How Do I Review AI Advertising Decisions?

The AI Decisions page provides a centralized log of the automated optimization actions performed by EVA’s advertising system. These decisions reflect the actions taken by EVA to improve campaign performance based on campaign data, optimization rules, and platform logic.

Instead of requiring advertisers to manually monitor every campaign change, EVA evaluates campaign performance continuously and applies optimization actions when conditions defined in the system are met. The AI Decisions page allows advertisers to review those actions and understand how the system is optimizing campaigns.

Using this page, advertisers can:

The AI Decisions interface functions as a transparency and monitoring layer for EVA’s automation system.

Step 1: Open the AI Decisions page

To review automated optimization decisions:

This page displays a table containing the decisions made by the AI optimization engine.

Each row in the table represents a single automated action applied to a campaign, ad group, keyword, or product target.

The table provides a historical record of optimization decisions generated by the platform.

Step 2: Understand the decision table structure

The AI Decisions page is structured as a decision log table that records the actions performed by EVA’s optimization system.

Each row in the table represents a specific decision taken by the AI.

The table typically contains columns that describe:

This table allows advertisers to review exactly what the system changed and when those changes occurred.

Because the table is chronological, it can also be used to analyze how optimization behavior evolves over time.

Step 3: Identify different decision types

The AI optimization system performs several different categories of optimization actions. These decisions correspond to different optimization mechanisms configured in the Advertising system.

Common decision types include:

Bid decisions

Bid decisions occur when EVA automatically adjusts bids in order to improve campaign efficiency or performance.

These adjustments typically result from optimization rules that evaluate campaign metrics such as performance efficiency or conversion behavior.

Example actions may include:

These decisions are typically generated by bid optimization rules.

Negative keyword decisions

Negative decisions occur when EVA determines that certain search terms are underperforming and should be excluded from campaign targeting.

In this case, the system automatically adds those search terms as negative keywords to prevent further spend on queries that are unlikely to convert.

These decisions help improve targeting efficiency by preventing ads from appearing for irrelevant or poorly performing search queries.

Status decisions

Status decisions occur when the system automatically changes the status of advertising entities.

Examples may include:

These decisions are typically triggered by optimization rules related to inventory availability or campaign performance.

Step 4: Filter decisions by type

The AI Decisions interface provides filtering tools that allow advertisers to focus on specific types of optimization actions.

Filtering can be used to isolate decisions such as:

This is useful when investigating a specific category of automation activity.

For example, if an advertiser wants to review all recent bid adjustments applied by the system, they can filter the table to display only bid decisions.

Filtering simplifies the analysis of large decision logs.

Step 5: Use search to locate specific decisions

The AI Decisions page also includes a search function that allows advertisers to locate specific decisions.

The search function can be used to find decisions related to:

This allows advertisers to quickly identify the optimization actions applied to particular campaigns or entities.

Step 6: Adjust the date range

The AI Decisions page includes a date range selector that allows advertisers to control which time period is displayed in the decision log.

This allows users to analyze decisions within a defined time window.

For example, advertisers may choose to review:

Changing the date range updates the table to show only the decisions that occurred within the selected time period.

Step 7: Customize table columns and presets

The AI Decisions table can be customized to display the information most relevant to the advertiser.

Customization options may include:

Column presets allow advertisers to save specific table configurations and quickly switch between different views depending on the type of analysis being performed.

For example, one preset may focus on campaign-level decisions while another focuses on keyword-level optimization actions.

Step 8: Review decision history

By combining filtering, search, and date range selection, advertisers can investigate how the AI system has optimized campaigns over time.

This historical view allows advertisers to:

The AI Decisions page therefore acts as a monitoring tool for EVA’s automated optimization system.