About the product
Mayster AI is an intelligent parts lookup assistant built for heavy equipment distributors, developed in partnership with Glimat, a parts distributor with thirty years in the market.
The challenge
Finding the right spare part for a machine like a Komatsu excavator means digging through massive OEM PDF catalogues, a process that takes a salesperson around ten minutes per enquiry. The knowledge lives in a handful of experienced staff, and every lookup interrupts a sale. For a distributor handling hundreds of enquiries a day, that time adds up to a real cost.
What we built
Mayster replaces the catalogue hunt with a three-second conversation in Polish. A salesperson types a colloquial query, for example a rubber turbo hose for a PC210, and instantly gets the exact catalogue number and the technical diagram:
- Natural language lookup that understands trade jargon, model shorthand and Polish grammar.
- Exact catalogue numbers with the matching OEM diagrams, not just probable answers.
- Diagnostic suggestions that rank likely causes and the parts to check first.
- Coverage of more than 30 million Komatsu parts, with a zero-hallucination retrieval design.
How we worked
An AI-native team of five engineers built the product, combining Node.js, Rust, Vue.js and Python. Accuracy is the core requirement: answers are grounded in the catalogue data itself, so the assistant returns a verified part number or nothing.
The outcome
Parts lookup went from ten minutes to seconds, and the platform operates on a hybrid SaaS model with more than ninety percent margins. The engagement is ongoing as coverage expands to further manufacturers.