Human Oversight SOP Template for AI SaaS

Purpose

Define when humans review AI outputs, who can override them, and how incidents are escalated.

Scope

Applies to: [feature list]

Roles

Role	Responsibility
Reviewer	Reviews flagged outputs before final action
Owner	Accountable for SOP updates
On-call	Handles incidents

When human review is required

• [ ] Fully automated actions affecting people
• [ ] Scores below / above threshold X
• [ ] User contest / complaint
• [ ] Novel edge cases

Override procedure

1. Reviewer opens the case in [tool]
2. Records reason for accept / reject / edit
3. Final action executes only after confirmation
4. Decision logged (who/when/why)

Logging (recommended always; legal if high-risk)

Store: input reference, model version, output, human decision, timestamp.

Art. 12 logging is a legal requirement for high-risk systems (Annex III duties from 2 December 2027 if applicable).

Review cadence

• Weekly sample review of [N] decisions
• Quarterly SOP update

Related

Free readiness brief · Recruiting checklist