Set Your Primary Language for Translated Conversations

By default, once your tenant has translation enabled, conversations are shown to agents translated into English. If English isn’t your first language, that’s an extra layer of friction between you and the customer. The View and reply in your primary language toggle removes that layer for you personally: it renders every conversation in your own language instead, and lets you compose in that language and have the reply translated for the customer on send.

This is a per-agent setting. Turning it on doesn’t change what anyone else sees — each teammate can choose their own language independently.

Prerequisite: tenant translation must be enabled

This toggle only exists if your tenant’s auto-translation feature is already turned on. If it isn’t, Settings → Language doesn’t show the toggle at all — there’s nothing for it to do. Auto-translation itself — what it is and how the tenant-wide English rendering works — is covered in What is auto-translation?. This article is about the personal layer on top of it.

Turn it on

1. Go to Settings → Language.
2. Set your Primary Language, if you haven’t already — this is the language conversations will be translated into for you.
3. Enable View and reply in your primary language. The toggle only appears when your tenant has translation switched on.
4. Click Save changes.

Once it’s saved, it applies to every conversation you open — no need to reassign anything.

What changes when it’s on

Reading a conversation

Instead of the tenant-wide English rendition, each message is translated on demand into your primary language the first time you view it. While a message is being translated you’ll briefly see a Translating… indicator next to it; it clears as soon as the rendition is ready.

Messages you already understand natively aren’t translated at all — if a customer writes in your primary language, or in any of your Secondary Languages, you just see the original.

Show original

Translated messages carry a Show original toggle. Use it to flip between the translated rendition and the customer’s original wording — useful when you want to double-check phrasing, tone, or something that reads oddly in translation (product names, slang, or a name that doesn’t translate cleanly). Toggling back and forth doesn’t re-request the translation; it’s the same rendition each time.

The toggle is conversation-wide rather than per-message, and there is one for each side of the thread: revealing the original on a customer message reveals it on every customer message in that conversation, and the agent-side toggle works the same way for your own replies.

Replying

You compose the reply in your primary language (your compose language). Translating it into the customer’s language is a per-conversation choice you make in the composer: under the reply box is a line reading Replying in — translate to with a checkbox. Tick it and Atender translates the reply as it sends; leave it unticked and the reply goes out exactly as you wrote it.

The checkbox only appears when the conversation has a detected language you don’t already read, and it starts unticked until you flip the Default: control beside it to Translate. From then on it is pre-ticked for conversations in languages you don’t read, and your per-conversation choice is remembered.

How this differs from tenant-wide auto-translation

• Where it’s configured — Settings → Translation (needs the AI settings permission) — Settings → Language (per agent)
• What it translates to — English, for everyone — Your own primary language, only for you
• Who it affects — All agents in the tenant — Only you
• Requires — — — Tenant-wide auto-translation to already be enabled

If you turn the toggle off, you go back to seeing the tenant-wide English rendition (assuming auto-translation is still enabled at the tenant level) rather than seeing raw, untranslated originals for every language.

Things to keep in mind

• If your tenant disables auto-translation entirely, the toggle disappears from Settings → Language and your setting stops having any effect — conversations show their original, untranslated content, same as everyone else’s.
• Translation happens per viewer, not per conversation. Two agents with different primary languages looking at the same conversation each get their own rendition; neither affects what the other sees.
• Switching your primary language re-triggers translation into the new language the next time you open a conversation — existing renditions in your old language aren’t reused.