Turn On Auto-Apply

This guide helps you roll out automation carefully so coverage increases only when trust does.

Fast Path

Start with the setup wizard, choose the closest preset to your real tolerance, launch in Shadow Mode first, and use Activity Log plus trust signals to decide when live automation is justified.

How To Use This

The most useful automation question is not "How aggressive can we be?" It is "Which decisions do we already make consistently enough that the system can take them first?"

Use the presets as operating stances:

• Conservative when mistakes are expensive and review bandwidth exists.
• Balanced when you want useful coverage but still need meaningful guardrails.
• Aggressive only when your team understands the risk tradeoff and the account can tolerate broader execution.

What To Watch For

Shadow Mode is the safest starting point because it shows how the system would behave without forcing you to trust it blindly. If shadow decisions regularly disagree with what your team would do, do not fix that with more confidence alone. Fix the mismatch first.

Live automation should be paused when the safety signals are telling you something changed in the account, the API, or the recommendation quality. Frequent pauses are a root-cause investigation, not a reason to simply loosen thresholds.

Common Problems

If Auto-Apply is enabled but nothing executes, the usual gates are pause state, a tripped circuit breaker, excluded recommendation types, excluded campaigns, or thresholds that the current queue does not meet.

If execution volume suddenly feels too high, reduce scope before you raise tolerance again. Narrowing by type or campaign is usually a faster control than rewriting the whole profile.

FAQ

What is the safest first move when setting this up?

Use Shadow Mode first and compare its decisions with the decisions your team would actually approve.

Does Auto-Apply remove manual control?

No. Manual review, history, and revert workflows still matter even after live automation starts.

Next Steps

Use Trust and Smart Tuning to judge whether automation is trustworthy enough to increase, or Advance Through Autonomy to plan the next maturity step.