Feed it your best page. The most content-rich page for what the campaign sells - specific beats general, text beats imagery.
Spend real minutes in Discovery. It's the highest-leverage screen in the wizard; every downstream decision is built from that brief. Sharpen the USPs especially - they become your headlines.
Use exclusions. Telling the AI what you don't sell is as valuable as what you do - it steers keywords and negatives at once.
Review every ad group, not just the first. Each has its own angle and its own copy; a wrong claim hides in the group you skipped.
Trust the quality gate's cuts. A shorter keyword list isn't a thinner campaign - it's a campaign with the waste already removed. Add proven terms you know; don't pad.
Launch paused, activate deliberately. The recommended flow: push, glance over it in place, then switch it on in Clicc.
Let the rest of Clicc take over. Post-launch, Search Terms grows the keyword and negative lists from real queries, and the alert rules watch spend, serving and performance - the build is step one of a managed lifecycle, not a one-off.
RELATED ARTICLES
Reviewing & correcting Discovery
Two minutes here shapes everything the AI builds.
After launch
Where the campaign lives once it's running.
Keywords & the quality gate
How keywords are found, enriched, and ruthlessly filtered.
