Google Ads creates 1-click tests to compare AI Max with traditional search campaign

Google Ads updated, on October 26, 2026, the frequently asked questions page about AI Max campaigns for search with a new section: the one-click experiment recommendation. The change was spotted by Barry Schwartz, of Search Engine Roundtable, who also cited the observation made by Hana Kobzova, of PPC News Feed, about the update.

The text added by Google is direct: eligible accounts can find, in the Recommendations tab, a shortcut to automatically create a test comparing the current search campaign against a version with AI Max enabled.

What the button actually does

According to Google Ads documentation, clicking "Create 1-click experiment" makes the system build, on its own, a 50% to 50% split between the control group (the search campaign as it stands today) and the treatment group (the same campaign running with AI Max). There's no need to manually configure budget, duration, or traffic distribution: Google takes on these parameters.

Hana Kobzova summarized the practical effect of the change:

Testing AI Max just got easier and more granular for Search campaigns. Eligible accounts can now use a 1-click setup to create a 50/50 experiment automatically through Recommendations. Additionally, advertisers can manually refine experiments at the ad group level.

Hana Kobzova, PPC News Feed

Google's own FAQ page conveys the same logic in product language:

If your account is eligible, you may find a 1-click experiment recommendation on your Recommendations page.

Google Ads documentation on AI Max

Fine-grained control still exists, just tucked one step further

Schwartz notes that, in the past month, Google also added to the same documentation an answer about more granular tests, outside the one-click flow. When the experiment is set up manually through the Experiments tab, instead of the automatic shortcut, it's possible to decide, ad group by ad group, whether Search term matching and Asset optimization are turned on or off in the treatment arm. The control group keeps the campaign's original configuration intact.

In other words: the 1-click experiment and the manual experiment aren't the same feature disguised in two ways. They are two levels of decision.

Why this is a growth problem, not just an Ads configuration one

For whoever decides where channel budget goes, the news isn't "Google automated one click": it's that Google is optimizing the adoption friction of AI Max, not necessarily the quality of the decision to test. A recommendation on the Recommendations screen is, by definition, designed to convert clicks into activation of the product that Google wants the account to use more.

This doesn't invalidate the feature. Manually running a 50/50 A/B test, with budget configuration and an evaluation window, consumes time from a performance analyst who could be interpreting results instead of setting up the experiment. Automating this setup is a real productivity gain in accounts with enough volume for a statistically relevant test.

The risk runs the other way: in small accounts, or in campaigns with low conversion volume, an automatic 50/50 split can dilute the budget to the point where neither arm generates reliable data. In that scenario, accepting the one-click recommendation without looking at the expected sample size is the same old mistake, just faster to make.

In short: the 1-click experiment solves the operational friction problem, not the problem of knowing whether it's worth testing AI Max in that specific account. That second decision remains the work of whoever understands statistical power, seasonality, and what a change in search term matching does to the range of queries the campaign captures.

What changes in the workflow of whoever manages search accounts

In practice, performance teams should treat the 1-click recommendation as a shortcut to validate, not one to accept with eyes closed. Before clicking, it makes sense to check:

• Whether the target campaign has enough conversion volume for the test to reach significance within a reasonable timeframe;
• Whether the business's seasonal moment won't contaminate the comparison between control and treatment;
• Whether, after being created through the automatic path, the experiment can still be manually refined by ad group, something Google's documentation confirms is possible even in parallel with the shortcut.

For GEO and for AI-powered search more broadly, the move is also a signal: Google is designing the path of least resistance so that more accounts test (and, predictably, adopt) broader matching behavior and AI-optimized assets in search campaigns. Whoever measures campaign performance should track, in upcoming reports, whether the AI Max treatment arm changes the mix of queries captured, not just the final CPA: that detail is where the trade-off between reach and brand control in paid search lives.