AI Moderation: Recommended configurations

Only use these configurations if you want to use AI Moderation. Auto-Approved means NO AI moderation and should only be used where you rely entirely on user reports.

Configuration A - High-Trust Professional Community
(B2B, professional networks, vetted technical forums)

• Balanced threshold (~75%)
• Selective Review for all content
• Board Context ON for technical boards
• Appeals ON, AI-Only OFF
• Max Appeals 1

Configuration B - Mixed-Trust Community with Growth
(Growing B2C or gaming communities)

• Balanced threshold (~75%)
• Required Review for all content
• Selective Review override for Trusted Members
• Board Context ON for boards with unique norms
• Appeals ON, AI-Only OFF,
• Max Appeals 2

Configuration C - High-Risk or Regulated Content
(Healthcare, financial advice, minors, legal/compliance)

• Conservative threshold (above 80%)
• Required Review for all content
• Board Context ON for all boards with detailed descriptions
• Appeals ON, AI-Only ON
• Max Appeals 1
• Custom Reason Codes configured with specific policy references

Configuration D - Fast-Moving Event or Launch 
(Product launches, live events, time-sensitive campaigns)

• Aggressive threshold (below 70%)
• Selective Review for all content
• Board Context OFF
• Appeals ON 
• Max Appeals 1 
• Use Auto-Approved (no AI) only IF dedicated moderators are actively monitoring throughout.

Configuration E - Testing Phase
(First implementing AI Moderation)

• Balanced threshold (~75%)
• Required Review for 1–2 test boards; Auto-Approved for remaining boards
• Board Context ON for test boards
• Appeals disabled initially, only enable after observing patterns
• Monitor AI Moderation Analysis modal for all flagged content