You are currently viewing automation policy in read-only mode. Only Organization Admins can modify execution authority and guardrail thresholds.

Automation Policy & Safety Gates

Configure autonomous decision authority, risk ceilings, and deterministic safety guardrails.

Determines whether AI recommendations can be dispatched automatically or require human sign-off.

Gate #3

MANUAL

Human review strictly required for 100% of recommendations. Zero autonomous mutation.

Recommended for new ad accounts

SUPERVISED

Autonomous dispatch permitted only for LOW risk actions that satisfy all guardrails.

Balanced safety and velocity

AUTONOMOUS

Full autonomous optimization within configured risk and budget caps.

Requires explicit Admin activation

Minimum Confidence Threshold80%

Recommendations with confidence score below this threshold will require manual human approval (Gate #4).

Maximum Budget Delta (%)±20%

Single-step budget changes exceeding this percentage are automatically blocked (Gate #7).

Maximum Single-Action Spend Change ($)

Dollar-denominated cap on single budget scaling actions (Gate #8).

Daily Autonomous Action CapMaximum number of autonomous actions allowed within a rolling 24-hour window (Gate #9).

Permitted Risk Level Ceiling

LOW Risk AllowedMEDIUM Risk AllowedHIGH Risk Allowed

Permitted Action Types (Gate #6)

Budget ChangesPause UnderperformersResume CampaignsBid Adjustments

Last updated by: System Admin