Itential

Most infrastructure teams are closer to getting agentic operations on live environments than they think.

John Capobianco explains the easiest way to get started: open a session in Claude Code, connect it to the right context, and start asking questions.

He walks through what comes next in the full demo with Netbox Labs, building and running two FlowAgents in real time. One agent reads device health state, diffs it against NetBox context, populates missing interfaces and IPs, then sends a report on completion. The second agent does the reverse: pulling intent from NetBox to push config changes downstream to the device.

No Python required. No hand-coded REST calls. RBAC and a full audit trail on every agent action.

See how he built it 👉 https://www.itential.com/resource/demo/netbox-flowai-agents/?utm_source=Facebook&utm_campaign=itential&utm_medium=social

Every automation pipeline starts as a happy path. Then a merge fails. An environment goes offline. You write a workaround. Then another. Then ten more.

A year later, that clean workflow is a technical debt machine, and most of what you're maintaining has nothing to do with the original outcome.

CTO Chris Wade breaks down how that process changes with the right agentic approach in this episode of Packet Pushers' Cloud Gambit. His take: a production-ready operating model pairs deterministic workflows for happy paths with AI reasoning that covers the exceptions. The resulting agent handles what you'd otherwise spend months coding around.

Less manual workarounds. More automation coverage you can trust. That’s what operationalizing agents should look like in production.

Watch the full episode: https://www.itential.com/resource/podcast/cloud-gambit-agentic-ops-how-infrastructure-teams-scale-reasoning-without-losing-control/?utm_source=Facebook&utm_campaign=itential&utm_medium=social