BlogAI search

How to show up in Google's AI Overviews and AI Mode

Benjamin Libor10 min read

To show up in Google's AI Overviews and AI Mode, rank for the questions behind the query, answer the question in the first lines of the page, keep the facts current, and be present on the sources Google already trusts for the topic. The links in an AI Overview come largely from pages Google already ranks for the query and its sub-questions, but not only from the top three, so a clear, current page on a solid site can be cited without holding position 1.

In short

  • AI Overviews are a generated summary above the classic results; AI Mode is a separate, conversational results page. Both cite a handful of links drawn from Google's index.
  • They appear mostly on informational, comparative and how-to queries, and rarely on plain brand or navigational searches.
  • The cited pages tend to be ones that already rank for the query or for the sub-questions Google breaks it into, so ranking still matters, but position 1 is not required.
  • The playbook is ordinary SEO plus a sharper page: the answer first, headings that match sub-questions, dated facts, structured data where it fits, and presence on the third-party sources for the topic.
  • Search Console counts AI Overview and AI Mode impressions and clicks, but at the time of writing it does not separate them from the rest, so appearance has to be checked directly and on a schedule.

What AI Overviews and AI Mode are

AI Overviews are the AI-written summary that Google shows at the top of some results pages, with a set of links beside or beneath it. The classic ten links remain below. Google launched them broadly in the United States in 2024 and has since extended them to many more countries and languages. For a brand, it is a sentence that either names you or does not, and a link that either points to you or does not.

AI Mode is a separate way of searching: a tab where the whole results page is a conversational answer, with follow-up questions and links woven into the reply. Google has described AI Mode as using query fan-out: it issues several related searches for one question and writes the answer from all of them. AI Overviews rely on the same idea, so think of the two as one system at two depths rather than two products.

How they differ from the classic results page:

Classic resultsAI Overviews and AI Mode
What the reader seesTen links with titles and snippetsOne written answer, with a few links
How many pages get exposureTen per page, more with featuresTypically a handful, chosen from across the sub-questions
What decides inclusionRanking for the queryRanking for the query or its sub-questions, plus having a passage that answers
What the brand getsA click, sometimesA mention in the text, a link, or neither
Where the traffic goesTo the ranked pagesMostly stays on the page; clicks go to the cited links

Which queries trigger them

Google does not publish a list, but the pattern is consistent enough to plan around. AI Overviews appear most on:

  • Informational questions: "what is a webhook", "how does single sign-on work".
  • Comparative questions: "Fernwick vs Slotwise", "best scheduling tool for agencies".
  • How-to questions: "how to reduce no-shows for client calls", "how to connect a calendar to HubSpot".

They appear much less often on navigational searches (your brand name typed to reach your site), on purely transactional ones, and on some sensitive topics. AI Mode answers almost anything, because the user has chosen it, but its sources follow the same logic.

The useful consequence: the queries that trigger an Overview are the ones your buyers ask before they know you. If you have a prompt list for measuring AI visibility, run it in Google and note which prompts show an Overview. Those are the ones where a classic ranking alone no longer guarantees exposure.

What can be observed, and what Google's own documentation says, add up to this:

Largely pages Google already ranks well. The cited links for a query overlap heavily with the pages ranking on the first page for that query. Google's documentation for site owners says there are no special requirements for appearing in AI features beyond the ordinary ones: the page must be indexed, eligible to appear in search, and allowed to show a snippet.

But not only the top results. Because of fan-out, a page that ranks for one of the sub-questions can be cited even if it is nowhere for the headline query. For "best scheduling tool for agencies", the Overview may cite a comparison list that ranks for that phrase, a pricing page that ranks for "scheduling software pricing", and a help article that ranks for "HubSpot calendar integration". The second and third did not need to rank for the headline query at all.

Pages with a passage that answers. Among the pages that rank, the ones cited have a sentence or a short section that answers a sub-question directly. A page that circles the topic without stating the answer is fetched and passed over.

Snippet controls apply. Google's documentation says the standard controls, nosnippet, data-nosnippet and max-snippet, limit what from a page can be shown in AI Overviews and AI Mode. A site that set a short max-snippet years ago may have quietly made itself unquotable. Google also says the Google-Extended robots token, which covers Gemini training, does not affect inclusion in Search or AI Overviews.

The playbook

None of this is exotic. It is SEO with a sharper page and a wider definition of "ranking".

1. Rank for the underlying queries

Take each buying question and write out the searches it fans out into: the "best" version, the "with [platform]" version, the "vs" version, the "how to" version, the pricing version. Those are your targets, and many are long-tail and uncontested. A page that ranks fifth for "scheduling software HubSpot integration" is more likely to be cited for the agency question than a page that ranks fifteenth for the agency question itself.

2. Answer the question in the first lines

Put the answer in the first paragraph, in plain words, before any context or storytelling. A pricing page says what it costs and for whom in its first lines; an integration page says what connects to what. The model is looking for a passage it can lift; give it one at the top.

