Uploaded July 2026 | Updated November 2026, 3 hours ago
A decade of intent-based frameworks, sources of truth, and infrastructure as code has produced real progress, yet most teams still can't automate beyond a handful of brittle pipelines. The problem isn't effort; it's that we've been missing an intelligence layer.
In this closing keynote, Itential Co-Founder & CTO Chris Wade makes a practical, hype-free case for agentic automation in network operations. We will explore the "smiling curve" of where LLMs actually perform (and why the messy middle of OS-specific, multi-vendor questions fails), the anatomy of a ReAct agent built from MCP, skills, and knowledge, and the critical distinction between build agents and operate agents: one path lets your team adopt AI while every outcome stays fully deterministic, and no agent ever touches production. Learn how to balance reasoning with determinism, engage your AI governance team, and move from human-in-the-loop toward human-out-of-the-loop on your own terms. Because AI isn't a bolt-on to network automation; it's the layer that finally delivers on its promise.
#AutoCon5 #networkautomation #agenticAI #netdevops #MCP
A decade of intent-based frameworks, sources of truth, and infrastructure as code has produced real progress, yet most teams still can't automate beyond a handful of brittle pipelines. The problem isn't effort; it's that we've been missing an intelligence layer.
In this closing keynote, Itential Co-Founder & CTO Chris Wade makes a practical, hype-free case for agentic automation in network operations. We will explore the "smiling curve" of where LLMs actually perform (and why the messy middle of OS-specific, multi-vendor questions fails), the anatomy of a ReAct agent built from MCP, skills, and knowledge, and the critical distinction between build agents and operate agents: one path lets your team adopt AI while every outcome stays fully deterministic, and no agent ever touches production. Learn how to balance reasoning with determinism, engage your AI governance team, and move from human-in-the-loop toward human-out-of-the-loop on your own terms. Because AI isn't a bolt-on to network automation; it's the layer that finally delivers on its promise.
#AutoCon5 #networkautomation #agenticAI #netdevops #MCP










