Content moderation

Overview

Content moderation is an AI-powered tool that helps companies make sure that their intranet sites are clear of offensive feed posts and comments.

Note that as of now, history and analytics of content moderation are not available, but they will be added in an upcoming release.

Content moderation must be enabled for your org by Simpplr. Reach out to your account representative to have this feature enabled.

Content moderation engine

Simpplr's content moderation engine is an AI-powered algorithm used on every feed post, comment, and reply to ensure they aren't obscene. It's built to detect these types of objectionable content:

How content moderation works

When a feed post (home or site), comment, or reply is submitted, it goes through the content moderation engine. If the engine doesn't flag anything, it's posted as normal. If it contains content that is flagged, the poster will be notified, given the reason for the flagging, and given the option to edit the post, or continuing posting it. If posted un-edited, content is sent to the queue for moderation. The content moderator can decide whether to keep the post, or hide it. Users can also report feed posts/comments as offensive, and give a reason for the report. These are also sent to the content moderator, who will decide to keep or hide the posts. 

Enable content moderation

By default, the engine is turned off. To enable content moderation, go to Application settings > Applications > Setup > Privileges > Content moderation. Select Use content moderation. Admins have the ability to manage content moderation but you can add additional users with content moderation permission.

Note that as of now, content moderation will only scan about the first 500 words (or one page of text) in any given post. 

The Access Control Groups (ACG) role for content moderation allows access to the content moderation queue. If someone is added as a feature manager for this system ACG, they will be able to access Manage features > Content moderation. 

Content moderation queue

Once content moderation is enabled for your organization, content moderations can view their queues by going to your User menu > Content moderation and accessing the Queue tab. App managers go to Manage > Content moderation. The latest reported content will appear in the queue first.

Click Remove comment to remove a comment. If removed, a comment will remain visible to moderators in the analytics section of Content Moderation.

Note that content moderators will be able to see all moderated activity across public, private and unlisted sites within the queue. However, if not already a member, they will not see any other information outside the literal content written. They cannot go to that site or make any changes.

 Notifications

All content moderators will receive in-app, actionable notifications for any reported content. The latest reported content will appear in the queue first. These notifications cannot be disabled.

Report inappropriate content

Users can also report inappropriate content. A modal will open, prompting a reason for the report, and the content will be added to the moderators' queue. 

Languages

Currently, content moderation supports any content written in the following languages: 

• US English

• Spanish

• Danish

• German

• French

• Portuguese

• Italian

AI content moderation

AI content moderation automatically monitors posts and comments across your intranet to identify and mitigate inappropriate or unsafe content. By integrating advanced machine learning, the system detects concerning material in real-time, hides it from the community, and routes it to a dedicated moderation team for final review. This ensures a proactive safety net without relying solely on manual reporting.

Scope of monitoring

The system is integrated into the most active social areas of the platform to maintain a consistent standard of conduct. This includes home and site feeds (posts, comments, and replies), content comments on pages, blogs, events, and albums, as well as recognition posts and their associated threads.

The moderation workflow

When a user submits a post or comment, the system initiates a rapid, two-phase screening process that typically completes in under half a second.

• Initial screening: The content is first filtered against a global list of blocked words and phrases.

• AI analysis: The system then utilizes AI to interpret meaning and identify nuanced issues such as hate speech, explicit content, violence, or self-harm.

• Queue management: If content is flagged as unsafe, it is instantly hidden from other users and moved into a review queue for human intervention.

It is important to note that the author is never prevented from clicking "post". If their content is flagged, it is simply sequestered from public view until a moderator validates the submission.

Moderator roles and actions

Only users assigned the content moderator role can access the review queue and historical reports. To protect privacy, moderators can only view flagged content that they already have permission to access; private or unlisted content remains hidden unless the moderator is a member of that specific group or site.

Once an item is in the queue, a moderator must decide on one of two outcomes:

1. Approve: The content becomes visible again to its intended audience.

2. Remove: The content is permanently deleted, with the option to remove all related replies in the same thread to clean up the conversation entirely.

Administration and insights

Administrators can enable this feature within the AI control center by toggling the "AI content moderation for content" setting. While active, the system uses Gemini by default to process checks securely within your organization's boundaries. Administrators can also track performance through detailed reports, which break down flagged items by source—such as home feeds or profiles—and compare AI-detected flags against user-reported items.

Best practices for implementation

For effective oversight, organizations should start with AI enabled to capture a broader range of problematic content with fewer false alarms than legacy systems. Maintaining a consistent, small moderation team ensures that decisions remain fair and aligned with company culture. Finally, checking moderation reports regularly allows you to refine your blocked words list and adjust internal posting guidelines based on real-world activity.