Google has published a help page for a new Search ad format that writes review summaries directly into paid listings across four English-speaking markets. The feature, called Review summaries, uses Google AI to pull positive feedback from online reviews and display it as short AI-generated commentary inside Search ads.
The catch: advertisers cannot edit the text. Google decides what appears, and only shows it when it predicts the extra copy will improve ad performance.
How the review summaries work
According to the help page, the system aggregates online reviews from across the web and generates summaries that are positive only. The copy is meant to help shoppers evaluate a business before they click.
Eligibility is narrow for now:
- Search campaigns only
- All ad positions on Google.com
- US, UK, Australia and India, in English
- All devices and verticals except sensitive categories such as pharmaceuticals
Google says it uses responsible AI practices and strong grounding requirements to keep summaries accurate. Grounding means tying generated text to source reviews, reducing the risk of fluent but false claims.
What advertisers can’t control
The help page is direct: advertisers cannot directly edit or customize the generated text. That is a meaningful shift. With automatically created assets, advertisers usually get campaign-level or account-level controls. Here, Google is framing the copy as third-party opinion, not advertiser speech.
The page also leaves big practical questions open. It does not say whether advertisers can opt out, which review platform feeds the summaries, whether a disclaimer appears in the ad, how reporting works, or how a mostly negative business is treated.
That last gap matters. A positive-only summary is a selection, not a balanced representation. Marketers will want to know whether a low-rated brand can still get a glowing line pulled from a handful of good reviews.
Why this matters now
Search ads are already under pressure from AI changes on the results page. A Seer Interactive study cited by PPC Land found paid click-through rates on queries that triggered AI Overviews fell 68% between June 2024 and September 2025. Extra trust signals inside ads could be Google’s way to claw back engagement, but the help page includes no performance data.
Accuracy is not hypothetical. NP Digital research from February 2026 found 47.1% of marketers encounter AI inaccuracies several times a week, and 36.5% have had incorrect AI content published publicly. An error in an organic AI answer is one thing; the same mistake inside a paid placement on your own ad is harder to brush off.
What to do about it
Start by checking your live Search ads in the four markets. If a review summary shows up, screenshot it and ask your Google rep how it is sourced and whether it can be disabled at account level. Keep a log of any inaccuracy, because paid copy you cannot correct is a brand-risk issue, not just an optimization one.
Also treat this as an early signal. Google is clearly still testing review formats in Search ads, from ‘What people are saying’ to ‘What customers love.’ The help page may be the first official documentation, but the visuals and controls are likely to keep moving.
Source: PPC Land
