The future of AI Network operations. Two Years On: Have We Been Replaced?

Looking Back: The 2023 Turning Point

Back in 2023, spurred by the release of ChatGPT, I wrote an article exploring the relevance of large language models (LLMs) to the field of network engineering. At the time, the use cases were limited, but the potential was clear.

Fast forward two years, and while the surface may seem calm, the undercurrents of change are undeniable. If you haven’t been following the developments closely, you might think not much has changed. But that’s exactly why I felt compelled to write this follow-up.

Introducing Angie: Agentic Network Engineering in Action

In my spare time (usually after the kids are in bed), I’ve been developing a custom Model Context Protocol server. The goal? To see what happens when you give a 2025-era frontier model access to the Cisco Catalyst Center API and ask it to troubleshoot a network fault.

Meet Angie

The result is Angie — an AI persona and toolkit designed to enable agentic workflows for network troubleshooting. On her very first run, Angie successfully diagnosed a faulty wireless client using data from just one API endpoint. And that’s just the beginning.

Angie troubleshooting a faulty wireless client

Even with only one endpoint implemented, Angie has already demonstrated her potential to reduce what might be the most critical metric in modern networking: Mean Time to Innocence — the time it takes to prove your network isn’t the problem.

The Bigger Picture: What Comes Next?

Expanding the Toolkit

There’s a whole ecosystem waiting to be integrated: SD-WAN Manager, Identity Services Engine (ISE), Secure Firewall Management Center — not to mention direct device access and third-party tools integrations. But even now, it’s easy to see the long-term value.

What’s most striking is that this was built in just a few evenings by one engineer. Now, imagine what Cisco is working on right now. What happens when they train custom models on every single TAC case, every piece of internal documentation?

It’s not a question of if they’ll offer “AI Network Engineers” as a service — it’s when. And at the current pace, I’d give it another two years. Until then, the field engineers still get the last laugh.

Until the robots arrive.

Ready for What’s Next?

Explore how Forfusion can help future-proof your network with intelligent, AI-ready managed services.

Talk to our team.

About the Author

Ezekiel Nye is a Senior Engineer within Forfusion’s Managed Services business unit, where he plays a pivotal role in supporting and operating complex IT solutions for customers across a wide range of industry verticals. With a strong foundation in enterprise networking and security, Ezekiel brings extensive expertise in infrastructure automation and programmability.

He specialises in leveraging coding and scripting tools to streamline processes, enhance operational efficiency, and drive innovation in network and security operations.

Original Linkedin Article