How to rank in Google AI Overviews and get your pages cited

Last updated August 25, 2026 · ~8 min read

The top of Google changed — quietly, then all at once

For many shopping and research queries, the first thing a buyer now sees on Google is not a list of links. It is an AI Overview: a short, AI-written answer, built from web pages and product data, with a handful of sources linked beside it. The classic results still sit below — but the summary, and the two or three products it happens to name, get the first look and a large share of the attention.

The good news: an AI Overview is grounded in Google's live index, not a frozen training set. If your page is crawlable, relevant, and clearly answers the question, you can earn a place in that summary as fast as Google re-crawls and re-ranks you. The whole game is getting into the retrieved-and-linked set for the questions your buyers actually type.

How an AI Overview decides what to show and link

Behind each Overview, Google is doing three things in sequence, and you can influence all three:

• Retrieve — it pulls relevant, well-ranked pages from the Search index and product data from the Shopping Graph. If you are not indexed or don't rank for the query, you are not even a candidate.
• Ground — it reads those candidates and writes a summary anchored to what they actually say. Facts hidden in JavaScript, images, or behind walls don't get read, so they can't ground anything.
• Link — it surfaces a small set of supporting sources next to the answer. Pages that answer the question directly, with structured, corroborated facts, are the ones it trusts enough to name.

In our October 2026 study of 60 shopping questions, 64% of the sites AI Overviews cited were outside the organic top 10 for that exact question, and brand sites appeared in about half of the overviews. See who AI Overviews cite for shopping questions.

What gets your product into an AI Overview

• Crawlable & indexable. Allow Googlebot, serve key facts in real server-rendered HTML, and keep the page indexable. If it isn't in the index, it can't be in the Overview.
• Relevant enough to surface. The page still has to be a strong result for the buyer's question — AI Overviews draw from pages Google already considers relevant, so ranking and topical fit remain the entry ticket.
• Answers the question directly. Lead with the answer a buyer is searching for ("best for wide feet", "works with X", "under $150"), not a paragraph of brand story. Direct pages get grounded; buried ones get skipped.
• Quotable facts, not adjectives. "IP68, 2-year warranty, 480g, $129" is groundable; "premium, best-in-class" is not. Google summarizes specifics.
• Product structured data. schema.org Product markup — brand, canonical URL, price and availability, genuine ratings and reviews — makes your facts machine-readable so the summary uses your price and spec, not a marketplace's.
• In the Shopping Graph. A clean Google Merchant Center feed puts your items in the product data Overviews draw on, so you're a first-party source rather than a mention on someone else's page.
• Reviews & corroboration. Independent evidence — genuine reviews, comparisons, mentions — that agrees with your claims raises Google's confidence to ground and link you. This is the E-E-A-T layer.
• Freshness. Current, dated content signals reliability; stale pages get passed over for newer-looking sources.

The step-by-step playbook

1. Measure where you stand. Ask Google's AI — and the Gemini model behind it — the real buyer questions in your category, and see whether it names your brand, model, or URL today. That is your baseline.
2. Clear the crawl path. Confirm Googlebot is allowed, the page is indexable and server-rendered, and the product is in your Merchant Center feed and Shopping Graph.
3. Answer the question directly. Restructure the page to lead with the answer a buyer wants, then support it with explicit specs, price, use cases, and an honest "who it's not for".
4. Make the facts machine-readable. Add clean Product JSON-LD (price, availability, rating), a canonical URL, and turn vague claims into quotable numbers so Google grounds on the right facts in one pass.
5. Earn corroboration. Build the independent reviews and mentions Google can also find; agreement across sources turns a claim into something safe to summarize.
6. Re-measure and iterate. Track whether the Overview and Gemini name you over time, and keep improving the pages and questions that move it.

AI Overviews vs classic Google SEO

Because AI Overviews are grounded in Search, good SEO helps you get found — but being named in the answer takes the AEO layer on top. See the full AEO vs SEO · SEO suite vs AEO platform, and the engine-specific playbooks for ChatGPT and Perplexity.

