AI Overviews Now Answer 'Reviews' Searches Too

You type your own business name plus the word "reviews" into Google, expecting the usual wall of star ratings and review sites. Instead there's a written summary on top, and you're a sentence or two inside it, described by whatever the AI could find. Nobody on your team wrote that sentence.

That's the shift this study puts a number on. Comparison searches, the ones people run right before they buy, used to be a blue-links neighborhood. They're now mostly an AI Overview neighborhood.

What moved, and by how much

Semrush didn't sample random searches. It picked 400 commercial keywords built around five phrases buyers use when they're choosing between options, then checked each month whether Google put an AI Overview on top.

• Alternatives ("X alternatives")
• Brand vs. brand ("X vs Y")
• Best X for Y
• Price and cost
• Reviews

Across all 400 keywords, coverage went from 45.75% to 88% in a year. Reviews was the dramatic one: under a third of "review" searches had an AI Overview in September 2025 (Semrush's methodology puts the baseline at 30.14%), and it was 90% by August 2026.

One more number worth reading slowly. When Semrush weighted keywords by how many people search them, coverage rose to 97.08%. The keywords still without an AI Overview were less than 3% of search volume, so the few holdouts are mostly rarely-searched phrases.

Why a local owner should care anyway

Look at the panel: banking, ecommerce, software. That's not your dentist, your landscaper, or your boutique. We aren't going to pretend the 88% applies to a "plumber near me" search. It doesn't, and our post on AI Overviews skipping most "near me" searches shows the local-intent split.

But your customers don't only search "near me." Before they call anyone, plenty of them ask comparison questions: is this company any good, who's better, what does it cost, is there something cheaper. Those are the same five shapes the study measured. The study says the habit of answering them with a generated summary is now close to universal in the categories it tracked, and it's reasonable to expect the shapes to spread. That expectation is ours, not Semrush's data, so hold it loosely.

The practical consequence is simple. For these searches, the summary is the first thing a buyer reads, and it's assembled from whatever your site, your profiles and third-party pages say. This is the second search described in our guide to Answer Engine Optimization (AEO): ranking on the page matters less than being the source the summary draws from.

Checking what the summary says about you

This takes twenty minutes and costs nothing. Run these searches in a private window, using your real business name and your real competitors:

1. "[your business] reviews"
2. "[your business] vs [closest competitor]"
3. "[your business] alternatives"
4. "[your service] cost" plus your region's name
5. "best [your service] for [your typical customer]"

For each one, write down whether an AI Overview appeared, what it says about you, and which pages it links to as sources. You're looking for three things: wrong facts (hours, services, prices), thin facts (a one-line mention where a competitor gets a paragraph), and sources that aren't yours.

Then fix what you control:

• Put a plain pricing or "what it costs" page on your own site. If you won't publish a number, publish the factors that move it. A summary can only repeat what exists.
• Write a short, honest comparison page covering who you're a good fit for and who you aren't. Summaries favor pages that answer "X vs Y" directly.
• Keep review profiles current and replied to. The summary leans on third-party review pages, so a stale profile is a stale answer. Our post on why your reviews have an expiration date covers the timing.
• Fix wrong facts at the source. The error is rarely in the AI. It's in an old directory listing the AI trusted.

Moves that won't help here

• Paying a vendor to "optimize for AI Overviews" without showing you the five searches above. If they can't show you what the summary says today, they can't show you what changed.
• Stuffing "reviews," "alternatives" and "vs" into every page title. A page that's about nothing but the keyword gives the summary nothing to quote.
• Writing fake comparison pages that trash named competitors. It's unsourced, it invites trouble, and it reads like what it is.
• Treating 97% as your number. It's a volume-weighted figure for a specific panel of commercial keywords. Your searches may look very different, which is why you check them yourself.

FAQ

Do AI Overviews appear on 88% of all Google searches?

No. The 88% is the share of 400 specific commercial keywords, in banking, ecommerce and software, that showed an AI Overview in August 2026. It says nothing about informational, navigational or local-intent searches.

Does this study cover local businesses?

No. The panel was U.S. desktop searches in three industries, none of them local services. It's useful as a direction for comparison-style searches, not as a measurement of local results.

What's the difference between keyword-weighted and volume-weighted coverage?

Keyword-weighted counts every keyword equally, which gave 88%. Volume-weighted counts popular searches more heavily, which gave 97.08%. The gap means the keywords without an AI Overview are mostly low-volume ones.

Why did "reviews" searches change so much?

Semrush reports the jump, from under a third to 90%, but the study doesn't explain why Google made the change. Anything beyond the numbers is guesswork.

Can I block my business from appearing in these summaries?

Not in any way we'd recommend. Opting out of summaries generally means opting out of being a source, and these are searches where buyers are deciding. Fix the facts instead.

How often should I check what the summary says about me?

Monthly is enough for most small businesses, and right after any change to your prices, hours or services. Keep the screenshots so you can see what changed.

Want someone to run those five searches every month and fix what they find? That's part of what we do on getting found on Google and in AI answers.