AI Overviews answer the buyer's question before any link gets clicked. Learn how Google picks cited pages, and the six moves that earn a local business a place.
By Heather Laskin · Published August 27, 2026
A growing share of Google searches now open with a written answer at the top of the page. The answer names two or three businesses and cites a handful of pages as sources. Every other result sits underneath. Learning how to rank in AI Overviews means learning how to become one of those cited pages, because the businesses named in the answer get the call and the businesses below it often get nothing.
The encouraging part: Google's own documentation says AI Overviews run on the same core ranking systems as classic search. No secret markup forces your way in, and no new trick replaces the fundamentals. The work is a sharpened version of good SEO, aimed at the way a written answer gets assembled.
Ranking in AI Overviews means getting your pages cited inside the written summary Google places above its results. Google's documentation says the same core ranking systems choose those pages, so the path is: rank well for the question in classic search, answer the question directly in the first hundred words, structure the page so a machine lifts a clean passage, keep structured data in sync with visible content, earn third-party mentions, and keep your business facts accurate everywhere. Then track whether you appear, because inclusion shifts from week to week.
An AI Overview is a written summary Google generates for a search and displays above the list of links. Instead of ten blue links waiting for a click, the buyer reads a paragraph or two, sees a few businesses named, and notices a small set of cited pages off to the side.
Two consequences follow for a local business.
The buyer's shortlist forms inside the answer. A homeowner asking about a repair cost reads the summary, notes the two companies named, and calls one of them. The company ranking third in the classic results never enters the decision.
The cited pages become the trusted sources. When your page is cited, Google is telling the buyer your site helped write the answer. When a competitor's page, a directory, or a review site is cited instead, those pages shape what the buyer believes about you.
This is why ranking in AI Overviews is now a separate goal from ranking in the classic list, and why the AI visibility gap most businesses do not know exists keeps widening for owners who only watch classic positions.
Google has published guidance on succeeding in AI search, and the headline is boring on purpose. The same core ranking systems behind classic search choose the content for AI features. A page must be indexed and eligible to appear in search before any AI surface will cite it.
Within the eligible pool, the selection favors pages with five traits.
| What Google looks for | What it means on your site |
|---|---|
| Unique, people-first content | Information a buyer gets from you and nowhere else: your prices, your process, your service area, your photos |
| Direct answers | A passage near the top of the page resolving the exact question in plain language |
| Clear structure | Headings phrased as questions, short paragraphs, lists and tables a machine lifts cleanly |
| Matching structured data | Schema markup describing the same facts the visible page shows, never more |
| Trustworthy page experience | Fast load, works on a phone, no intrusive interstitials |
One more mechanic matters. An AI Overview answers the main question plus several sub-questions the buyer would ask next. Each sub-answer draws from the page with the cleanest passage on that point. Your page does not need to win the whole query. It needs to win one passage.
This is also why a first-place classic ranking does not guarantee a citation. The page in position five with a direct answer block beats the page in position one with the answer buried in paragraph nine. If you rank organically and still never appear, the gap is almost always structure, not authority.
Everything below assumes your page reaches at least the first page of classic results for the question you want to own. AI Overviews cite pages outside the top positions, but they cite them for sub-questions where the passage quality wins. The bulk of citations still go to pages already ranking well.
For a local business, classic ranking work means the same short list it always has: a Google Business Profile filled out completely, a consistent name, address, and phone across every directory, service pages that target the specific terms buyers search with one intent per page, and reviews arriving steadily.
If your page sits on page three for its target question, fix the classic ranking before anything else here pays off. The AI visibility checklist walks the order of operations.
The single highest-impact edit on any page is the opening. State the question's answer directly, in one or two sentences, before any background or story.
A plumber's page about water heater replacement should open with the price range and the main factor moving it, not with a paragraph about how important hot water is. A dentist's page about implants should open with the direct answer to the cost or timeline question the page targets.
Answer engines extract passages. A passage resolves a question on its own, without pronouns pointing back to earlier text, without throat-clearing, and without a sales line. Write the opening so it reads correctly when lifted out and shown alone, because lifted out and shown alone is exactly where it ends up.
Formatting is ranking criteria now. The patterns earning citations:
Notice what this rules out. Long narrative intros. Clever headings withholding the topic. Key facts spoken only in a video or baked into an image. If the information is not in selectable text under a descriptive heading, the answer engine never sees it.
This formatting work overlaps heavily with answer engine optimization, and the what is AEO guide covers the discipline in full.
Google's AI search guidance calls this out by name: structured data must match the visible content on the page. Schema describing facts the page does not show is a trust violation, and pages get filtered for it.
For a local business the priority schema types are LocalBusiness with your name, address, phone, hours, and service area, Service schema on each service page, FAQPage on genuine question-and-answer sections, and Article on editorial content like this one.
Run every page through Google's Rich Results Test after adding markup. The test confirms the schema parses and shows what Google extracted. What it extracts is what the AI features read.
Your website tells Google what you say about yourself. The rest of the web tells Google whether to believe it. AI Overviews lean on this distinction hard, because the summary format has room for only a few names and the wrong recommendation costs Google the user's trust.
Third-party signals feeding the choice include reviews on Google and industry platforms, listings in the directories your trade respects, mentions in local press and community sites, and citations on the pages answer engines already quote for your category.
The practical move: find the pages AI Overviews cite for your target questions today. Search your five highest-value buyer questions and note which directories, review sites, and publications appear. Get your business present and accurate on exactly those sources. Presence on a cited source puts your name inside the evidence the next answer gets written from.
For the broader picture of how AI engines pick businesses to name, see why AI assistants skip your business.
An AI Overview states facts with confidence. Your hours, your services, your prices, your service area. When the answer is wrong, the buyer acts on the wrong information: shows up after closing, calls expecting a service you dropped, or prices you out of contention before the first conversation.
Wrong facts in AI answers trace back to a source somewhere: an old directory listing, a stale review, a forgotten page on your own site. Audit the facts quarterly. Search your business name, read what the answers say, and correct the sources feeding any error.
Freshness matters the same way. Pages updated with current information earn citations over identical pages last touched years ago. A date-stamped price guide refreshed twice a year outperforms a timeless page saying nothing specific at all.
Three tactics circulate as AI Overview shortcuts. None hold up.
The first is keyword stuffing. Repeating the target question fifteen times does not make the page more citable. It makes the prose worse, and passage extraction favors pages a person would finish reading.
The second is schema without content. FAQ markup on questions the page never visibly answers violates Google's matching rule and risks a manual action, not a citation.
The third is mass-produced AI text at scale. Google's guidance asks for unique, non-commodity content, and a page indistinguishable from a thousand others gives the answer engine no reason to cite yours over any of them. The differentiator the selection rewards is first-party specifics: your real prices, your real photos, your real process.
Classic rank tracking will not show AI Overview presence. You need a separate record: your list of buyer questions, checked on a schedule, noting whether an AI Overview appeared, whether your business or page was named, and who was named instead.
Run the check monthly at minimum, weekly on your two highest-value questions. Note the cited sources each time, since a new citation appearing for a competitor tells you where the next round of work goes. The full setup is in the AI overview tracking guide, and the wider measurement framework sits inside AI search for local and small businesses.
Doing the first sweep by hand takes an afternoon. If you want the baseline built for you across a full question set, run a free audit and use the result as your starting row.