Inside 15 international marketing teams: 8 trends in AI and translation

Posted: 16 Nov 2026

Over the summer and autumn of 2026, I spoke with marketing leaders at 15 international companies, including LG Electronics, Selkirk Sport, Siderise, Sonata Software and IFF. I asked how they create, adapt and translate content for new markets, how they use AI and where translation still slows them down.

Their setups vary a great deal, from a content team of two working across 16 languages to global brands with colleagues in every region. Even so, the same patterns came up in conversation after conversation.

Here are the eight biggest trends we heard. 

1. The “global master + local adaptation” model is now the norm

 

Most of the multi-market brands we spoke with now build one complete global version first, then let each market adapt it. Local teams spend their time removing what doesn’t apply rather than starting from a blank page, which saves translation spend in every market.

2. Local knowledge is coming in earlier

 

Every team interviewed relies on local colleagues of some description to judge whether content will work in their market. The teams with the fewest problems ask for their input when they write the brief, not at sign-off, when changing course costs time and money.

3. Localisation is being thought of as more than just language

 

An accurate translation can still miss its market. The teams we spoke with shared examples of words that mean different things in different countries, audiences that expect different levels of detail, and campaigns that landed in the wrong season.

4. AI is everywhere, but it’s rarely left unsupervised

 

Every team uses AI, whether for first drafts, edits, research or turning one piece of content into several. How far they go varies widely, but only one brand we spoke to lets AI-produced content reach customers without a real person checking it first.

5. A content-risk tiering model is emerging

 

More experienced teams decide how much care each piece of content needs before choosing AI, a linguist, or a specialist agency. An internal presentation and a legal contract don’t need the same process, and treating them the same way either wastes money or takes unnecessary risks.

6. The bottleneck is now review capacity, not translation itself

 

A first draft is quick to produce, but finding a colleague with the language, product knowledge, and time to check it is much harder, and we heard this challenge most often.

7. Terminology management and glossaries are becoming foundational, not optional 

 

The right word in a market is often not the literal one, even between UK and US English. Teams are building glossaries so that every translator, person or tool uses the approved terms, and so that updates cost less.

8. Highly technical, regulated, or industry-specific content resists generic translation

 

Generic translation struggles with highly technical, regulated or industry-specific content. Teams shared examples where accuracy depends on product knowledge and rules, such as technical data sheets, country-specific building regulations, and healthcare content that must meet different rules in the UK, the EU, and the US. Some technical terms have no direct equivalent in another language, so you have to adapt the wording rather than translate it word for word.

Three quick wins you can do this month to improve your translation workflow:

  1. Give each of your main content types a risk level before you decide who, or what, translates it.
  2. Bring one reviewer in each priority market into your next brief, before the creative work starts.
  3. Write down 50-100 of your most important terms, with the approved wording for each market.

Want a deeper dive into these trends?

Our full report goes into each of the eight trends in more detail, with real examples from the teams we spoke with and an idea to take to your own team for each one.

Email “TRENDS” to info@comtectranslations.co.uk and we’ll send you the full report.