Automated Logistics & Procurement

The problem

Ripple Collective runs six companies out of shared warehouses and production facilities. That sounds efficient until you realize procurement was happening over WhatsApp threads, approvals lived in someone's inbox, and delivery notes were filled out by hand with no way to trace who ordered what for which entity. One company's urgent restock would collide with another's bulk order, and nobody had visibility into what was already in transit. Route planning was done on gut instinct, which meant drivers were crisscrossing the city when a single optimized loop would have cut fuel costs in half.

The approach

I built a system that gave each company its own procurement flow while keeping everything under one operational roof. The key was making the shared infrastructure feel dedicated to each team without duplicating resources.

• Designed a multi-tenant procurement pipeline where each company submits requests that route to the correct approver based on entity, category, and amount thresholds
• Built automatic delivery note generation that pulls from approved purchase orders, tags items by destination warehouse, and produces printable documents with QR codes for receiving confirmation
• Created a route optimization engine that batches deliveries across all six companies, clusters stops by geography, and generates driver-ready sequences that minimize total distance
• Added real-time inventory visibility so procurement managers can check what's already on hand before submitting a new request, preventing duplicate orders across entities
• Implemented approval escalation logic: if a request sits untouched for more than 24 hours, it bumps up the chain automatically instead of dying in someone's notifications

The outcome

Procurement cycles dropped from days to hours. The approval bottleneck essentially disappeared because managers could review and approve from their phones with full context attached. Delivery notes went from a source of confusion to a reliable audit trail. Drivers stopped calling the office asking where to go next. The route optimization alone saved enough on fuel and time that it justified the entire project within the first two months. Most importantly, the six companies stopped stepping on each other's toes when it came to shared warehouse space and overlapping supplier relationships.

What I learned

Multi-tenant systems are deceptively hard. The technical challenge isn't building the features, it's modeling the permissions and boundaries correctly so that shared resources don't create shared headaches. I also learned that the people closest to the problem (warehouse staff, drivers, procurement clerks) will tell you exactly what's broken if you actually watch them work for a day instead of just collecting requirements in a meeting room.