Industrial AI · Food ManufacturingIllustrative demoDemo
AI-Powered Manufacturing Operations AssistantManufacturing Operations Assistant
An AI assistant that investigates production downtime by combining operational data, maintenance knowledge, inventory, and equipment telemetry.
System architecture
Step 1 / 7Project implementation
Existing enterprise systems
Web UI
React · TypeScript
AI Orchestrator
NestJS · TypeScript
External AI service
External LLM
LLM API · Tool Calling
PostgreSQL
PostgreSQL · SQL
RAG Service
Enterprise Knowledge · REST API
Inventory API
Spare Parts · REST API
IoT Telemetry API
Industrial Telemetry · REST API
Project implementation
Web UI
React · TypeScript
AI Orchestrator
NestJS · TypeScript
External AI service
External LLM
LLM API · Tool Calling
Existing enterprise systems
PostgreSQL
SQL
RAG Service
REST API
Inventory API
REST API
IoT Telemetry API
REST API
All four systems connect independently to the AI Orchestrator.
Step 1 — User request
User Request
1. Enter Question
2. Send Request
POST/api/investigations
{
"question": "Why did maintenance downtime increase on Packaging Line 3 this quarter?"
}3. Request Accepted
The AI Orchestrator validates and accepts the investigation request.
Investigation started
4. Start AI Investigation
The AI Orchestrator sends the user's question, system instructions, and available tool definitions to the external LLM.
User question
Why did maintenance downtime increase on Packaging Line 3 this quarter?
System instructions
Use approved tools to retrieve evidence. Do not invent operational data.
Available tools (7)
PostgreSQL
- resolve_equipment
- get_downtime_statistics
- get_maintenance_faults
RAG Service
- search_knowledge
Inventory API
- get_spare_parts
- get_inventory_movements
IoT Telemetry API
- get_telemetry
User Request