Three languages, one company: English, Chinese and Thai in the same workspace
A small company in Bangkok or Singapore or Kuala Lumpur often has three first languages in it before it has twenty people. The founders might work in English and Chinese. The operations lead is Thai. The two engineers hired last quarter each answer in a different one of the three.
Nobody planned this. It is what hiring in the region looks like when you hire for the work rather than for the language.
What multilingual actually costs a small team
The cost is rarely the dramatic thing people expect. Teams like this usually settle into a shared working language within a few weeks, and it is usually English, and the conversation works.
What wears people down is the surface of the tools they work in. A person whose English is fine for a message can still be slower at a settings page written in it, and much slower under time pressure. Menu labels, confirmation dialogs, the difference between two buttons that both sound like save: that is the layer where a second language is a tax, and it is the layer people never complain about because complaining about it sounds like admitting something.
The second cost is quieter still. Somebody reads a screen, is fairly sure of what it says, and does not act. A week later it turns out they were waiting for someone to confirm something that the screen had already told them.
What follows the browser
The interface in Opitor is available in English, Chinese and Thai, and it follows your browser language.
Two things about that are worth spelling out, because they are what make it useful rather than a checkbox on a feature list.
It is per person, not per company. The company does not pick one interface language and impose it. Your colleague sitting next to you can have the same screen in a different language at the same time, looking at the same todo.
And it is not a setting anybody has to be walked through. A person who opens the product in a browser that is already set to Thai sees Thai. For the person you are onboarding, this means their first ten minutes are spent on what the product does rather than on finding where the language lives.
What it does not do
Here is the boundary, stated plainly, because it is the thing people assume in the wrong direction.
Messages are not rewritten. What a person writes is what everybody else reads, word for word, in the language they wrote it in. If your operations lead writes a message in Thai, it arrives in Thai. If you answer in English, it arrives in English.
That is a deliberate limit and not an oversight. An internal message is often a commitment, and a commitment that has been rephrased by a machine between one person and another is a commitment two people now remember differently. The version everyone can point at should be the version somebody actually wrote.
Your chief of staff sits inside that same interface, so the labels, the buttons and the dates around it follow the same language setting. That is an interface claim. It is not a claim about the words inside anybody’s messages.
So what the three languages solve is the surface. The conversation itself is still something the team has to decide how to run, and that is the rest of this guide.
Habits that carry across three first languages
These are cheap, they are boring, and they are the ones that hold up in the third busy week.
Agree on one language for anything shared. Todo titles, deadline names, the short sentence on a quarterly goal. Pick the one most of the team is most comfortable reading under pressure. It is usually English, but the point is that it is agreed once, out loud, rather than negotiated silently in every thread.
Keep private conversation free. The agreement covers shared objects. A direct message between two people who both prefer Thai should be in Thai. A rule that makes people work in a second language for a conversation that concerns only them buys nothing and costs goodwill.
Write dates as numbers. The twelfth of March is the one thing on this list that goes wrong silently. Month names are exactly the words a tired reader half-recognises, and the numeric forms differ by convention across the same team. Put the date on the todo, where it is a date rather than a sentence, and let each person’s own screen render it.
Keep goal sentences short. A quarterly goal that runs to two clauses and a subordinate one is hard to read in your own language and genuinely ambiguous in your second. Short subject, short verb, one number if there is a number.
Put the decision in the first line. Small-company messages often bury the ask under two paragraphs of context. That is a mild annoyance for a first-language reader and a real cost for everybody else, because they have to read the whole thing to find out whether it was addressed to them.
Name each thing once. A project called one name in English and a different one in Chinese is two projects as far as anybody searching is concerned. Pick the name, write it down, and use it in all three languages.
Where the seams still show
It would be dishonest to present this as solved. Two things remain.
Nuance in a second language is hard, and disagreement is mostly nuance. When two people are talking past each other in a written thread, the fastest fix is still a short call in whatever language the two of them share, followed by a written summary of what was decided. The summary belongs in the shared language, in writing, on the item it concerns.
And reading speed is not the same as writing speed. A person can read English comfortably and still take much longer to compose a careful reply in it. If a thread is running fast and somebody has gone quiet, that quiet is often composition, not disagreement. A chief of staff that drafts helps here, because the draft removes the blank page, and the person keeps the decision about what is actually sent.
What to tell a new person on day one
Three sentences, said out loud rather than written in a handbook nobody opens.
Your screen will be in your language, and you do not have to do anything to make that happen. Shared things, which means todo titles, dates and goals, are written in our agreed language. Everything else, including your own messages, is yours to write however you write.
That is the whole policy, and a small company does not need more of one.
Where Opitor fits
Opitor’s public materials describe an AI-native operating system for small companies, where every person gets their own AI chief of staff. The interface is available in English, Chinese and Thai, and follows your browser language, so each person can work in their own while the company shares todos, deadlines and quarterly goals. Messages, direct and group, carry what people write. Product materials describe each chief of staff as following up todos and deadlines and drafting AI replies, which are signed in purple and sent only when you choose; production model context remains unverified. Materials also describe a self-created reminder as eligible for scheduled sending at its due time. More on the described model is on what is an AI chief of staff and in the questions and answers.
Updated 2026-09-25