Case Study: Lexcelera achieves autonomous, higher-quality machine translation and faster turnaround with SYSTRAN

Lexcelera is a language service provider who recently join t'Works. They were looking for a MT solution to improve the MT output and to train their own translation models in order to master the terminologies of their clients

Lexcelera, an ISO 9001–certified language services provider with a network of experts across 60 countries, faced the dual challenge of mastering industry-specific terminologies (construction, energy, medical, etc.) while adapting to digitalization and AI-driven workflows that shift linguists’ roles toward data preparation and post‑editing. To meet demand for fast, consistent translations for large corporate clients, Lexcelera turned to SYSTRAN, deploying SYSTRAN Pure Neural Server alongside SYSTRAN Model Studio and the SYSTRAN x MemoQ connector for pre‑translation in their CAT tool.

Using SYSTRAN technologies they’ve relied on for over a decade, Lexcelera now trains custom neural models autonomously, integrates MT directly into MemoQ, and improves raw output quality—gaining 10–15 BLEU points per training and regularly winning internal benchmarks. SYSTRAN’s solution cut turnaround time and costs, made post‑editing more efficient, and gave Lexcelera the domain accuracy and scalability needed to serve large multinational customers.

Lexcelera

Laurence Roguet

Managing Director