Swiggy’s Vidura

This platform contains all the data. ML models use the data platform to generate time predictions. The optimisation algorithm then analyses all predictions and matches drivers with orders. The app shows this to the customer in real time.

Address Accuracy

In e-commerce and food delivery, delivery in time is the key differentiator. An address is provided by the customer. Companies use data science and ML models to break it down. It is related to the addresses close by. Then a geocode is developed. It is a point on the map. The delivery boy arrives at the location. An app tells what the actual location is against what is geocoded. This feedback to the ML model increases the efficiency levels more and more over a period of time.

ML is used to validate customers’ addresses. Address quality scores are computed. City pin code mismatches are corrected. Suggestions are provided to the users to rectify the address.

Locus

It provides a proprietary geocoding solution based on natural language processing. It reads addresses, categorises them in zones, sorts them for delivery. It offers 95 per cent accuracy in metros.

Warehouse

E-commerce warehouses process lacs of shipments daily. They use AGV or automatic guided vehicles (cobots). They work alongside the employees. It helps to meet the scaled up demand. The efficiency is improved. Algorithms are used to figure out the optimum placement of inventory in and across warehouses to ensure the shortest delivery distances.

Box Sizing Algorithms

Amazon uses box sizing algorithms at fulfilment centres to calculate the right size of box for the order.

Picking Optimisation Software

It determines the most efficient route an employee should take to pick products stocked in shelves to ready them for dispatch.

Blackbuck and IoT

Blackbuck, a logistics start up uses IoT sensors for live vehicle tracking, to suggest the best route and monitor driver behaviour. A dashcam captures driver’s images. A careless or sleepy driver gets an alert.