Google Ads AutoPilot settings explained
Understand automation levels, the actions you allow, minimum AI confidence, risk, budget limits, learning evidence, daily limits and the waiting period.
- Make sure primary conversions represent real business outcomes such as purchases or qualified leads.
- Allow synchronisation to finish before judging new recommendations.
- Always read the campaign and account context instead of acting on the title alone.
- Change one decisive factor at a time when possible and let the measurement period finish.
Why this matters
In Advisionr this feature is part of a wider decision process. Current Google Ads data, account settings, previous changes and measured outcomes are evaluated together. The same feature can therefore lead to a different recommendation in two accounts, which is exactly what you want from account-specific optimisation.
For the best result, first make sure synced data is current and that your primary conversions represent real business value. Then use Advisionr’s explanation to understand why a recommendation exists. A change is only truly useful when you can also measure afterwards whether it improved the result.
Settings and recommended choices
The best setting depends on your account, budget and risk tolerance. Use these values as a practical starting point rather than a universal rule.
| Setting | Recommended | Why |
|---|---|---|
| Minimum AI confidence | 90% production | Higher values allow fewer but better-supported automatic actions. |
| Maximum risk score | 30/100 production | Blocks suggestions above your chosen risk level. |
| Budget change per action | 10–20% | Smaller steps are easier to control and measure. |
| Daily actions | 10 | Limits the total automation pace. |
| Waiting period | 24 hours | Prevents rapid repeated changes to the same campaign or item. |
| Learning evidence | On for conservative use | Adds historical proof requirements on top of fixed safety. |
How to get the best result
Advisionr performs best when data, goals and responsibilities are clear. Use this working method to improve the quality of analyses and decisions.
- Allow synchronisation to finish before judging new recommendations.
- Only broaden AutoPilot after checking goals, budgets and the action types you allow.
- Review History regularly to see both executed and deliberately blocked actions.
Common mistakes
- Judging an optimisation on one day or one metric without context.
- Making many changes close together and losing causal clarity.
- Opening AutoPilot limits before conversion tracking and budget safety limits are verified.
- Expecting every analysis to become a One-Click action even when evidence is insufficient.
- Confusing an automatically closed suggestion with an executed Google Ads change.
Frequently asked questions
How often should I review {title}?
A weekly review is a good minimum for active accounts. Review more often during major changes, but avoid reacting daily to normal volatility.
Why can Advisionr analyse something without creating an action?
Analysis and execution are deliberately separate. The opportunity may lack volume, overlap with a running measurement, have no safe way to carry it out, or not meet your certainty and risk limits.
Can Full AutoPilot always execute this automatically?
Only when the action type is enabled and every fixed safety check passes. Full AutoPilot does not mean every possible change is executed blindly.