There are several ways to detect and set a user's language. One of those ways is automatically detecting an end user's language from an email message sent to Zendesk Support. Automatic detection works only for requests from unregistered (new) users and users who do not currently have a language setting selected.
Email-based language detection is useful if you provide support via email only since your users don't use help center, never sign in, and aren't able to set a language preference. The only other way to handle language for email-only end users is to have them send email support requests to a language-specific email address such as email@example.com.
This article contains the following sections:
How it works
Email-based language detection is automatically enabled when you have more than one language in your Zendesk account.
- If a language is declared in the Accept-Language email header, that language is set for the user.
- If there isn't a language declared in the header, an email text scan attempts to detect the language. The length of the email message can affect the language analysis. There's no guaranteed length at which the language will always be detected, but longer messages produce better results.
- If the text scan can't detect the language (for example, the message was too short to determine the language), no language is detected.
- If the detected language is one of your supported languages, then the user's language preference is set.
- If a language can't be detected by analyzing the email text or header, then the user's language is set to the Zendesk account's default language.
The email language detection feature currently supports the following languages:
I have a question about automatic language detection when a new user is created automatically from an request received by E-mail.
We do not have German as a localized language in our Zendesk.
We have, English (US), French, Spanish and Ducth.
If someone writes us in German should the language be defaulted in English or another language based on the below point:
For new users, their language is automatically detected by the text contained in the body of their first email sent to your Support account. The system scans the email content to look for the closest matched language enabled in your account if the language is not supported.
With your use-case, I quickly run a test on my end to see how the system will detect the language, and the new user who writes in German was set with the English language.
This says it's enabled for French. Is it French-Canadian as well?
@... I think you can find it under Canadian French. (It's where I had to find Canadian English - it wasn't with the other 'English' languages for some reason...)
What if the end-user is created from an outgoing e-mail? We noticed that the end-user was created with "English US" as language, even if we contact the customer in another language.
Any way to go around that?
Outbound email message is not part of end-user's language detection. The only way is for your agents to edit/create the user and set the correct language on their profile.
I'm wanting to know if we can use this language-detecting feature to help in sorting out tickets using view, trigger, automations, etc.? Like if ZD detects ticket's language to be Japanese, it would assign the ticket to a specific group.
Hope to hear from you
It won't do this natively, Zendesk doesn't have language properties assigned to Groups. Thus it won't know that the group you named 'Japanese Support' is intended for Japanese language tickets. You need to create rules to assign these tickets when they come in.
I would look at creating a Trigger for:
Ticket IS Created
Requester Language is Japanese (this is based on what the system detected)
Assign to Japanese Support Group
(optionally) assign skill if you used Skill Based Routing)
(optionally) tag the ticket with a language tag or set a custom field value on the ticket 'Language' for later reporting
Put this early in your trigger list so that it assigns to the correct group as early as possible in your workflow.
Hi Dan Ross
Thanks for your quick reply and looking into my question but what I meant is that are we able to put this language-detecting feature into one of the conditions for any rules to be created, that is, for example, an Email written in Japanese be automatically routed to a specific group of Japanese language support?
Hope it becomes clearer now and appreciate your prompt revert
Yes, you can put it in Triggers (I haven't tested Automations) but in our experience it's got a pretty high miss rate for accurately detecting language.
@Gavin can you show me how you put the language detection into the trigger?
Can trigger support to detect language in email title instead of email body?
Our users are used to sending emails with long titles, and in mail body only pictures or no any content. So Zendesk always detect the emails as not the language in email title.
Can Zendesk help?
Simply select "Language" from the trigger conditions.
No, ZD does not offer this level of control.
Currently, the language detection is natively designed to read body text and not subject.
I can only think (as a workaround), a trigger with Subject Text - Contain "string of words" or specific words that can be normally found in the subject text, to route those emails/tickets to the specific Group.
This might not be useful for you if the emails subject does not have pattern. But if does, a very customisable Trigger like:
I hope this helps.
Thanks for the feedback and suggestion.
We have global business all around the world, and users will write emails with many language and many different words, there is no clear pattern of the subject.
Is there any plan from Zendesk to detect the language of email's subject?
I've tested with traditional Chinese content email (With English subject) and the user language wasn't able to detected as in Chinese. Please kindly advise.
That's our experience as well, Christy. Unfortunately this is considered normal by ZD.
We use the Organizations feature to organize customer accounts by language/region, and then we route by Organization. However it required working with our Engineering team to accomplish matching the account to the appropriate Org.
Just looking to clarify how identifying user language works. I would like to report to our execs a breakdown of languages we get tickets in. So, ideally, I need Zendesk to tell me that out of these 80 tickets, 10 are in English, 5 are in Spanish, etc, etc. However, if sounds like right now, however, all the language identification is done on the user side. So the best I could do is say "We got 50 new users who submitted tickets, 10 spoke English, 10 spoke Spanish, etc."
My understanding after reading all these help articles is to do so I have to manually create an automation for every language I want to identify. So if I think we get tickets in ten languages, I have to create 10 automations, and if we get tickets in a new language, I have to identify that we're getting those tickets and then create a new automation for that language.
Is that correct?
In Zendesk Explore, try using the Attribute for 'User Locale' in conjunction with a metric for counting the number of Tickets. That should break down the languages for you by the # tickets created attached to that requester's language.
If you have requesters who submit in multiple languages, I would suggest creating a trigger for each support language that sets a value in a custom drop down field for Ticket Language. This way you get an accurate count going forward of the ticket language, even if the requester's locale settings change.
Ah, I got it! Thank you. I'll start off by automating "Is english" and "is not english," and then move from there.
Please sign in to leave a comment.