Guides
How to Rank in Google AI Overviews (and AI Mode)
A step-by-step method for getting your pages cited inside Google AI Overviews and AI Mode. Which queries trigger them, what the cited pages have in common, and how to restructure yours.
To be cited in a Google AI Overview, a page needs to rank in the top results for the query, answer the question directly in an early passage, be clearly structured with headings, lists and tables, carry the right schema, and come from a recognised entity. Overviews mostly quote pages that do all of those at once. This guide walks through each requirement and how to meet it.
What an AI Overview is, and where it appears
An AI Overview is the AI-generated answer Google shows above the organic results for many searches. It is assembled from several web pages, and each part of the answer is attributed to the page it came from. AI Mode is the conversational version, where the whole results page becomes a chat that asks and answers follow-up questions, again citing pages.
Overviews appear most often for:
- Informational questions: "how long does a TPO roof last".
- Comparisons: "TPO versus EPDM roofing".
- Process questions: "how to choose a commercial roofer".
- Local recommendation questions, increasingly: "best commercial roofer in Atlanta".
They appear less often for pure navigational searches and for some sensitive categories.
Step 1: find the queries that trigger Overviews for your business
Take your target queries and search them in a clean browser. Record three things for each: whether an Overview appears, which pages it cites, and where you rank in the organic results underneath. Queries where an Overview appears and you rank in the top ten but are not cited are your fastest wins. Queries where you do not rank at all need classic ranking work first.
Step 2: understand what cited pages have in common
Look at the cited pages for each query. Most share the same traits:
| Trait | What it looks like on the page |
|---|---|
| Ranks already | Usually in the top ten, often top five, for the query or a close variant |
| Answers immediately | A direct answer in the first 40 to 60 words of the relevant section |
| Question-shaped structure | H2 and H3 headings that match how the query is phrased |
| Extractable formats | Numbered steps, comparison tables, short definition sentences |
| Specific and current | Numbers, ranges, dates, named options; a visible updated date |
| Clear entity | Organization or LocalBusiness schema, an author or company that Google recognises |
Your page needs to match or beat the cited pages on each row.
Step 3: restructure the page passage by passage
Overviews quote passages. Work on the page one passage at a time.
Rewrite the opening. Under the H1, answer the main question in two or three sentences. If the page is about cost, give the range in the first sentence.
Turn subheadings into questions. "Our process" becomes "How does a commercial roof replacement work?" Google matches these against the query and its rewrites.
Answer each heading immediately. The first sentence under every heading should be the answer. Context and caveats come after.
Format for extraction. Steps become numbered lists. Options become tables with the same columns for each option. Definitions become one clear sentence.
Add the numbers. Prices, timelines, dimensions, percentages, with a source or a basis. Overviews prefer passages they can quote as fact.
Cover the follow-ups. AI Mode asks follow-up questions. A page that answers "how much", "how long", "which option" and "what can go wrong" in one place is preferred over four thin pages.
Step 4: add schema and structure signals
Schema does not by itself put a page in an Overview, but it makes the page easier to interpret and attribute:
- Organization or LocalBusiness site-wide, so Google knows the entity.
- Article or Service or Product on the page, matching its purpose.
- FAQPage for a real question and answer section. It no longer produces rich results on commercial sites, but it describes the passages cleanly.
- BreadcrumbList so the page's place in the site is explicit.
The Schema Markup Generator writes all of these.
Step 5: get the page into the ranking range
Overviews draw from pages that rank. If you are on page two, the passage work alone will not get you cited. The classic levers still apply: internal links from your strongest pages, a clean technical foundation, and links and mentions from relevant sources. The free audit flags the technical blockers.
Step 6: keep it current and track it
Overviews favour fresh pages. Review target pages quarterly, update figures, and keep the modified date honest. Then check each target query weekly: does an Overview appear, and are you cited in it? Report citations next to the clicks those queries send, because the citation is what protects the click.
What about lost clicks?
Overviews reduce clicks for many informational queries. That is exactly why citations matter: the cited page keeps a link and the trust of being chosen, while uncited pages lose the click entirely. The practical response is to be the source for the questions that lead to a purchase, and to make sure the cited page has a clear next step for the reader who does click.
Frequently asked questions
How do you get cited in Google AI Overviews?
Rank in the top results for the query, answer it directly in an early passage, structure the page with question-shaped headings, lists and tables, add Organization and page-level schema, and keep the content current. Overviews mostly quote pages that do all of these together.
Do AI Overviews reduce traffic?
For many informational queries, yes. Being cited protects the click, because the citation carries a link and the trust of being chosen. Pages that are not cited lose the click entirely.
Is this different from featured snippet optimisation?
It builds on the same foundations, direct answers and structured pages, but Overviews combine several sources and prefer pages that cover related follow-up questions. Optimise for the combined answer, not a single snippet.
Does this help with Google AI Mode?
Yes. AI Mode uses the same retrieval and rewards the same page patterns, with extra weight on pages that answer follow-up questions in one place.
