Project Title: Inventory Management Automation System (n8n + AI Agents + PostgreSQL)

Manual inventory management breaks down for a predictable reason: nothing keeps the stock count, the spreadsheet, and reality in sync once things start moving fast. Businesses end up overselling, missing reorder points, or making decisions on outdated numbers.

I designed and built an end-to-end inventory automation system to solve this, structured as three independent modules rather than one large workflow:

1. Real-Time Inventory Engine Processes every inventory event sales, returns, damage, manual adjustments as it happens. Validates the event, updates stock, logs the movement for auditability, and checks reorder thresholds. When stock drops below the reorder point, it automatically creates a purchase request, routes it through an approval step, and notifies the supplier once approved.

2. Inventory Intelligence Pipeline Runs on a schedule, fully decoupled from live operations. Pulls inventory, movement, and purchase data, calculates KPIs, generates AI-driven insights, and delivers an HTML dashboard and PDF report automatically.

3. Conversational Inventory AI An AI agent that answers natural-language questions about inventory by querying PostgreSQL directly with read-only queries answers come from real data, not estimation.

Each module has different triggers, failure tolerances, and scaling needs, which is why they're built to run independently rather than as one workflow. PostgreSQL sits underneath all three as the single source of truth, so the dashboard, the AI answers, and the actual stock count never drift apart.

My Role: End-to-end design and development architecture, workflow logic, database schema, AI agent integration, and reporting pipeline.

Tools & Technologies: n8n · PostgreSQL · Gemini / OpenAI · JavaScript · Webhooks · HTML/PDF Reporting · Email Notifications

Links: GitHub repository: https://github.com/anshudainsights3/inventory-automation.N8N/tree/main — full n8n workflow JSON, database schema, SQL scripts, architecture diagrams, and README.