/11 · GLOBAL · Languages

Thúy Hiền

Language & Localisation Specialist · Chạm AI Agency

Removing the language barrier from every project with an international side.

Translating the words correctly does not mean saying the right thing. Thúy Hiền owns localisation: keeping the message and brand voice intact when content moves into another language, instead of swapping words from a dictionary.

The Chạm AI bilingual website is the working example: more than 30 English pages running alongside the Vietnamese ones, each properly written rather than machine translated, with reciprocal hreflang so search engines understand these are two versions of the same content.

The rest of the work is the terminology set: agreeing how core concepts are translated so every document, email and AI Agent reply uses one consistent name.

Responsibilities

Core work

Website localisation

Moving content into the target language while keeping voice and intent, with correct hreflang declarations.

Brand terminology

Fixing how core concepts translate so documents, website and AI Agent all use the same term. A glossary usually runs 4 columns: source term, approved translation, terms to avoid and a usage note.

International documents

Writing and editing capability profiles, quotes and correspondence for overseas partners.

Machine translation review

Auditing machine output to remove context errors, wrong terminology and misleading phrasing.

Need to talk about this area?

Contact Chạm AI → Services English home Bản tiếng Việt

How is localisation different from translation?

Quick summary: Thúy Hiền leads language and localisation at Chạm AI Agency: moving websites, documents and brand content into another language while keeping the message and voice intact. Unlike word for word translation, localisation also handles cultural context, a shared terminology set, and hreflang declarations so search engines read the language versions correctly. The Chạm AI site currently runs more than 30 English pages alongside the Vietnamese ones this way, each pair carrying 3 hreflang lines: vi, en and x-default.

Why not run a multilingual site on machine translation?

Machine translation breaks context exactly where it matters most: sales messages and service names. Bulk machine output also creates thin pages, and without hreflang search engines may treat the two versions as duplicates.

What is hreflang and why must it be reciprocal?

Hreflang tells search engines that a page has versions in other languages. The declaration must point both ways: the Vietnamese page to the English one and back. A one way declaration is ignored.

What problem does a terminology set solve?

It fixes how core concepts are named. When the website, capability profile and AI Agent use three different translations for the same service, neither customers nor search engines recognise them as one thing.