The Steering Wheel of AI Translation

T

ranslated’s VP of AI Solutions, John Tinsley, explains why larger content windows have finally given the industrial world real control over AI translation quality, why proprietary technology is the only way to fully leverage that control, and why multi-model alternatives are selling sophistication they can’t deliver.

What’s the biggest challenge in AI translation today?

For years, if the output was bad, you had three moves: more data, a glossary, and some post-processing rules. No real steering wheel.

That’s changed. Context windows are now huge. We can translate whole documents instead of forcing content through segment-by-segment workflows, bring in external context the way a human would, give the model style guides and instructions, and have the model actually use them. But an LLM does not make a translation product. Productization is where the real work is.

What’s Translated’s answer to that?

It starts with proprietary technology. If you’re building on the same model as your competitors, you don’t own the evolution cycle. You can’t move on quality, speed, or direction yourself. That’s a hard ceiling.

TranslationOS is that complete production system: Lara for pre-translation, T-Rank for translator assignment, Matecat for production, real-time analytics, a feedback loop for continuous improvement, and full transparency throughout. It integrates with major TMS platforms for multi-vendor environments.

How is the approach winning?

Translation buyers are increasingly looking for transparency, control, and systems that improve over time. That’s why organizations move away from monolithic legacy providers toward more adaptive approaches.

The “alternative” sold right now is multi-model, best-of-breed. Mix and match LLMs, AI that judges AI, then judges the judgment. It sounds sophisticated. It isn’t. You get watered-down versions of everything, inconsistent results, and no feedback loop. The winners build translation products, not wrappers.

About John

John Tinsley is the VP for Al Solutions at Translated. He’s an Irish entrepreneur, computer scientist, and translation expert. He founded Iconic Translation Machines, an award-winning language technology software business which pioneered the commercial deployment of Neural Machine Translation technology. John grew the business for almost a decade before selling it to RWS in 2020 in one of the largest technology deals in the language industry. He holds a PhD in Machine Translation and a degree in Applied Computational Linguistics, and is a regular public speaker on topics related to language, translation, and business.