e-Catalogue: when 8,000 dealers can search for a part the way they search for a flight.

Mahindra's dealers were searching for parts by flipping through printed catalogues — sometimes hundreds of pages thick. Wrong-part orders were a real cost, and parts updates took a printing cycle to reach the field. The brief was simple: make searching for a part feel like searching for a hotel.

A Next.js app with an Elasticsearch index of every part across every model line. Search supports part number, plain English, and image cross-reference. Filters cover model, year, assembly, and vendor.

On the operations side, we built a catalogue publishing tool that lets the parts team push updates with a single click — and roll them back just as fast if something looks wrong.

Median lookup time dropped from about 90 seconds to under 15. Wrong-part orders fell almost in half within the first quarter. The printing budget dropped — but the more interesting number was dealer adoption: 2,400 dealers were actively searching by the end of week one, without any forced rollout. Once one dealer in a region found something fast, the others followed.