Kudzu Canopy NOC includes an AI-assisted analysis layer that looks across KPIs, alarms, infrastructure state, map references, and operator notes to produce actionable findings. The platform generates network insight problems, proposes solutions, runs agent sessions, builds reports, and offers interactive investigation through the Nomi chatbot. Together, these capabilities help you move from “something is wrong” to “here is what to do about it.”
Metrics and alarms tell you that a condition exists, but they do not explain why it happened or what you should do next. The AI layer bridges that gap by correlating evidence across the network, identifying root causes, and generating structured recommendations with feasibility scores and implementation guidance. This reduces the time between detection and resolution, especially for complex multi-factor problems that span devices, gateways, zones, and integrations.
A network insight problem is a structured finding about a network condition. It represents a specific operational issue that the AI has identified by analyzing KPIs, alarms, map data, and infrastructure state. Each problem includes a title, a detailed description, supporting evidence, and references to the affected entities — devices, gateways, metrics, zones, H3 cells, or integrations.
A network insight solution is a proposed action for addressing a network insight problem. Solutions include:
A single problem can have multiple candidate solutions. You choose the most appropriate one based on your operational context.
Agent sessions are interactive or scheduled reasoning sessions where the AI gathers evidence, runs analysis, and produces outputs. Sessions can be triggered automatically by the platform or started manually when you use the Nomi chatbot or the Investigate action on an alarm. Each session maintains context about the evidence it has examined and the conclusions it has reached.
Reports are longer-form generated documents that summarize a network’s operational state over a reporting period. Because report generation can involve data gathering, analysis, and AI writing, reports are produced asynchronously.
Reports move through a defined lifecycle:
| State | Meaning |
|---|---|
| Draft | The report has been requested but generation has not started. |
| Rendering | The platform is gathering data and generating content. |
| Failed | Generation encountered an error. You can retry. |
| Needs Input | The AI requires additional context before continuing. |
| Completed | The report is ready to view, edit, or print. |
Completed reports can include supporting charts and dataset evidence alongside prose, giving readers traceable findings rather than unsupported conclusions.
Nomi is the platform’s built-in AI assistant. You can access Nomi in three ways:
/ keyboard shortcut from anywhere in the interface.When you open Nomi from a Get Insights button, the chatbot receives the full context of the specific item — the metric’s history, the gateway’s state, the alarm’s affected entities — so you can ask targeted questions without having to explain what you are looking at.
Use the Get Insights button on an alarm’s metric detail page to start an AI-assisted investigation session with full context about the alarm, its metric history, and the affected entities.
| Entry point | What happens |
|---|---|
| Action Center suggestions | AI-generated optimization suggestions grouped by underlying problem, with solutions, checklists, and entity references. |
| Get Insights button | Opens Nomi pre-focused on a specific metric, device, gateway, or alarm. |
| Investigate button (Alarms) | Starts an interactive AI debugging session for a specific alarm. |
| Reports page | Lists generated and in-progress reports with status filters and regeneration controls. |
| Chat with Nomi | Opens a general-purpose conversation with the AI assistant. |