Settings
Last updated
MangoApps Administrators Guide
Last updated
Admins can choose from three options to set up moderation for their domain: moderate only flagged content, moderate all posts and interactions, or moderate flagged content alongside content containing specific keywords. By default, the first option is selected, allowing real-time posting with moderation of flagged content only. In the second option, all content requires approval before becoming visible. The third option moderates flagged and keyword-triggered content.
Additionally, AI-powered moderation can enhance content monitoring with negativity tolerance and harm analysis. Admins can customize moderator permissions and decide which users or teams are subject to moderation, with notifications keeping users and admins informed of moderation outcomes.
Admins have a choice of 3 options to set up Moderation for their domain. These include:
Moderate only flagged content
Moderate all posts, status updates, questions, comments, and replies
Moderate flagged content in addition to content that contains the following keywords
In this option, content is posted and visible in the domain in real time. Any user can use the Flag action to send the content to the Moderation Queue for review. Moderators can view any content that is present in the Moderation Queue and edit it. Moderators can publish or remove the flagged content.
In this option, all posts, updates, questions, comments, and replies enter the Moderation Queue for approval before being visible in the domain. Moderators can decide to publish or remove all content in the Moderation Queue.
When published:
User posted content is visible in the domain and notifications are pushed out.
The item in the Queue moves to the Published status view.
When removed:
User posted content does not become visible in the domain.
The item in the Queue moves to the Removed status view.
In this option, once the content is approved, the user can not edit the content, but they can still delete it.
In this option, all comments, replies, updates, questions and posts when created will be checked if they match one of the keywords.
If a match is found, the content will not be available & visible in the domain in real time.
If a match is not found, then the content will be published and visible in the domain in real time.
If the content ends up in the moderation queue, the remaining processing on approval or decline is the same as option 2.
In this option, once the content is approved, the user can’t edit the content but they can still delete it.
Keyword-based Moderation will not automatically translate to other languages. Non-English languages will require separate keywords in that specific language.
AI-powered Moderation can be enabled within Option 3. AI Moderation settings will be in addition to the defined keywords.
AI-enabled Moderation currently only applies to English text.
The following settings can be configured when AI Moderation is enabled:
Observation Mode allows you to understand AI detection behavior without disrupting users’ content publishing flow. AI-detected content that would normally get flagged for Moderation will be automatically published, but still appear in the Queue for observation and testing of AI Moderation. AI-detected content in Observation Mode can be visible in the Published status view of the Queue.
Live Mode will enforce AI settings in your domain in real-time. AI-detected content will be flagged for Moderation and will not be visible until published in the Queue
This option allows you to set your desired tolerance for negativity to moderate content that contains varying levels of negative emotions based on AI Sentiment & Emotion Analysis.
When selected, you can choose to set your negativity tolerance at one of the following levels:
Low - Moderates all negative emotions (e.g. Sadness, Fear, Disgust, and Anger)
Medium - Moderates Fear, Disgust, and Anger
High - Moderates Disgust and Anger
Very High - Moderates only Anger
The slider can only be configured if the Negativity Tolerance checkbox is selected. Otherwise, the slider will be disabled.
This option detects and flags content that contains Harassment and/or Discrimination based on AI Harm Analysis.
When Harassment is selected, MangoApps AI will detect and flag content that expresses the following:
Personal Insults
Threats
When Discrimination is selected, MangoApps AI will detect and flag content that expresses the following:
Sexism & Gender Bias
Racism & Ethnic Bias
Religious Bias
From this section, admins can choose which user role can serve as moderators.
Network & Moderation Admins (default): Network and Moderation admins are able to take actions on flagged items in the Moderation Queue in the Admin Portal.
Team Admins: Team admins will have access to a team-level Moderation Queue within their Admin Tools in the User Portal to publish and remove content flagged within their team.
In this section, choose which users/teams should be moderated.
All users (default): Content posted by all users except network and moderation admins will be moderated.
Only moderate content from the following users: Select specific users whose content will get moderated. Only content published by these users will undergo moderation.
Do not moderate content from the following users/teams: Select specific users and/or teams whose content will not get moderated. Content published by these users/teams will not be moderated.
Moderation items will only be visible in the Queue for up to 60 days. After 60 days, items in the Pending and Removed status views will be deleted along with their original content (deleted content is available in the trash can until the trash can limit expires).
The Moderation module has notifications for both domain/moderation admins and the person who posted the content, keeping everyone in the loop.
Domain and moderation admins are notified at the top of every hour if new content has been submitted for moderation.
The user who posted the content is notified when their content is approved or declined.
When the user posted content is pending (or rejected) in the moderation queue, the author of the content can see the items in their profile under Pending Moderation.
This tab will be empty if no content of the user is in the moderation queue (in pending or rejected state).
A user can not see this tab for other users. They can only see it on their own profile.
Q: How do moderation notifications work for admins?
A: Admins can receive notifications when an item is set to be moderated. These notifications can be sent within the domain and via email. However, unlike other notifications, moderation alerts are not sent in real time. Instead, they are scheduled to be sent on the hour. For example, if content is added to the moderation queue at 12:05 p.m., admins will receive the notification at 1:00 p.m.
