Approval design guide
How to give people useful control over AI decisions
Human-in-the-loop AI means a person reviews a defined decision before the work continues. Show that person the proposed action and the evidence needed to judge it. Let them approve, reject, correct, defer, or ask for help. Keep their decision in the work history.
- Good fit
- Use human review when policy interpretation, customer consequence, irreversible action, uncertain evidence, or professional accountability makes automated commitment inappropriate.
- Pause when
- Adding a generic approval button is not meaningful oversight if the reviewer lacks source evidence, cannot change the recommendation, or faces a queue too large to examine responsibly.
When this approach helps
- The system can prepare a case reliably but cannot justify final judgment across exceptions.
- Different roles own content quality, security, legal, financial, or customer-impact decisions.
- Reviewers spend most of their time finding context instead of evaluating the proposed action.
Follow these steps
- 01
Define the review question
State the exact question, allowed outcomes, reviewer permissions, and deadline. Ask for approval of the specific action that needs judgment.
- 02
Put the evidence beside the action
Show source links, excerpts, check results, applicable rules, uncertainty, and the proposed change together. Separate observed facts from AI interpretation. Highlight missing information.
- 03
Let the reviewer correct or stop it
Allow edits, requests for evidence, reassignment, and stopping. Record a reason without requiring an essay for routine decisions. Make urgent exceptions visible without bypassing the rules.
- 04
Check whether review works in practice
Measure waiting time, review time, corrections, disagreement, missed exceptions, and abandoned tasks. Reduce volume or add staff if reviewers cannot examine the work properly.
Documents and records to keep
Decision packet
A compact, source-linked view separates facts, recommendation, policy checks, uncertainty, missing evidence, and the exact action awaiting authority.
Role and escalation map
The map names the primary reviewer, backup, specialist escalations, response target, and what the system does while the decision is pending.
Review quality receipt
The receipt records the presented evidence, reviewer outcome, corrections, reason code, action result, and time needed without claiming that approval alone proves quality.
Common questions
Which AI decisions should always involve a person?
Require a responsible person where law, policy, professional duty, irreversible impact, unclear evidence, or meaningful customer consequence demands accountable judgment. The boundary should be based on the action and context, not the model brand.
How can teams prevent approval fatigue?
Narrow the decision, group related evidence, remove low-value notifications, automate only well-proven cases, use risk-based thresholds, and measure queue health. If reviewers routinely approve without reading, reduce volume or authority.
Does human approval make an AI system safe?
Not automatically. Safety also depends on source quality, reviewer competence, interface clarity, available time, permission enforcement, monitoring, and recovery. Approval is one control whose effectiveness must be tested in operation.