Meet Azi
Azi is the AI interface to your OpenIndustrial workspace. She has Total System Awareness - understanding not just your data, but your architecture, relationships, and operational context.
Other AI sees fragments. Azi sees your whole system. That total awareness enables intelligent proposals. The governance ensures those proposals require your approval. We invented this pattern.
The Four Surfaces
Azi operates across the Four Surfaces of the Control Plane:
| Surface | What Azi Does | Why It Matters |
|---|---|---|
| OBSERVE | Watches data streams, detects patterns, understands schemas | AI that sees what you see |
| PROPOSE | Generates queries, configurations, and actions | Recommendations, not actions |
| GOVERN | Queues every proposal for your review | Every action audited |
| QUERY | Translates natural language to KQL | Ask anything, safely |
The Four Surfaces aren't just our framework - they're the framework for governed AI. We invented this architecture.
Azi is MCP-Native
Azi communicates via Model Context Protocol (MCP) - the industry standard for AI integration:
- Today: Claude natively, GPT/Copilot via adapters
- Tomorrow: Any AI that adopts MCP plugs right in
- Your choice: Use whatever MCP-native AI fits your needs
This is "Bring Any AI" in action. We bet on the standard, not on any single vendor. When better AI arrives, it works with your Control Plane. Fire us and keep your AI connections.
Azi Context Levels
Azi adapts to where you are in the workspace. Each context level gives Azi different knowledge and capabilities:
| Context | What Azi Knows | Panel Title | Example Questions |
|---|---|---|---|
| Workspace | All connections, surfaces, high-level structure | "AZI'S UNDERSTANDING" | "What connections do I have?" |
| Surface | Data flowing to this surface, queries within it | "AZI'S UNDERSTANDING" | "Create a warm query here" |
| Query Manager | Specific query, KQL, results history | "WARM QUERYS WITH AZI" | "Update this to filter by temp > 70" |
The Azi panel title changes based on context. "AZI'S UNDERSTANDING" at workspace/surface level, "WARM QUERYS WITH AZI" inside the Query Manager.
What Azi Can Propose
Azi proposes actions across the system:
| Category | Examples |
|---|---|
| Connections | Create new data connections, configure protocols |
| Surfaces | Create execution boundaries, route data flows |
| Queries | Build KQL, create warm queries, add filters |
| Configurations | Set up alerts, define thresholds, annotate fields |
Every proposal requires your approval.
Fire Azi (and Keep Running)
If you remove Azi, your system keeps running. Your data stays yours. Your queries keep executing. Azi is a teammate - not a tether.
What stays:
- All your data in your Azure tenant
- All configured connections and surfaces
- All warm queries and their KQL
- Full audit trail of past decisions
What goes:
- Azi's ability to propose new actions
- Natural language query translation
- AI-assisted configuration
You can always re-enable Azi later - or use a different MCP-native AI.
Learn More
Understand the governance loop in detail - how Azi proposes, how you review, and why the audit trail matters.