Set a goal — CTR, CPA or CPV — and the platform tunes the campaign toward it: lowering bids on placements that cost too much per result, stopping placements with clicks but no results, moving budget to the best placements and blocking sources that fail traffic checks. Every automatic change is listed with its time.
What it needs
Conversion data. AI Optimization steers toward outcomes, so it needs the tracker wired up and conversions coming back. Without that it has nothing to optimize against — see Setting up conversion tracking.
Where its actions appear
In the Automated Rules log, alongside your own rules. Nothing it does is hidden.
How it differs from your own rules
- Rules do exactly what you specified, at thresholds you chose.
- AI Optimization determines the thresholds itself from conversion data and keeps revisiting them.
Many operators run both — rules as hard safety limits, AI Optimization for tuning inside them.
Setting expectations with advertisers
It starts acting once conversions arrive, not the moment it's switched on. An advertiser who enables it on a fresh campaign and expects instant change will be disappointed; tell them to give it real data first.