Knowledge Base

Table of Contents

Language Detection

(60+ languages supported)
When a customer sends you a message, Sparkcentral automatically detects the language used. We support over 60 languages. The language is added to the customer profile as a contact attribute. Customer service agents will always have the option to set a language manually or change it.

The purpose of detecting the language:

  • Let the autoresponder respond to customers in their language
  • Route conversations to teams with specific language skills
  • Report on the languages used by your customers
  • Store the customer language in your CRM

How does language detection work?

Once we receive a message from the customer, it will be processed by our AI language engine. The AI will need enough content to determine the language. If the language isn’t recognized based on the first customer message, we will try again on the second message. We will keep trying until it has succeeded, or the customer service agent has manually set the language.

Example 1: Success on the first message

MessageLanguage detection
“Hello, I have a question about my invoice. Can you please help me out?”97.5% English, language set to English

Example 2: Success on the third message

MessageLanguage detection
1 – “Hi”Will not be processed, the message is too short (min. 25 characters)
2 – “Hello, I have a question about an invoice”87.8% English below required certainty threshold of 90%
3 – “Can you please explain why the amount is higher than last month? It seems to me that some additional charges have been applied.”99.3% English, language set to English

How to set up language detection?

Enable language detection in your account

Language detection is disabled by default. When enabling this feature, you will have to decide what languages you want to detect automatically. The fewer languages you select, the better it will function. We recommend selecting no more than ten languages.

Making the language visible

Once the language detection is enabled, it will automatically store the detected language in a contact attribute called “Language”. You can choose to show the language as part of the Contact Attributes.

Go to SettingsAgent WorkspaceContact Attributes to manage your contact attributes. Use the toggle to make the contact attribute visible.

Routing based on language

If you are already familiar with using teams and skills-based routing, you know that you can route conversations based on topics, and customer segments. With language detection, you can now create customer segments using the customers’ language. This enables you to route a conversation to a different team based on the language.

Autoresponder with multiple languages

You can now add multiple languages for autoresponders. One of the languages is used as a default. The default language is used if the customer language is unknown. Simply start by adding a new autoresponder in one language. And then click on the + Add Translation tab, choose a language from the list and start filling the fields.

Language in CRM API

You can also retrieve the customers’ language from your CRM. Or the other way around, store the customer language in your CRM. We have added a variable named spark-language in the CRM API. More details on how to integrate with CRM can be found here.

Language reporting and exports

Conversations dashboard

On the Conversations dashboard, we have added a “Resolved Conversations per Language” chart. This will give you an overview of how many conversations have been taking place in what language.

Language in conversations export

We have added a Language column to the Resolved Conversations export. This field will only be populated if the language for a conversation is known.

See the full details of the resolved conversations export here.

Was this article helpful?

4 0

We're always happy to

Did you know you could help us too?

We love feedback from our users. It’s incredibly important for our business. That’s why a positive recommendation from your company could really make a difference!