Manual and automatic control are explicit
You can evaluate individual suggestions yourself. Automatic mode applies the configured eligibility and guardrail policies; it does not simply approve every suggestion.
Can I see why Nexus recommends an ad change and decide whether to apply it?
Review advertising changes before applying them
Yes. The AI Optimization workspace shows campaign suggestions with supporting detail, an apply or dismiss decision, and a history of changes. You can review suggestions yourself or configure automatic optimization within the available policies. Simulation and proof views help inspect assumptions and recorded outcomes; they do not guarantee better results.
Review advertising changes before applying them
Start with connected campaign results
Connect the relevant advertising accounts and conversion data. Open AI Optimization and load the pending suggestions for the campaign activity available in the account.
Inspect the proposed change
Open a suggestion to review the affected campaign or asset, its supporting detail, and the proposed action. Decide whether the available evidence and business context justify changing the campaign.
Apply or dismiss the suggestion
Use Apply for a change you want or Dismiss for one you do not. Check the result returned by the operation. Bulk controls can process selected suggestions while reporting individual failures.
Set the limits for automatic decisions
If you want automatic application, configure Auto-Optimization Mode and its policies. The pending view distinguishes eligible changes from changes on hold and shows the available reasons for a hold.
Review change history and outcomes
Return to History to inspect changes after they run. Compare the available recorded outcomes with the decision you made and review the supported rollback option for a change that needs reversal.
Use simulations as planning tools
Run a simulation or compare saved snapshots when evaluating an experiment configuration. Keep predicted outcomes separate from the campaign results actually recorded after a change.
A hold can be useful information
The pending view explains available eligibility and hold reasons, helping staff distinguish an actionable suggestion from one waiting on more evidence or a policy condition.
Measurements depend on the connected data
Campaign metrics, attributed outcomes, and experiment evidence determine what the analysis can evaluate. Missing or unreliable conversion data limits the conclusions.
Before you get started
Questions about this workflow
Can I review a suggestion before it changes my ads?
Yes. Open the suggestion detail and apply or dismiss it. Automatic mode is a separate configuration choice.
Can I see changes that already ran?
The History view retains change records. Review the available result and supported rollback action for the specific change.
Does the simulation show guaranteed future performance?
No. It models the selected scenario. Use it alongside the actual results and attribution available for the campaign.
Walk through your work in Nexus
Bring an example of the work you manage today. We’ll demonstrate the steps, customer view, and setup relevant to your team.