Order-to-Dispatch System

The plant made thousands of precision brass parts a day and managed them on spreadsheets and a whiteboard. Jamnagar is India’s brass parts capital, and this manufacturer turns brass rod into inserts, fittings and terminals for electrical and plumbing companies in India and abroad.

Orders arrived by email and phone and were tracked in spreadsheets. Production was planned on a whiteboard, supervisors walked the floor to find out where an order had reached, and dispatch dates were promised without knowing whether the machines were free. Brass is expensive, but nobody could say how much rod had gone into an order or how many turnings had come back. Quality reports were on paper, and every invoice was typed again into the accounting software, just as GST e-invoicing made those invoices harder to get wrong.

We spent the first month following real orders from purchase order to truck, through sales, planning, stores, the machines, quality and dispatch.

• Delivery dates were guesses. Sales couldn’t see machine loads, so promises and capacity rarely matched.
• Brass was counted by the tonne, not by the batch. Losses only showed up at stock counts, too late to know where they happened.
• The real bottleneck moved every week, from one set of CNC machines to plating to inspection, and nobody could see it coming.
• Supervisors were the plant’s memory. When one was absent, the status of half the floor went with them.

Every order starts as a sales order entered once, from the buyer’s purchase order: part number, drawing revision, quantity, finish and requested date. Repeat parts carry their routing, the list of operations and machines each part needs, so the system knows the work behind every order.

As the order is entered, the system checks brass rod stock for the right alloy and size, and the free capacity on every machine the part needs. It then suggests a dispatch date. Sales promises dates the plant can keep, and when a buyer asks to pull an order forward, the planner sees exactly what moves to make room.

Planners schedule work by machine and shift on one production board. Each machine shows its load for the coming days, changeovers are planned to group similar parts, and the board warns when an order is at risk of missing its date.

When a machine breaks down or an urgent order arrives, the planner moves jobs on the board and every affected date updates. The bottleneck of the week is visible on Monday, not on Friday.

Supervisors work from tablets. Every batch carries a QR code from the moment rod is issued, and scanning it opens the digital job card for the next operation: bar cutting, CNC turning, drilling and threading, and finishing.

The job card shows the drawing, the machine and the checks for that operation. Supervisors record good output, rejections with their reasons, and time. Inspectors record dimensional checks at first-piece, in-process and final inspection, with measurements and photos, and a batch can’t move on until its inspection is passed.

Brass is the plant’s biggest cost, so the system tracks it to the gram:

• rod issued to each batch, by alloy, size and lot;
• good parts produced, by count and weight;
• turnings recovered from each machine, weighed and returned to stores;
• the gap, shown for every batch, so a loss is found the day it happens.

Parts that go out for nickel or tin plating travel with a GST job work challan generated by the system. Quantities and weights are tracked out to the plater and back, and the details for the ITC-04 job work return are ready when it is due.

At dispatch, the system prepares the packing list from the batches that passed final inspection, with each box labelled by part, quantity and batch. It generates the GST e-invoice and the e-way bill through an authorised GST API provider, and sends the invoice to the accounting software, Tally, so nothing is typed twice. Buyers receive dispatch details, with test certificates where they need them, the same day.

The system is a Laravel and MySQL application with a Vue.js interface for planners, sales, stores and accounts, and a Flutter app on rugged tablets for the shop floor. Redis keeps the production board live across screens, and a Python service calculates capacity and suggested dates. GST e-invoices and e-way bills are generated through an authorised provider’s API, invoices flow to Tally, and everything runs on AWS, with the shop-floor tablets working through short network drops.

The plant moved over department by department. Orders and planning ran alongside the old spreadsheets for a month until planners trusted the board. The shop floor followed, one machine group at a time, with supervisors trained on the tablets beside their machines. Dispatch and accounts switched last, once the first e-invoices had been checked line by line.

The plant now runs from one live view, from purchase order to truck. Dispatch dates are based on real capacity, supervisors update progress from the floor instead of chasing it, and every gram of brass issued to a batch is accounted for. Every part can be traced to its rod lot and inspections, and invoices, challans and e-way bills come straight from the records, without retyping.