Common mistakes that keep you out of the Overview

• Blocking or hiding the facts. Disallowing Googlebot, or rendering price and specs only in JavaScript, removes you from the candidate set entirely.
• Confusing Google-Extended with Googlebot. Overviews use the Search index — blocking Google-Extended does not opt you out of them, and does not help you either.
• No direct answer. If the buyer's question isn't answered plainly on the page, a more direct competitor gets grounded and linked instead.
• Adjectives over facts. Unquotable marketing copy gives the summary nothing specific to lift.
• Missing structured data & feed. No Product markup and no Merchant Center feed means Google grounds on a marketplace listing rather than your page.
• Guessing instead of measuring. Assuming you're in the Overview without probing is the most expensive mistake — you can't fix what you don't measure.

A note on Googlebot, Google-Extended, and control

It's worth being precise, because the robots settings trip people up. AI Overviews are generated inside Google Search from the normal Search index, so the access that matters is ordinary Googlebot — keep it allowed and the page indexable. Google-Extended is a separate control for whether your content trains and grounds Gemini and Vertex AI products; changing it does not add or remove you from Search or from Overviews. And there is no button that forces inclusion: you make yourself eligible by being crawlable, relevant, structured, and corroborated — then measure the result.

Can you measure this?

Not directly. Google offers no official way to query AI Overviews or report your inclusion rate in them; third-party trackers can only sample. What you can check is your own page — crawlable, indexable, answering the question, with Product structured data that matches the visible copy — and what the Gemini API answers to the same buyer questions. The Gemini API is a different surface from AI Overviews and from the Gemini app; treat it as one observation, not a stand-in. Searching for your own query in Google and keeping a dated screenshot is the honest way to watch Overviews themselves.

From AI answers to sales

Being named in an AI Overview is one step, not the result. What matters is whether the buyer’s question was answered well enough to move them toward a purchase, and what happens next. Today you can follow AI answer checks, AI referral visits in Google Analytics 4, and the leads and opportunities in Hasmord’s CRM, each with the buyer questions behind it. The questions buyers raise in sales conversations point back to the product information: add evidence, sharpen the claim, then see how the market and sales respond.

Hasmord is the AEO-native sales platform. Buyers ask questions across AI assistants, search, and sales conversations. Hasmord uncovers where your answers about your products and services fall short, turns evidence into better answers, finds potential customers for you, and connects every lead to your sales process. Hasmord checks your product information and runs answer checks through the Gemini API, among other model APIs. The Gemini API is not AI Overviews, and Hasmord does not report whether an Overview names you.

Guides by engine: get cited by ChatGPT · ChatGPT shopping product data · Perplexity shopping · rank in Perplexity answers · appear in Gemini · what each engine cites (study)

Frequently asked questions

How do Google AI Overviews decide which products to show?

An Overview is generated from Google's Search index and Shopping Graph: it retrieves relevant, well-ranked pages and product data, grounds an AI summary on them, and links the supporting sources. To be one, your page must be crawlable and indexed, relevant for the query, and clear and corroborated enough for Google to ground a claim on it.

Do I need to allow Google-Extended to appear in AI Overviews?

No. Overviews in Search are built from the regular index, so ordinary Googlebot access is what counts — keep Googlebot allowed and the page indexable. Google-Extended only controls whether your content trains and grounds Gemini and Vertex AI; it doesn't add or remove you from Search or Overviews.

How is appearing in AI Overviews different from ranking #1 in Google?

Ranking makes you a candidate, but the Overview rewards a page that answers the question directly and quotably, backs claims with facts and reviews, and is corroborated — then links you as a source. You can rank on page one and be left out, or be named from a spot that isn't #1.

What structured data helps products appear in AI Overviews?

schema.org Product markup — brand, canonical URL, price and availability, and genuine ratings and reviews — plus a clean Google Merchant Center feed so your items sit in the Shopping Graph. It doesn't force inclusion, but it lets Google ground on your real price and spec instead of a marketplace's.

Can I measure whether AI Overviews mention my product?

Not directly: Google offers no official way to query AI Overviews or report your inclusion rate in them, and third-party trackers can only sample. Hasmord can check your product information and run answer checks through the Gemini API (and the OpenAI, Perplexity and Anthropic APIs), but the Gemini API is not AI Overviews. For Overviews themselves, run your own buyer questions in Google and keep dated records.

Why isn't my product showing up in Google AI Overviews?

Usually because the page or its facts are blocked from Googlebot or hidden in JavaScript, it doesn't rank for the question, it uses adjectives instead of quotable specs and price, it lacks Product structured data and a feed, or reviews and corroboration are thin.

Explore more guides in our AEO resource library, or see how Hasmord connects AEO to sales on the Hasmord homepage.