All Collections
Artificial intelligence
How Howspace AI works when there are several languages combined in one chat?
How Howspace AI works when there are several languages combined in one chat?

This article explains how Howspace AI behaves with several languages in one chat and how that affects the results.

Updated over a week ago

Results may vary. We recommend having one language per chat.
If you have two different languages, we recommend using two different chat widgets or pages.

Behavior depends on the feature

Word cloud

Word cloud shows multiple languages and counts the words. You'll end up seeing words in multiple different languages. You might see words like "kieli" or "language" in the same cloud.


Summary picks the most used language: it will count which language was most used (number of paragraphs) in the chat and summarize only those discussions.

Theme clustering

It tries to make clusters and you will see different languages inside clusters. You will see multiple languages within clusters and they should give pretty good results.

Again we do not have a translation service so it's possible that there are two separate clusters: one about "Finnish language" and another about "suomen kieli".

Sentiment analysis

Currently it picks and shows only one language in the results.

Workspace analytics

Workspace analytics will treat topics of different languages separately and does not translate them.

A notion from our AI developer

Officially we do not support having multiple languages in one chat, but that doesn't mean it would not work in your specific case.

Did this answer your question?