Governance Officer¶

Users responsible for risk, control, and audit of AI usage. Usually a different role from the system administrator.

Start: automatic redirect right after login (or click the XGEN logo at the top-left)
Procedure: Dashboard · Governance Officer view
Governance Officers (with admin.governance:*) see governance-only widgets — Risk-Level Policy status, Banned-Term Policy status, and the agent approval queue — on the main screen.

Review and approve agents that already cleared System Administrator approval (governance approval, stage 2)¶

Start: Admin Center → AI Governance → Agentflow Approval
Procedure: AI Governance · Agent Approval
Flow: only agents that already passed stage 1 (Agent Management — Deployment Approval) appear in this queue. Items pushed in by a risk-threshold breach in Risk Review join the same queue.
Queue view: click any of the four stat cards (All / Pending / Approved / Rejected) at the top to filter instantly.
Review items: node layout, author, department, parameters (look at PII-exposing nodes first), category.
Decision: Approve (comment optional) or Reject (comment required). Both write governance_reviewed_by + governance_review_comment + timestamp to AI Service Change History permanently.
Outcome: approved agents flip to servable immediately. Rejected agents return to the author, who fixes them and resubmits from stage 0.

Start: Dashboard · Agent deployment/approval status widget — the Governance pending counter
Procedure: click the counter → open the Agent Approval queue and use the Pending stat card to batch-process
Operations: a rising counter while Deployment pending stays normal means stage 1 is fine but governance is the bottleneck. Sweep the queue at least weekly.