AI Overviews Now Cover 'Best' and 'Reviews' Searches

You search "best [thing you sell] for [your kind of customer]" and Google answers before it shows a single link. That used to be the moment a buyer clicked three results, opened tabs, and picked a shortlist. Now the shortlist arrives pre-assembled, and either your name is on it or it isn't.

That's the shift this study puts a number on. It is also easy to over-read, so this post covers what was measured, what wasn't, and what is worth doing about it.

What Semrush tracked

The panel was 400 keywords in three industries, sorted into five buying-stage patterns: alternatives ("X alternatives"), "best X for Y," pricing and cost, reviews, and brand-versus-brand ("X vs Y"). Semrush checked the same keywords monthly for twelve months and recorded whether an AI Overview appeared.

Three numbers carry the story:

• Keyword-weighted coverage: 45.75% in September 2025, 88% in August 2026. Every keyword counts the same.
• Volume-weighted coverage: 97.08% in August 2026. The keywords without an AI Overview were 12% of the panel but under 3% of the estimated search volume, which means the searches people run most are the ones that got covered.
• Reviews: 30.14% to nearly 90%, a 59-point rise and the biggest of any group. A year earlier a reviews search triggered an AI Overview less than a third of the time.

"Best X for Y" ended highest at 96.12%. Alternatives gained 34 points and passed 82%. Brand-versus-brand started highest (63.41%) and gained least, to over 86%.

Why the buying-stage searches moved

These are the queries where someone has stopped learning and started choosing. "Reviews," "best," "alternatives," "price": nobody types those out of idle curiosity. Google has decided a synthesized answer is the right first screen for a person comparing options, and the data says it made that decision across the board in about a year.

For an owner, the practical translation is simple. The click that used to go to a ranked page now goes, at least in part, to whichever sources the summary pulls from. Semrush says those sources include a brand's own pages, third-party sites, YouTube videos and review platforms. Your homepage is one possible source out of several, and not always the one Google trusts most on a question like "is [your business] any good."

What this does and doesn't tell a local shop

Be careful here, because a lot of coverage of this study skips the caveats.

The panel was banking, ecommerce and software, on U.S. desktop. A plumber, a groomer or a bakery was not in it. Semrush also reports 15 industry-and-modifier combinations with no per-cell keyword counts, so an individual figure like banking reviews (13.79% to 89.66%) rests on a small base. And the advice in the article about how brands should respond is the authors' interpretation, from a company that sells the tools to do it.

Local searches behave differently. A separate hand-run Whitespark study of "near me" queries found AI Overviews on only about 15% of local-intent searches while a local pack appeared about 93% of the time; we broke that down in why AI Overviews skip most "near me" searches. Both can be true at once. The "near me" search at 11pm is still a map-pack fight. The "best [service] for [situation]" and "[business] reviews" search, the one a careful buyer runs before calling, is the one drifting toward an AI summary.

So take the direction seriously and the exact percentages loosely. Nobody has published the equivalent number for local services, and I'm not going to guess one.

Where a small business can show up in a comparison summary

Four moves hold up whether the real local number turns out to be 40% or 80%.

1. Own the facts a summary would quote. If someone asks which option fits their situation, the summary needs plain statements to lift: who you serve, what you do and don't do, where, and roughly what it involves. Put those in sentences on the page, not buried in a PDF or implied by a stock photo. The same habit is why question-style headings keep getting cited.

2. Treat review platforms as part of your site. The study names review platforms as a common source. Recent, specific reviews on the profiles you already have count toward what a summary says about you; old reviews lose weight, so a steady trickle beats a burst. Ask every customer, in the same boring way, every time.

3. Get described consistently elsewhere. A summary that sees your name, services and location stated the same way on your site, your Business Profile and a couple of directories is less likely to garble you. A summary built from mismatched sources is how businesses end up with the wrong hours or the wrong service list. This is a large part of what brand mentions across the web are for.

4. Check what the summary says before you decide it's a problem. Search "best [your service] in [your region]" and "[your business] reviews" in a signed-out window and read the answer. Note who is cited and whether the details are right. Ten minutes of reading tells you more than any dashboard will, and it is the first step in checking what AI says about your business.

Reactions that won't help

• Rewriting your whole site because a percentage went up. The study is about a different kind of business. Change what is demonstrably thin, not everything.
• Chasing a "rank for the AI Overview" trick. Google has published no switch for this. Anyone selling one is selling a theory.
• Buying reviews or trading them. The Federal Trade Commission's fake-review rule is aimed squarely at that; see what the rule means for small businesses.
• Ignoring your classic rankings. Pages that rank are still a main source of what summaries cite. The two jobs overlap, and the pillar guide on Answer Engine Optimization (AEO) covers where they split.
• Stuffing "best in town" claims onto the homepage. Summaries lean on third-party descriptions of you. A self-awarded superlative reads as what it is.

FAQ

Do AI Overviews show up on "reviews" searches now?

In Semrush's panel of 400 commercial keywords, yes: about 90% of "reviews" searches in August 2026 showed one, up from roughly 30% a year earlier. That panel covered banking, ecommerce and software on U.S. desktop, so treat it as a direction for local businesses, not a measurement of them.

Does this mean nobody clicks through to websites anymore?

The study measured how often an AI Overview appeared, not how often people clicked. Clicks are a separate question; see our look at searches that end without a click. A summary appearing does not mean your visit count drops by the same amount.

Does this apply to local searches like "plumber near me"?

Mostly not yet. A manual study of local-intent searches found a local pack far more often than an AI Overview. The buying-stage searches around a local business, such as "best" and "reviews," are where the Semrush trend is more likely to apply, though no one has published local numbers for it.

How do I find out if my business is mentioned in these summaries?

Search the comparison and review phrases your customers would use, in a signed-out browser window, and read the answer and its cited sources. Repeat monthly and write down what changed. Google Search Console's generative AI report can show impressions inside those features for your own site.

Should I pay for an AI-visibility tracking tool?

Not before doing the manual check above. Tools are useful once you know which queries matter and want them tracked on a schedule. Keep in mind that this study came from a vendor of exactly those tools, which is a reason to read it carefully, not to dismiss it.

What is the single cheapest fix?

Make sure your Business Profile, your site and your top review platform all say the same thing about what you do and where. It costs an afternoon, and every summary that mentions you draws on those sources.

If you'd rather someone else watch this and keep those sources lined up month to month, that is what our Search Engine Optimization and AEO work covers.