Share this
Table of Contents
At AutoCon 4, we introduced FlowAI to the community. It was the first time we showed how an agentic operating model can be applied to real infrastructure with governance, determinism, and security at the center rather than bolted on after the fact. The energy in Austin made something clear. Teams are ready for AI in their operations, but only if it can be trusted. Only if it works with the systems they rely on every day. Only if it respects the realities of production infrastructure.
FlowAI is designed for exactly that world.
Over the past year, we have been working toward a simple idea. Give infrastructure teams a safe way to attach an AI reasoning layer to the Itential Platform so they can operate networks and hybrid environments with far more intelligence and far fewer manual burdens. The conversations at AutoCon confirmed that this is not a future concept. This is what the industry has been waiting for.
Below is a deeper look at what FlowAI is, how the framework works, and why the components matter.
What Is FlowAI?
FlowAI is a framework of applications inside the Itential Platform and Itential Automation Gateway that allows teams to create, deploy, and govern agents that can take action on infrastructure. These agents are not abstract assistants. They are fully governed operational entities that understand intent, reason through the required steps, and execute through the deterministic automation engine, the Itential Platform, that customers already trust.
FlowAI does three things that teams consistently asked for at AutoCon 4:
- It connects AI reasoning to real infrastructure actions with full control.
- It gives teams the power to build their own agents using natural language.
- It integrates external vendor agents through MCP while applying the same safety boundaries.
The result is a unified environment for agentic operations that keeps humans informed, maintains deterministic execution, and expands what automation can achieve.
The FlowAI Framework
FlowAI consists of a coordinated set of applications that together define, govern, and operate agents across hybrid infrastructure. Each part of the framework plays a specific role in how an agent is created, controlled, observed, and extended.
FlowAgent Builder
FlowAgent Builder is the design environment where new agents are authored. Teams describe an outcome or task in natural language and then shape the agent’s persona, reasoning model, scope, and access. The Builder combines these elements into a fully structured agent blueprint that can operate safely within production environments. This approach lets teams create powerful agents while keeping governance and architectural clarity intact.
FlowAgents
FlowAgents are the agents themselves, the cognitive layer that performs work across networks and cloud environments. Each agent carries a defined persona, intent, and set of skills and capabilities. When activated, a FlowAgent uses reasoning to interpret the task, evaluates the available tools, and executes changes through the deterministic automation engine. These agents are purpose built, governed, and designed for real infrastructure operations.
FlowAgent Manager
FlowAgent Manager is the operational control plane for all agents running inside the Platform. It is where teams set boundaries around what an agent can access, how it behaves, and which capabilities it is permitted to use. This is also where operators gain full visibility into agent status, mission history, permissions, and operational posture. FlowAgent Manager ensures that every agent operates within clear, enforceable policy.
Agent Missions
Agent Missions provides complete observability into how agents behave once they are deployed. Every mission includes detailed traces of the agent’s reasoning, tool selection, workflow execution, and final result. This level of telemetry gives teams confidence in how agents make decisions and how those decisions translate into infrastructure action. It is the transparency layer that supports trust and continuous improvement.
Agent Tools
Agent Tools defines the action surface that agents can use. This catalog includes workflows, automations, API Integrations, and MCP connected capabilities. By curating this set of tools, teams determine exactly what agents are allowed to do. This ensures that every action taken by an agent remains safe, predictable, and consistent with enterprise rules.
How FlowAI Connects to MCP
At AutoCon 4, one topic kept coming up. Teams know a wave of vendor-built agents is coming, and they want the freedom to use them, but not at the cost of losing governance. FlowAI delivers that by combining capabilities across both the Itential Platform and the Itential Automation Gateway.
FlowMCP Gateway extends the Itential Automation Gateway to securely invoke external infrastructure agents and MCP compatible tools such as NetBox MCP or Selector MCP. It gives the platform a safe, structured way to call into third party agent ecosystems without exposing infrastructure directly. The gateway brings external intelligence under the same governance umbrella customers already rely on. Every outbound request, every response, and every tool invocation passes through the Automation Gateway’s controls, data handling rules, and execution guardrails. External MCP tools become callable, but never unbounded.
FlowMCP Server handles the inbound side. It is the enterprise grade version of the open source MCP Server we launched in May 2025, expanded with the capabilities large organizations need. FlowMCP Server provides centralized management of multiple MCP instances through virtual MCP servers, full persona-based access control, and multiuser authentication and authorization. It receives intent from any external agent, validates the request, applies policy, and translates the intent into structured actions the Itential Platform can execute deterministically. It keeps reasoning outside the execution boundary while giving enterprises a unified control point for every MCP connected agent.
Together, FlowMCP Gateway and FlowMCP Server form the bridge between FlowAI and the emerging ecosystem of vendor agents. They allow internal FlowAgents and external MCP agents to operate side by side, all within a single governed and deterministic execution model. For many teams at AutoCon 4, this was the unlock. You do not have to choose between innovation and safety. You can bring outside intelligence into your automation strategy without weakening your operational posture.
Why Determinism Still Matters
FlowAI brings AI reasoning into infrastructure, but it never hands execution over to an unpredictable model. Reasoning happens in the agent. Execution happens in the deterministic automation engine (the Itential Platform) that has been hardened over more than a decade. This separation is essential for production infrastructure. It is the difference between experimentation and operational reality.
Operators can trust that every change remains traceable, reversible, and controlled. AI provides the intelligence. Itential provides the safety.
What Comes Next
AutoCon 4 made something unmistakably clear. The industry is ready to move past slideware visions of AI and into real operational models that respect how infrastructure actually works. The conversations last week were some of the most direct and energizing I have had in years. Teams want intelligence with boundaries. Flexibility with determinism. Innovation that does not compromise safety. FlowAI was built for exactly that world, and the reaction in Austin showed that the momentum is already here.
We are still early in this shift, but the direction is set. FlowAI gives organizations a practical, governed path to bring reasoning and automation together in a way that elevates operations rather than disrupts them. The next phase will be shaped by the teams putting this technology to work, the vendors contributing new agents, and the community pushing the model forward.
If last week was any indication, the energy is real and the possibilities are wide open.
So here is to building the future of agentic operations together. And as always, keep automation weird.
Check Out FlowAI on Packet Pushers
I recently joined Ethan Banks from Packet Pushers on the Tech Bytes podcast for a deeper technical walkthrough of Itential’s FlowAI release. We dug into how FlowAI brings AI reasoning together with deterministic, governed workflows, why that pairing matters in real production infrastructure, and how teams can adopt agentic operations without giving up control or governance.
If you’re looking for a clear, engineer-level view of what FlowAI actually does, this conversation delivers.
Listen to the podcast episode below or dive into the complete episode details here. Plus, don’t miss Packet Pushers’ full blog on the launch of Flow AI here.
Peter Sprygada
Chief Architect ‐ Itential
Peter Sprygada serves as the Chief Architect at Itential after serving as the Chief Technology Officer at Pureport where he was responsible for their multi-cloud network as a service interconnect platform. Prior to Pureport, Sprygada was a Distinguished Engineer for Red Hat, where he played the role of Chief Architect for the Ansible Automation Platform. Sprygada also held senior technical and leadership positions at Arista and Cisco, as well as several networking startups.
