Full Track: Author Your Production Intelligence

You chose complete control. Build your production system from first principles.

---

What You'll Build

By the end of this track:

Component	Scope
Connections	All production data sources
Surfaces	Plant → Line → Cell → Station hierarchy
Warm queries	Complete OEE, quality, downtime library
Team	Roles for ops, quality, maintenance, engineering
AI integration	Full MCP connection

---

Phase 1: Design Production Architecture

Step 1.1: Map Your Hierarchy

Plant
├── Line 1
│   ├── Cell A (Machining)
│   │   ├── Station 1
│   │   └── Station 2
│   └── Cell B (Assembly)
│       ├── Station 1
│       └── Station 2
├── Line 2
│   └── Same structure
└── Support
    ├── Quality Lab
    └── Maintenance

Step 1.2: Identify Data Sources

Level	Data Source	Metrics
Station	PLC/Sensor	Cycle time, state, counts
Cell	Cell controller	Throughput, WIP
Line	MES	OEE, yield, batch tracking
Plant	Aggregation	Total output, efficiency

Step 1.3: Create Connections

Ask Azi:

Azi proposes all connections. Review and approve each.

---

Phase 2: Create Surface Hierarchy

Step 2.1: Plant Surface

Ask Azi:

"Create a plant-level surface that aggregates all line production data."

Azi proposes:

• Surface name: Plant-Production
• Nodes: Line1Total, Line2Total, QualityOverall
• Description: Plant-wide production aggregation

Step 2.2: Line Surfaces

Surface	Nodes	Purpose
Line-1	CellA, CellB, LineOEE	Line 1 detail
Line-2	CellA, CellB, LineOEE	Line 2 detail

Step 2.3: Cell and Station Surfaces

For deep operational visibility:

Surface	Nodes	Purpose
Line-1-Cell-A	Station1, Station2	Machining detail
Line-1-Cell-B	Station1, Station2	Assembly detail

---

Phase 3: Build Comprehensive Query Library

Standard Production Queries

Query	Purpose	Level
plant-oee	Overall plant OEE	Plant
line-oee	Line-specific OEE	Line
station-cycle	Station cycle analysis	Station
downtime-pareto	Top downtime causes	Plant/Line
quality-fpy	First pass yield	Line
throughput-hourly	Hourly production	Line

Step 3.1: Plant OEE Dashboard Query

Ask Azi:

"Create a query that calculates real-time OEE for the entire plant, broken down by line."

Azi proposes:

ProductionData
| where Timestamp > ago(1d)
| summarize
    AvailableTime = sum(iff(Status != "Planned Downtime", Duration, 0)),
    RunTime = sum(iff(Status == "Running", Duration, 0)),
    ActualUnits = sum(Units),
    GoodUnits = sum(iff(Quality == "Pass", Units, 0))
    by Line
| extend IdealCycleTime = 30 // seconds - adjust per line
| extend Availability = RunTime / AvailableTime
| extend Performance = (ActualUnits * IdealCycleTime) / (RunTime)
| extend Quality = GoodUnits / ActualUnits
| extend OEE = Availability * Performance * Quality * 100
| project Line, Availability = round(Availability * 100, 1),
          Performance = round(Performance * 100, 1),
          Quality = round(Quality * 100, 1),
          OEE = round(OEE, 1)

Step 3.2: Station-Level Analysis

Ask Azi:

"Build a query that identifies bottleneck stations based on cycle time variance."

StationData
| where Timestamp > ago(8h)
| summarize AvgCycle = avg(CycleTime),
            StdDev = stdev(CycleTime),
            Count = count()
    by Station
| extend CV = StdDev / AvgCycle // Coefficient of variation
| order by CV desc
| project Station, AvgCycle, StdDev, CV, Count

Step 3.3: Quality Traceability

Ask Azi:

"Create a query that traces all process steps for a specific lot number."

ProductionData
| where LotNumber == "{lotId}"
| join kind=leftouter QualityData on LotNumber
| project Timestamp, Step, Station, Operator,
          Measurement, Status, InspectionResult
| order by Timestamp asc

---

Phase 4: Establish Team Permissions

Step 4.1: Define Roles

Role	Access	Scope
Engineering	Full access	All surfaces, all queries
Operations	Line surfaces, run queries	Production data
Quality	Quality surfaces, edit	QC-specific queries
Maintenance	Downtime surfaces	Equipment-focused
Management	Plant surface, view	Dashboards only

Step 4.2: Invite Team Members

Ask Azi:

"Invite quality-manager@company.com as Editor for Quality surfaces."

Step 4.3: Configure Surface Permissions

Surface	Admin	Editor	Viewer
Plant-Production	Engineering	Operations	Management
Line-1	Engineering	Operations	-
Quality-Control	Engineering	Quality	Operations

---

Phase 5: Connect External AI Tools

Step 5.1: Generate MCP Configuration

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

Step 5.2: Configure Claude Desktop

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

Step 5.3: AI-Powered Analysis

Ask Claude:

"Why did Line 1 miss its production target yesterday? Show me the contributing factors."

Claude queries OEE components, downtime events, and quality data—then correlates them to identify root cause.

---

What You Built

Component	Count
Connections	3+
Surfaces	6+ (plant → station)
Warm queries	6+
Team roles	Configured
MCP integrations	1+

---

Category Metrics

Metric	Your Result
Time to First Governed Query	~1 hour
Query library	6+ queries
Surface hierarchy	4 levels
Audit coverage	100%
AI connected	MCP-native

---

What Governed Production Intelligence Looks Like

Before Control Plane	After Control Plane
Different OEE per system	One definition, one number
Tribal knowledge on metrics	Documented, governed queries
AI recommendations without context	AI insights with full data lineage
Manual root cause	Data-driven, traceable analysis

---

Next Steps

If you want to...	Go to...
Try the Hybrid track	Hybrid Track →
See query patterns	Patterns →
Learn more about MCP	MCP Guide →
Explore pharma	Pharma Journey →