This feature is available only on the Pro plan.
Setting up language attribution for conversations
Lokalise Messages automatically detects a language of an incoming conversation. It uses an AI to detect a language based on its content instead of making assumptions based on the customer's geolocation or country specified in the metadata.
Once the installation process is completed, a new custom field will be added:
By default the following languages are populated:
Language | Tag |
English | lokalise_english |
Arabic | lokalise_arabic |
Portuguese | lokalise_portuguese |
Spanish | lokalise_spanish |
French | lokalise_french |
German | lokalise_german |
Italian | lokalise_italian |
Chinese (Traditional)" | lokalise_chinese_traditional |
Chinese (Simplified) | lokalise_chinese_simplified |
Russian | lokalise_russian |
Dutch | lokalise_dutch |
Estonian | lokalise_estonian |
Tagalog (Filipino) | lokalise_tagalog_filipino |
Hebrew | lokalise_hebrew |
Malay | lokalise_malay |
Other | lokalise_other (languages which are not added to the list would receive |
You can add additional languages that you want to track. Language names must be specified as shown in the following table:
Afrikaans | Georgian | Lithuanian | Sindhi |
Albanian | German | Luxembourgish | Sinhala (Sinhalese) |
Amharic | Greek | Macedonian | Slovak |
Arabic | Gujarati | Malagasy | Slovenian |
Armenian | Haitian Creole | Malay | Somali |
Azerbaijani | Hausa | Malayalam | Spanish |
Basque | Hawaiian | Maltese | Sundanese |
Belarusian | Hebrew | Maori | Swahili |
Bengali | Hindi | Marathi | Swedish |
Bosnian | Hmong | Mongolian | Tagalog (Filipino) |
Bulgarian | Hungarian | Myanmar (Burmese) | Tajik |
Catalan | Icelandic | Nepali | Tamil |
Cebuano | Igbo | Norwegian | Tatar |
Chinese (Simplified) | Indonesian | Nyanja (Chichewa) | Telugu |
Chinese (Traditional) | Irish | Odia (Oriya) | Thai |
Corsican | Italian | Pashto | Turkish |
Croatian | Japanese | Persian | Turkmen |
Czech | Javanese | Polish | Ukrainian |
Danish | Kannada | Portuguese | Urdu |
Dutch | Kazakh | Punjabi | Uyghur |
English | Khmer | Romanian | Uzbek |
Esperanto | Kinyarwanda | Russian | Vietnamese |
Estonian | Korean | Samoan | Welsh |
Finnish | Kurdish | Scots Gaelic | Xhosa |
French | Kyrgyz | Serbian | Yiddish |
Frisian | Lao | Sesotho | Yoruba |
Galician | Latvian | Shona | Zulu |
Language-based routing
You can add a trigger to automatically assign tickets to a team that speaks a particular language.
Go to Admin > Objects and Rules > Triggers and create a new Trigger.
Set the necessary conditions, including "Language (Lokalise)" value.
In the actions select a group that should be assigned to the new ticket.
Language dimension for conversation analytics
A custom attribute adds a custom field and tags like "lokalise_english" that could be used with Zendesk Explore to build language based reports and dashboards.
Changing conversation language
Neural networks that we are using to perform machine translation sometimes fail to understand the language of the message (for example, if it contains typos or content in several languages).
Agents can change the language of a conversation using the sidebar widget. See Change the conversation language via the sidebar widget section in the tutorial for agents.