The ABCs of AI

In this episode of The Art of Network Engineering, @permitipandyandy is joined by returning guest @johncapobianco2527 (Head of AI & DevRel at @Itential and Mike Bushong to cut through the AI noise and give network engineers a practical on-ramp.

We talk about why network automation adoption is still stuck (the famous “70% aren’t automating” stat), why the “automate or die” narrative backfires, and what recent advances in AI mean to network engineers.

John breaks down the concepts network engineers keep hearing but rarely get explained clearly:

LLMs (and why “using ChatGPT like Google” misses the point) RAG (how to ground answers in your real data) Agentic AI (what “agents” actually do) MCP (Model Context Protocol)

Then we get tactical: the Hello World for network engineers isn’t “write an app.” It’s connecting an agent to your source of truth (like NetBox) so you can ask natural-language questions like: “What circuits do we have in Atlanta?” or “What IPs are available in this subnet?”

We also cover a sane safety approach: start read-only, use guardrails, and only expand into config changes once you trust the workflow.

If you’ve felt the eye-roll every time someone says “AI,” this one is for you.

John Capobianco: https://www.linkedin.com/in/john-capobianco-644a1515/

Mike Bushong: https://www.linkedin.com/in/michaelbushong/

Links & Resources: https://linktr.ee/artofneteng

Subscribe and stay tuned for the next episode in this AI series, where we’ll build the “Hello World” workflow step-by-step.

This episode has been sponsored by Meter. Go to https://meter.com/aone to book a demo now!