There’s a version of every product launch where the team writes a blog post and moves on. This isn’t one of those launches.
This week, the most important conversation in enterprise networking is happening in Las Vegas. Cisco Live US 2026 has become the stage where the next era of network operations is being defined, and we couldn’t think of a better moment to take the wraps off what we’ve been building.
Today, we’re announcing the general availability of FlowAI, with broad availability beginning July 1, and a live debut on the Cisco Live show floor all week. After six months in private preview with leading enterprises across telecom, financial services, and utilities, FlowAI is ready for production. AI agents that reason, plan, and act on real enterprise infrastructure, inside the same governance, audit, and execution controls our customers have trusted Itential to deliver for nearly a decade.
That second part is what makes today different.
Every enterprise we’ve talked to over the last year is being asked the same question: what are we doing with AI? And underneath that question is a quieter one that security and operations teams are asking before anyone signs off on anything: how do we let a probabilistic system touch production infrastructure without breaking it?
That question is why most agentic AI for infrastructure is still stuck in sandboxes and chatbots. The market has split into two camps, and neither one solves the problem.
On one side, there are platforms for building agents. Wire up an LLM, define some tools, give it a prompt. Easy to demo. Impossible to deploy. There’s no governed execution layer underneath, so the agent can either suggest things to a human or call an API and hope for the best. No RBAC. No audit trails. No approval gates. No rollback. Not safe for production infrastructure, and the security teams know it.
On the other side, there are platforms for deterministic execution. Mature, governed, battle-tested. The kind of platforms that have run carrier and bank infrastructure for years. But they were built before agents existed. The most they can do is let an AI via MCP call a workflow as an external trigger. The agent isn’t part of the platform. It’s a guest, not a citizen.
Neither model works. Build-only platforms can’t be trusted in production. Execute-only platforms can’t reason. Enterprises end up duct-taping the two together and calling it agentic operations.
Itential is the only platform that does both. You build agents inside the same platform that executes them. The reasoning layer and the execution layer share one identity model, one governance model, one audit trail, one set of policy controls. Agents aren’t bolted onto automation. They aren’t connected to automation. They are a native execution model inside the same governed engine that already runs the world’s most regulated infrastructure.
That’s not a feature comparison. It’s an architecture comparison. And it’s the reason FlowAI is in production today while most agentic AI for infrastructure is still a demo.
FlowAI runs on top of that machinery. Not next to it. Not bolted on. Through it.
The Itential Platform is the agentic operations platform for infrastructure teams. Every action flows through one governed engine, whether the trigger is a human, a workflow, an API call, or an AI agent.
FlowAI is what makes the platform agentic:
Agents reason and execute. Every action governed by default. No exceptions for AI.
This isn’t a roadmap. It’s already running in production.
Lumen is building production-ready agents in minutes, inside the same governance and access controls they’ve relied on for years. A Tier-1 service provider called it the best POC they’ve ever done. Engineers across the FlowAI Innovation Program built complex production agents in under two minutes. Platform teams prototyped agents for software upgrades in 30 minutes – work that would have taken months through scripts and custom code.
Across every customer in the innovation program, the same operating pattern emerged: teams started by building agents fast – in minutes, not months – to reason through complex multi-vendor, multi-domain tasks that had previously taken weeks of scripting and tribal knowledge. Once an agent proved out the logic, teams converted that reasoning into deterministic workflows for scale, using spec-driven development – with the same governance, the same audit trail, the same execution engine. Reason first. Operationalize next. FlowAI is what makes that progression possible without sacrificing control.
Agentic operations isn’t a future trend. It’s a present reality, already running in production at some of the most demanding enterprises in the world. The platforms that win the next decade will be the ones that already knew how to govern production infrastructure before AI showed up. Everyone else is retrofitting.
Itential isn’t retrofitting. We’ve spent ten years building the foundation, and FlowAI is what runs on top of it. Today that foundation – paired with agentic reasoning, MCP-based interoperability, and the autonomy controls enterprises actually need.
All week in Las Vegas, the Itential team is showcasing FlowAI in action with live demos, customer conversations, and real production scenarios. If you’re at Cisco Live, here’s where to find us:
We’ll be at AutoCon 5 in Munich the following week. And if you can’t make either event, we’d still love to talk. The conversation about how agentic operations is reshaping the infrastructure landscape is the most interesting one happening in our industry right now.
FlowAI is here. The foundation was already built. Let’s go.
See how Itential connects AI reasoning to governed execution across your entire infrastructure.
