There are four levels of moderation process: pre-moderation, AI moderarion, users’ moderation, and in-flight moderation.
Pre-moderation (or automated moderation) is the first step in moderation, during which the system prohibits users from using certain forbidden words or phrases in a chat. This mechanism works based on several lists, which are updated constantly. If a message is blocked for sending, users receive the offer to re-phrase the message and try again.
There are Blocked words list, Blockes URLs, and Block list of Nicknames (you can a word, for example the name of the brand, users will be able to text in a chat but won’t be able to put in a nickname).
Please remember that except for lists available for administrators, there are default lists. They are pre-loaded and not open for editing. There is a set of 100% inappropriate words. These lists we prepare before the launch, therefore your moderators can work with lists that are available for editing on the admin panel. They can add new words and lists constantly, as they monitor the chat feed.
Exceptions are valid words that can be spelled partly as inappropriate and get blocked. So, if you remember such words, add them to this list. Replacements are the characters that users can use to type inappropriate words (@ instead of a, 3,14 instead of P, etc.)
Pre-moderation: how do blocking lists work
We have prepared pre-moderation lists for following languages.