Case study
Japanese Truck Manufacturer Boosts Resolution Speed with NLP
How SquareShift built a BERT-based NLP model that matches quality issues across English and Japanese for a Japanese truck manufacturer.
A leading Japanese original equipment manufacturer of commercial trucks, serving domestic and international markets.
Its large service network and complex quality-assurance process needed faster issue resolution across both English and Japanese.
Best-matching historical issues surfaced for every new quality report. Resolves quality issues consistently across English and Japanese. Replaces manual database searches for quality engineers.
Key servicesHow SquareShift delivered it.
The challenge
Dealers reported fresh quality issues constantly, but matching a new issue to a similar historical one meant manually sifting through a large, bilingual database. Issues were logged in both English and Japanese, so search and comparison were slow even before quality engineers found the right match — delaying resolution and denting customer satisfaction.
What we delivered
SquareShift translated Japanese issues into English for uniform processing, then built a custom BERT-based NLP model that compares new issues against historical data and surfaces the top three closest matches. Engineers reuse or adapt those matches instead of starting research from scratch.
The model runs behind a custom UI and backend on Microsoft Azure, so quality engineers work from one bilingual search experience instead of two separate manual processes.
The payoff
Quality engineers now get the three best-matching historical issues automatically, in both English and Japanese, instead of manually searching a large database. Translating and matching issues through a single BERT-based model cut the time between a dealer report and a resolved issue.
Bilingual search isn't translation plus keyword match — it's building one relevance model that treats an English report and its Japanese twin as the same problem, not two.
AI & Generative AI Practice Lead, SquareShift
Where this work sits