Hybrid Track: Govern Without Replacing

You already have an industrial backbone. This path adds governance without disruption.

---

What You'll Achieve

By the end of this track, you'll have:

• Governed queries over your existing data
• Full audit trail of every query execution
• API endpoints for integration with LIMS, dashboards, or AI tools
• MCP connection so Claude or Copilot can query your data

All without touching your existing control systems.

---

The Approach

Layer	What Stays	What's Added
Control	Your SCADA, DeltaV, PLCs	Nothing—we don't touch this
Data	Your historians, databases	IoT Hub connection to OI
Query	Ad-hoc SQL, manual reports	Governed warm queries
AI	Scattered tools	MCP-native AI on governed data

---

Phase 1: Connect Your Data

Step 1.1: Create IoT Hub Connection

Option A: Using the UI

1. Open your workspace
2. Drag a Connection node from the Node Bank onto your canvas
3. Click the node to open the Inspector
4. In Settings, configure for Azure IoT Hub
5. Enter your connection string
6. Commit your changes (click the save icon in the top-left)

Option B: Ask Azi

Step 1.2: Configure Data Flow

Your historian or SCADA system forwards data to IoT Hub. OI receives it as a stream:

Source	Protocol	OI Connection Type
OSIsoft PI	MQTT via Edge	IoT Hub
Aveva Historian	OPC-UA via Edge	IoT Hub
Rockwell FactoryTalk	MQTT	IoT Hub
Custom sensors	Direct MQTT	IoT Hub

---

Phase 2: Create Monitoring Surface

Step 2.1: Create Surface

Option A: Using the UI

1. Drag a Surface node from the Node Bank onto your canvas
2. Click the node to open the Inspector
3. In Settings, name it: Cleanroom-Monitoring
4. Draw a connection line from your IoT Hub connection to the surface
5. Commit your changes

Option B: Ask Azi

Step 2.2: Verify Data Flow

1. Open the surface
2. Check the Inspector panel
3. Confirm live data is flowing

You should see your temperature, pressure, and batch data appearing in real-time.

---

Phase 3: Build Compliance Queries with Azi

This is where governance becomes real. Every query you build is tracked, versioned, and auditable.

Step 3.1: Create Temperature Compliance Query

Ask Azi:

"Create a warm query that checks if cleanroom temperature stayed between 18-22°C in the last 24 hours."

Azi proposes:

TelemetryData
| where Timestamp > ago(24h)
| where Zone == "Cleanroom"
| summarize
    MinTemp = min(Temperature),
    MaxTemp = max(Temperature),
    AvgTemp = avg(Temperature)
| extend Compliant = MinTemp >= 18 and MaxTemp <= 22

You review, then approve or edit.

Step 3.2: Create Batch Tracking Query

Ask Azi:

"Build a query that shows all events for a specific batch ID, ordered by time."

Azi proposes:

BatchEvents
| where BatchID == "{batchId}"
| order by Timestamp asc
| project Timestamp, EventType, Zone, Operator, Details

Step 3.3: Create Threshold Alert Query

Ask Azi:

"Create a query that flags any pressure readings above 1.5 bar in the mixing zone."

Azi proposes:

TelemetryData
| where Zone == "MixingZone"
| where Pressure > 1.5
| project Timestamp, Pressure, DeviceId
| order by Timestamp desc

---

Phase 4: Access Your APIs

Every warm query automatically gets a REST API endpoint.

Step 4.1: Find Your Endpoint

1. Open the APIs menu in the top bar
2. Select Warm Query Manager
3. Choose your warm query from the dropdown
4. Copy the endpoint URL and code examples

Step 4.2: Test with cURL

curl -X GET 'https://www.openindustrial.co/api/workspaces/explorer/warm-queries/temp-compliance' \
  -H 'Authorization: Bearer YOUR_TOKEN'

Step 4.3: Integrate with Your Systems

System	Integration
LIMS	REST API call on batch completion
Dashboard	Polling or webhook
Alerting	Trigger on threshold queries

---

Phase 5: Connect MCP-Native AI

The "aha moment"—any MCP-native AI can now query your governed data.

Step 5.1: Get MCP Configuration

1. Open the APIs menu in the top bar
2. Select API Keys to generate a JWT token
3. Use the configuration below with your workspace ID and token

Step 5.2: Configure Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "pharma-factory": {
      "command": "npx",
      "args": [
        "-y",
        "@openindustrial/mcp-server",
        "--workspace", "YOUR_WORKSPACE_ID",
        "--token", "YOUR_TOKEN"
      ]
    }
  }
}

Step 5.3: Ask Questions

Now you can ask Claude:

"Was batch 112A compliant with temperature requirements?"

Claude calls your governed warm query and returns audited results.

---

What You Built

Component	Status
IoT Hub connection	✓ Connected to existing historian
Cleanroom surface	✓ Organizing your data
Compliance queries	✓ 3+ governed warm queries
API endpoints	✓ REST access to all queries
MCP integration	✓ Claude querying governed data

---

Category Metrics

Metric	Your Result
Time to First Governed Query	~30 minutes
Queries built	3+
Audit coverage	100%
AI connected	Claude Desktop

---

Before vs After

Before Control Plane	After Control Plane
"AI said it was fine"	"Here's the query, audit log, and approval"
Data scattered in silos	Unified through governed surfaces
No query traceability	Full history of every execution
AI locked to vendors	Any MCP-native AI works

---

Next Steps

If you want to...	Go to...
Try the Full track	Full Track →
See query patterns	Patterns →
Deep-dive on MCP	MCP Guide →
Learn more about Azi	Meet Azi →