still need to learn it?

The consensus among industry veterans, including Freitas, is that foundational knowledge remains indispensable, though its application is changing. Understanding the CLI is no longer just about being able to configure a switch; it is about having the "first principles" knowledge required to supervise and troubleshoot the AI.

When an AI agent makes a decision or suggests a configuration, a human engineer must have the underlying expertise to recognize whether that suggestion is sound. Without the rigorous training provided by certifications like the CCNA, an engineer risks becoming overly reliant on a "black box" system. Freitas emphasizes that while AI agents change the speed and scale of operations, they do not eliminate the need for the logical framework that manual CLI training provides. The difference is that instead of spending 90% of their time typing commands, engineers may now spend 90% of their time designing architectures and verifying the AI’s output.

The Advent of AI Agents in the Field

The discussion highlights a critical distinction between traditional automation and the new wave of AI agents. Traditional automation, such as Python scripting or Ansible playbooks, is deterministic. It follows a set of pre-defined rules to achieve a specific outcome. If the environment changes in a way the script didn’t anticipate, the automation often fails.

AI agents, as Freitas describes them, are more adaptive. Drawing from his extensive experience at Cisco, Freitas explains how these agents can analyze vast amounts of telemetry data in real-time to identify patterns that a human might miss. They can correlate events across the entire network fabric, providing insights into "Day 2" operations—the ongoing maintenance and troubleshooting phase—that were previously impossible to achieve manually.

This transition into AI-driven operations is fundamentally changing the "mental model" of the profession. In the past, troubleshooting often involved a "hunt and peck" method: logging into individual devices, checking interface statistics, and looking for anomalies. An AI-augmented engineer, however, interacts with the network as a singular, holistic system. They use AI agents to query the state of the entire infrastructure, allowing for a proactive rather than reactive approach to network health.

The Changing Mental Model of the Engineer

One of the most striking points made by Freitas is the psychological shift required by today’s engineers. For many, the ability to navigate a CLI with speed and precision has been a point of professional pride—a "badge of honor" that separates the experts from the novices. Relinquishing that control to an AI agent can be a difficult transition.

However, Freitas argues that this shift is necessary for the industry to move forward. As networks become more complex—spanning on-premises data centers, multiple cloud providers, and thousands of edge devices—the manual approach simply does not scale. The "new" network engineer must be comfortable with abstraction. They must move up the stack, focusing on policy, security posture, and user experience, rather than the minutiae of individual port configurations.

This change in the mental model also involves a shift toward data literacy. Engineers are increasingly required to understand how data flows through AI models and how to interpret the probabilistic outputs of an AI agent. It is no longer enough to know that a command worked; one must understand the data that led the AI to recommend that command in the first place.

Cisco’s Role and the Future of Network Operations

As a leader in the networking space, Cisco is at the forefront of integrating these AI capabilities into its product ecosystem. Sergio Freitas’s insights reflect a broader corporate strategy to move toward "Autonomous Networks." This vision involves a self-healing, self-configuring infrastructure where AI agents act as a constant companion to the human operator.

For the engineering community, this means that the tools of the trade are evolving. The podcast discussion underscores that the relationship between the engineer and the network is becoming more conversational and intent-based. This doesn’t mean the CLI is dead—it remains the ultimate "source of truth" for what is happening on a device—but it is no longer the primary interface for daily operations.

Community and Continued Learning

The evolution of the network engineer is not a solitary journey. Organizations like Packet Pushers have built entire communities around the idea of navigating these technological shifts together. Through platforms like their Slack community, professionals can discuss the practicalities of implementing AI, share experiences with new tools, and help one another adapt to the changing requirements of the job market.

As Freitas concludes, the rise of AI agents is not a threat to the network engineering profession, but rather an invitation to innovate. By offloading the repetitive, syntax-heavy tasks to AI, engineers are freed to focus on the creative and strategic aspects of networking. The "mental model" of the future is one where the human provides the vision and the ethics, while the AI provides the speed and the scale.

For those just starting out with their CCNA, the message is clear: learn the fundamentals, master the CLI, and understand the protocols—but do so with the awareness that these skills are the foundation for a much larger, more complex role as an AI-enabled architect of the digital world. The CLI is the beginning of the journey, not the destination. The future of networking lies in the successful partnership between human intuition and machine intelligence.