Itential introduces FlowAI and outlines layered AI infrastructure model

Itential presents a layered operating model for Artificial Intelligence (AI) in network and infrastructure operations and previews FlowAI, a framework intended to add AI reasoning while preserving deterministic execution and operational control.

Research Overview

At AutoCon 4, Itential’s chief architect described a structured approach for introducing AI into network and infrastructure workflows. The presentation framed the proposal as a three-layer model linking existing instrumentation and orchestration with an added AI reasoning layer.

Key Findings

The session identified three core requirements for AI in infrastructure: secure deployment, governance and traceability, and maintaining deterministic behavior where needed. Itental also highlighted the need to integrate AI without assuming native AI interfaces on existing devices.

Technical Breakdown

The model starts with an instrumentation layer that captures device state and exposes operational interfaces, followed by a deterministic execution layer for repeatable orchestration. A separate AI reasoning layer consumes context and model outputs to propose or enact changes while relying on audit trails and control gates.

Product Update

Itential announced FlowAI, a framework aligned to the layered model that emphasizes instrumentation and gateway support for Model Context Protocol (MCP) servers. FlowAI includes an agent builder and flow agents intended to execute changes through the layered controls described in the session.

Operational Impact

The presenter described a staged transition from Human-in-the-Loop (HITL) to human-on-the-loop and, where appropriate, to autonomous operations, with operators retaining the ability to intervene. He recommended designing smaller, discrete workflows to improve safety and control when agents act on infrastructure.

Leadership Perspective

Itential’s speaker framed AI as a tool to be applied where it fits and emphasized preserving security and auditability as primary constraints. He positioned the FlowAI framework as a practical implementation aligned with those constraints and invited technical discussion at the event.

The overall takeaway is that Itential proposes a layered, control-first method for adding AI reasoning to infrastructure operations, and the presentation previews FlowAI as an implementation of that method; this “Blog Signals brief” is a fact-based summary of the vendor blog.