3. Use headings that mirror the sub-questions

Each ## heading should read like one of the searches from step 1. "How the HubSpot integration works", "What it costs", "Who it is for". A reader scanning the page finds their question; so does a system looking for the passage that answers it. Writing content that AI answers quote goes deeper on the page structure.

4. Keep facts current, with dates

Overviews lean towards pages that look maintained. State the date a fact was last checked, update the pricing page when the price changes, and remove or redirect pages that describe features you no longer have. A stale page is the source of the wrong facts that then get repeated in answers, which is a separate problem to clean up (finding and fixing wrong facts in AI answers).

5. Add structured data where it matches

Organization, Product, FAQPage, HowTo and Article markup do not buy inclusion, but they make the page's facts unambiguous to a system that reads a lot of pages fast. Add it only where it matches the visible content; markup that says one thing while the page says another does harm.

6. Be present on the sources Google trusts for the topic

For comparative queries, the Overview often cites a review platform, a comparison list and a forum thread alongside vendor pages. Those sources already rank, and they will be cited whether or not you are on them. Being accurately listed there is the fastest way into Overviews for "best" and "vs" questions, because you ride a page that is already trusted. Source gap analysis is the method for finding which sources those are.

7. Check nothing is blocking the snippet

Look at the page's meta robots tags and the site's robots.txt. Remove nosnippet and short max-snippet values unless there is a legal reason, and make sure the relevant pages are in the sitemap and indexed.

How to check whether you appear

Searching once from your desk is the most common mistake. Overviews vary by country, by session and by day, and AI Mode varies even more because the fan-out is generated each time.

  • Search in the right country. Set the browser's location to where your buyers are, or use a tool that can specify the country. An Overview that shows in the United States may not show in Germany for the same query, and the sources differ.
  • Use a clean session. Signed out, no history, a fresh window.
  • Repeat. Run the same query several times across a week before you call a result. Keep a simple log: query, country, date, Overview shown or not, brands named, links cited.
  • Try the phrasings. "best scheduling tool for agencies" and "which scheduling tool should an agency use" can trigger different Overviews with different sources.
  • Automate it. A tool that runs the queries on a schedule from a fixed country and stores every answer turns an anecdote into a series. It is also the only practical way to watch AI Mode, where there is no stable page to look at.

What to measure in Search Console

At the time of writing, Google's documentation says impressions and clicks from AI Overviews are counted in the Search Console performance report, that AI Mode is counted there too under the Web search type, and that neither is reported separately. So Search Console will tell you that a query sent you traffic, but not whether it came from an Overview link, a classic blue link or an AI Mode reply.

What you can still do with it:

  • Compare query groups. Keep the list of queries where you have confirmed an Overview appears, and watch their impressions, clicks and click-through rate against the queries where none appears. A rising average position with falling click-through on the first group is the usual signature of an Overview taking the click.
  • Watch pages, not only queries. A page cited in an Overview often gains impressions on queries it never ranked for, because the fan-out surfaced it. Unexplained new queries in a page's report are worth a direct check.
  • Date the changes. Note when you rewrote a page's opening or added a section, and read the data from that date forward.

GA4 does not separate it either: a click from an Overview or from AI Mode arrives as ordinary organic traffic from google.com. Search Console measures the clicks; a direct, scheduled check measures the appearance; you need both.

Questions people ask

Can I opt out of AI Overviews but stay in Google search?

Not cleanly. Google's documentation says the snippet controls, nosnippet, data-nosnippet and max-snippet, limit what can be shown from a page in AI Overviews and AI Mode, but those same controls also limit the classic snippet. There is no switch that removes a page from AI features alone while leaving its ordinary listing unchanged.

Does blocking Google-Extended keep me out of AI Overviews?

No. According to Google's documentation, Google-Extended controls whether content is used to train and ground Gemini models, and it does not affect whether a page appears in Google Search, including AI Overviews. Inclusion follows Googlebot and the normal indexing and snippet rules.

Do AI Overviews appear on searches for my brand name?

Rarely for the bare brand name, which Google treats as navigational. They appear more often on brand-plus-question searches such as "Fernwick pricing" or "Fernwick vs Slotwise", where the answer is written from your pages and from third-party pages about you. Check those variants, because they are where a wrong fact about your company tends to surface.

Why do I rank first but not appear in the AI Overview?

Usually because the Overview was written from the sub-questions and other pages answered those more directly, or because your page buries the answer. Check that the first lines state the answer and that no snippet restriction applies. Ranking first for the headline query earns a fetch, not a citation.

Written by Benjamin Libor, founder, echo.

Something wrong or out of date? Write to hello@echo-aeo.com and we'll fix it.

See what AI says about your brand. Free, in about a minute.

The free AI score checks how ready your website is for AI assistants. Echo itself tracks your customers' questions across the assistants every day and turns the answers into actions.