Multilingual team: working without a shared language

A subsidiary opened in Germany, a team acquired in Spain, a developer hired in Portugal, a production partner in Asia: a 50-person company can end up with colleagues who share no language, or who share a very uneven English. Meetings get longer, emails are reread three times, and some people stay silent because they are not sure they will say it right.

Translation tools have improved a lot, and they help. But a multilingual team works first thanks to a few simple rules on the working language, writing and meetings. This article goes through them. It complements our other articles on translation and new markets, which focus on customers: this one looks inside the company.

Reading, listening, speaking: very different levels

In a multilingual team, people often assume everyone “speaks English”. Measurements show a more nuanced reality, and above all an uneven one depending on the skill. The EF English proficiency ranking illustrates it for France:

This ranking is based on people who chose to take the test: it shows a trend, not an exact portrait of your team. But the gap between reading and speaking shows up in many teams. Practical consequence: a written decision, read at one’s own pace, gets through better than a decision made orally at the end of a meeting, when those who were following with effort have already dropped off.

What meeting tools do in 2026

The two big office suites now translate meetings, each with its own conditions.

• Microsoft Teams shows live translated captions. According to Microsoft’s help, the feature is part of Teams Premium and Microsoft Copilot, add-on licences: if the organiser has one, all participants can use it.
• Google Meet goes further with voice translation, available to businesses since 27 January 2026 according to the Google Workspace announcement: everyone speaks their language and the others hear a translation that keeps the speaker’s tone. At launch, one language pair per meeting, between English and Spanish, French, German, Portuguese or Italian.

These tools help a great deal in everyday meetings: a colleague can follow in their own language instead of guessing. They do not solve everything. A mistranslated trade term goes unnoticed, a joke falls flat, and a three-language meeting still exceeds some tools’ limits. And what is said in a meeting goes to a translation service: the questions of translation and GDPR apply just as they do for a document.

Choosing a working language, and what stays in each person’s language

A team needs a reference language for what everyone must understand the same way. It does not need to do everything in that language.

For exchanges with foreign customers, the workflow described in translating customer support emails without hiring also works between colleagues: everyone writes in their language, translation bridges the gap, and what commits the company goes through someone who reads the language.

Writing to be understood, and translated

A simple text is easier to understand in a foreign language and translates better by machine. A few rules are enough:

1. One idea per sentence, with no innuendo or irony.
2. No idioms, and no internal acronyms left unexplained.
3. The same words for the same things, kept in a shared glossary, like the brand glossary used to translate the website.
4. The decision written first, the context after.
5. Explicit dates, figures and names: “Friday the 24th” rather than “next week”.

These rules also serve the tools: a clear text gives a better machine translation, whatever the engine, and the method in DeepL, Google or an LLM helps choose the one that best understands your trade.

Meetings that leave nobody out

Stage four is the most useful. A summary generated by AI from the transcript saves time, provided someone who attended the meeting reviews it before it is sent: a summarising tool can turn a hesitation into a decision. To check whether translations hold up without rereading everything, the method of our article on translation quality applies to internal documents too.

Documentation in several languages

Procedures, onboarding for newcomers, tool rules: documentation kept in one source language, then translated and synchronised, prevents each country from having its own drifting version. A translation memory only retranslates the sentences that changed, and the principles of translating software or an app apply to the internal tools the team uses every day. Training videos are subtitled following AI subtitles and dubbing for export.

Legal terms and documents that commit the company, employment contracts, regulations, agreements, follow the precautions of translating a contract with AI: where it stops.

What tools do not replace

Bilingual team members are bridges: they spot a misunderstanding the translation let through, and explain what belongs to culture rather than language. Language courses keep their value, because a direct conversation, even an imperfect one, builds a relationship a tool does not. And for a negotiation, labour relations or a crisis, a professional interpreter remains the right choice: the stakes justify someone whose job it is.

When the multilingual team also serves customers

A team that speaks several languages is an asset for opening markets. Its members know what customers in their country look for, which helps redo keyword research abroad rather than translate it, localize rather than just translate prices and proof, and review product pages translated for export. It helps choose which European market to open first, understand what changes beyond language in an English-speaking market or in China and Southeast Asia, and decide which blog articles to translate first. The choice between machine and reviewed translation is detailed in translating your website with AI: automatic or reviewed?, and the budget in website translation cost.

Where to start

Write down the team’s working language and the list of what must be done in it, then turn on translated captions in your meeting tool and test them in an everyday meeting. Start a shared glossary of the twenty internal terms that raise the most questions, and introduce written minutes after every decision meeting. To equip the translation of your documents and messages with your data rules, tell us about your team and your languages.

Frequently asked questions

Sources

1. Microsoft, live captions in Teams (new tab)support.microsoft.com
2. Google Workspace Updates, speech translation in Meet (new tab)workspaceupdates.googleblog.com