How Does AI Max for Shopping Beta Work?
Published
Written by Christopher Krassnig, Founder & CEO of ZenoX Media. Answers every question here from the accounts ZenoX runs, not from theory.
The short answer
AI Max for Shopping is a beta that uses your Merchant Center feed to match ads to what shoppers want. Text customization can rewrite product titles. Final URL expansion can add text ads and other landing pages. You can use text customization alone. Our advice at ZenoX: check access, feed language and bidding before you plan a test.
Source: Google Ads Help
When Did Google Announce It, and Is It Still a Beta?
Google announced AI Max for Shopping on 30 April 2026 in its launch post. The current Help page calls it "a beta suite of features". We checked both on 12 November 2026. A beta label does not mean every account has access.
What Changes When I Turn It On?
The beta has two parts: text customization and Final URL expansion. Google limits text customization to English feeds. Final URL expansion only works with a Target ROAS bid strategy, and it needs text customization on.
Google's launch post says you can turn Final URL expansion off at any time. Then your ads stay Shopping ads only. That matters if you want to test new titles before you let Google pick other pages.
What Would a Useful Feed Example Look Like?
Our example: a store sells a navy linen shirt. Before the test, compare the feed and the product page for material, color, size and care. Remove any claim you cannot prove.
Then read the new ad titles against those facts. A title that calls the shirt waterproof fails our check unless the product proof backs it. More reach is not worth a false promise.
What Should I Check Before and After a Test?
Our preflight: confirm the feature shows in your account, pick which settings to test, note the current budget and bidding, and save the starting reports. Check your allowed landing pages and tracking links. Keep the rest of the campaign steady while it runs.
Google's reporting guide names four reports: product titles, assets, landing pages and search terms. In the landing pages report, "Automatically selected" marks pages Google AI picked. Use these reports to see what actually served.
When Does This Not Apply?
Do not plan text customization for a feed that is not in English. Do not plan Final URL expansion without Target ROAS. This is also separate from the Editor 2.14 change, which only covers new Search campaigns. For cross-channel work, review Performance Max.
Where Can I Check the Rest of My Setup?
Start with the Editor 2.14 checklist. Then read Turn off Final URL Expansion, Track AI Max Traffic and Write an AI Brief.
Where Can I Get Help Applying This?
For help with the account, see our Google Ads agency, Shopify Google Ads service and Google Shopping service. To learn with other store owners, visit the Google Ads Ecom Lab on Skool.
Common Questions
- Is AI Max for Shopping Generally Available?
- No. Google's Help page calls it a beta. Check that it shows in your account before you schedule a launch.
- Can I Test New Titles Without Other Landing Pages?
- Yes. Google says you can use text customization on its own.
- Does This Replace Performance Max?
- No. Google's launch post keeps Performance Max as the recommended choice for cross-channel campaigns.
- Can I Judge Success from Clicks Alone?
- Our advice is no. Write down the sales result you want before the test. Then check sales value and cost next to the pages and ads that served.
Related questions
Keep Going
- How Do You Turn off AI Max Final URL Expansion?
- Which ValueTrack Parameters Identify AI Max Traffic?
- How Do You Use AI Brief in Google Ads?
- What Is the Fastest Way to Fix Low Click-Through Rate on Google Shopping?
- Why Do Some Dropshipping Products Get Zero Impressions on Google Shopping Even After Merchant Center Approval?
Keep reading
More from the ZenoX team.
On the call, we look at your account live. If we can grow it, we will show you how. If we cannot, we will tell you that too, and point you at someone who can.
We work with brands serious about scale. If that is you, let's talk.
