Moderate social activity
Classify items and apply only the moderation actions supported by the destination.
Social moderation can label incoming activity and, where supported, flag or hide it. Classification is a prediction and can be wrong. Platform support determines what an action can actually do.
Configure a rule
-
Open Inbox → Moderation or its settings in Social.
-
Review the available categories, confidence threshold and daily usage budget.
-
Create a rule for a specific classification and choose the action offered.
-
Begin with Flag / Review when evaluating accuracy.
-
Save and inspect classified examples and the action history before broadening behavior.
Understand the effect
Flagging an item for your team differs from hiding a network comment. Some platforms support hiding only for particular comment types; an unsupported action may fall back to a less intrusive result. Read the actual outcome shown on each item. Do not treat a high confidence score as proof. Review false positives, especially when a complaint or legitimate question could be hidden.
Undo and monitor
Open the item's detail and use Restore / Undo where available. Confirm the result on the relevant surface. Inspect the activity record for both automated and manual actions. If the budget is exhausted or the connection lacks permission, new items may need manual review. This guide concerns Social moderation; the separate Inbox manual remains at docs.inbox.diglab.ai.