Content Governance & Moderation

Content Governance & Moderation

1. What it is

Content Governance & Moderation are two platform-level admin surfaces that protect the quality and safety of content across your tenant. Content Moderation screens user-generated content in real time — blocking, holding, or flagging posts, comments, chat messages, and AI inputs before or as they are published. Content Governance runs periodic scans across published content in 16 apps, surfacing stale pages, broken links, overdue reviews, and app-specific compliance gaps.

• Enablement: both surfaces are available to every admin under Admin → General Settings — no separate app license is required. Individual governance scans are enabled per app in the governance settings.
• What it is not: these are admin-facing compliance tools. End users do not interact with either surface directly; they see the effects (a blocked post, a held comment) but never the queue or findings list.

2. Standing it up

Content Moderation

1. Go to Admin → General Settings → Content Moderation → Settings.
2. Under What to Moderate, choose a moderation type. The default is Moderate only flagged content (users report content manually). The other modes are Moderate all content (every piece of content is held for review before publishing) and Keyword + AI moderation (automated screening via keyword lists and optional AI).
3. If using Keyword + AI moderation, add custom keywords in the Custom Keywords field (comma-separated). The platform also checks a built-in library of offensive word stems and patterns.
4. Optionally enable AI Moderation to add LLM-based screening. AI moderation starts in Observation mode (logs without blocking). Switch to Live mode when you are satisfied with the detection quality.
5. Under Which Users / Teams?, configure user scope: by default All users are screened. You can switch to Exempt specific users / teams (screen everyone except named users/teams) or Only specific users (screen only named users).
6. Admins and super admins are always exempt from screening.
7. Verify by going to Admin → General Settings → Content Moderation → Queue — this is where flagged and held items appear for review.

Content Governance

1. Go to Admin → General Settings → Content Governance → Settings.
2. Each registered app appears as a card. Toggle Enable Auto Governance for each app you want scanned.
3. Set the review interval per app — items not updated within this window are flagged as stale.
4. Optionally enable broken link checks and gap analysis per app.
5. The governance scan runs automatically every Monday at 7:00 AM UTC. You can also trigger an on-demand scan from Admin → General Settings → Content Governance → Findings → Run Scan (throttled to once every 5 minutes).
6. Verify by checking the Findings tab for results after the scan completes.

3. How it fits together

Content Moderation vs Content Governance — these two surfaces solve different problems and should not be confused. Moderation is real-time and acts on individual submissions (a broadcast comment, a chat message, an AI prompt) at the moment of creation. Governance is batch and acts on published content that has gone stale or broken over time. Moderation asks “should this be published?”; governance asks “is what we already published still valid?”

Moderation types — the three moderation modes determine how content enters the review pipeline. Moderate only flagged content is passive: content publishes normally and users can report it, which creates a finding in the queue. Moderate all content holds every piece of content for admin review before it publishes. Keyword + AI moderation runs automated checks (keywords, patterns, optional AI) and only holds or blocks content that triggers a rule.

Observation vs Live AI mode — when AI moderation is enabled, it can run in two modes. Observation mode logs what the AI would flag without actually blocking or holding content — use this to evaluate detection quality before enforcing. Live mode actively blocks or holds content the AI flags.

Governance adapters — each app that participates in governance registers an adapter with the central registry. The adapter defines what content to scan, how long before content is considered stale, and what app-specific checks to run. Currently 16 apps have governance adapters: Libraries, Wikis, Sites, Drive, Workspaces, Comms Hub, News Feed, Policy Hub, SOP Hub, Contracts, Inspections, Onboarding Hub, Safety Hub, Mango GTM, Mango IQ, and Workforce Bots.

Moderation policies — each content surface (broadcasts, recognition, chat, Ask AI, forms, etc.) registers a moderation policy that controls which checks apply to that surface. Policies resolve settings through a precedence chain: per-app configuration overrides the global Content Moderation settings, which override the platform defaults.

Severity escalation — governance findings that remain open are automatically escalated. Low-severity findings escalate to medium after a configurable number of days (default 30). Medium findings escalate to critical after another configurable window (default 60 days). Findings rated High by an adapter are excluded from auto-escalation.

