Helping advertisers test campaign changes before going live

Advertisers managing large-scale campaigns had no reliable way to evaluate changes before applying them to active campaigns. Existing testing workflows were complex, difficult to trust, and rarely reused – and outside of the advertisers’ primary workflows.

I led the design of a new AI-powered testing experience that helps advertisers identify opportunities, create experiments, monitor results, and apply recommendations directly within Ads Manager.

The core challenge wasn't automation—it was trust. Because recommendations can affect significant advertising spend, I designed the workflow around transparency, review points, and user control. The AI agent suggests experiments, explains its reasoning, monitors outcomes, and recommends next steps while keeping advertisers in control of final decisions. Advertisers can see how a change would impact their campaign, before adopting it.

Detailed case study

Available at interview presentation