4. Running it

Reviewing the moderation queue

1. Go to Admin → General Settings → Content Moderation → Queue.
2. The queue shows items that are Pending Review. Filter by source type, issue type, or severity.
3. For each item, choose Approve (publishes the content and notifies the author) or Reject (removes it and notifies the author). You can add notes to either decision.
4. A separate panel below the queue shows unscreened publishes from the last 7 days — content that went live because the automated screen did not return in time (the system fails open). These are informational records, not items awaiting a decision.

Working the governance findings list

1. Go to Admin → General Settings → Content Governance → Findings.
2. Filter by app, issue type, severity, or status. Snoozed findings are hidden by default; toggle “Show snoozed” to include them.
3. For each finding: Acknowledge (mark as seen), Resolve (mark as fixed), Dismiss (mark as not actionable), Assign (delegate to a user), or Snooze (defer until a future date — the finding stays open but drops out of the active worklist).
4. Use bulk actions to acknowledge, resolve, or dismiss multiple findings at once.
5. Export the current filtered view to CSV from the findings page.

Triggering an on-demand governance scan

1. Go to Admin → General Settings → Content Governance → Findings.
2. Click Run Scan. The scan is queued as a background job.
3. A cooldown prevents re-running within 5 minutes of the last scan.

Reviewing governance analytics

Go to Admin → General Settings → Content Governance → Analytics for summary stats, daily trend, breakdowns by issue type/app/severity, top affected containers, and resolution time. The Gap Analysis sub-page (available when gap analysis is enabled for at least one app) shows zero-result search rates and top unmet queries.

Reviewing moderation analytics

Go to Admin → General Settings → Content Moderation → Analytics for summary stats, daily trend, breakdowns by issue type/source/severity, top flagged users, moderator activity, average review time, and AI security metrics (input blocks, PII detections, conversation safety flags).

5. Settings

Content Moderation settings

Found at Admin → General Settings → Content Moderation → Settings.

Setting	Default	What it changes
Moderation type	Moderate only flagged content	How content enters the review pipeline: user-flagged only, hold everything, or keyword + AI screening
Custom Keywords	(empty)	Comma-separated list of additional words to block, merged with the built-in offensive word library
AI Moderation enabled	Off	Whether AI-powered content analysis runs on screened content
AI Mode	Observation	Whether AI detections log only (Observation) or actively block/hold (Live)
AI Harassment detection	On	Whether AI screens for harassment
AI Discrimination detection	On	Whether AI screens for discrimination
AI Negativity tolerance enabled	Off	Whether AI applies negativity scoring
AI Negativity tolerance	50	Slider (0–100) mapping to sensitivity: 0–25 strict, 26–50 moderate, 51–75 relaxed, 76–100 lenient
User scope	All users	Who is screened: all users, exempt specific users/teams, or only specific users
Notify moderators on queue	On	Whether moderators receive in-app + email alerts when an item enters the queue
Notify authors on decision	On	Whether content authors receive email when their item is approved or rejected
Team admins can moderate	On	Whether team admins can review items in the moderation queue
Output PII scanning	On	Whether AI-generated output is scanned for PII (SSN, credit cards, bank accounts)

Content Governance settings

Found at Admin → General Settings → Content Governance → Settings, configured per app.

Setting	Default	What it changes
Enable auto governance	Off (per app)	Whether the weekly scan includes this app
Review interval (days)	90 (Libraries adapter)	Items not updated within this window are flagged as stale
Check broken links	On	Whether external URLs in content are verified during scans
Notify on governance review	On	Whether container reviewers receive notifications when findings are detected
Enable gap analysis	Off	Whether search-log analysis runs for this app to surface knowledge gaps
Notify on gap summary	On	Whether admins receive a digest of gap analysis results
Escalation: low → medium (days)	30	Days before a low-severity finding escalates to medium
Escalation: medium → critical (days)	60	Days before a medium-severity finding escalates to critical

6. More help

• Content Governance FAQ — specific questions about governance scans, moderation decisions, and troubleshooting
• Audit Logs & Activity History — tracks who changed what across the platform, including moderation decisions
• Ask AI — the assistant answers questions about Content Governance & Moderation from these articles.