How to Appear in Google AI Overviews
The short answer
AI Overviews are generated from pages that already rank well for the query and its related sub-questions, then summarised into one answer with a small set of linked sources. The fastest route in is fixing pages that already sit in the top ten but answer the question too slowly.
Table of contents
What AI Overviews actually are
An AI Overview is the summarised answer Google generates at the top of a results page, with a handful of source links attached. It sits above the organic listings you have spent years earning, and it answers the question before anybody scrolls.
The mechanism underneath is worth understanding. Google breaks your query into several related sub-questions, runs them, collects pages that rank well across that set, and synthesises an answer from the overlap. The sources it links are drawn from pages already performing in conventional search.
That last point is the good news. Unlike a standalone AI assistant, this is not a separate universe with its own trust set. It is built on the index you already compete in. If you rank, you are eligible. If you do not, you are not.
Which pages get picked as sources
Across the queries I watch for clients, the pages that get pulled in share a fairly consistent profile.
- They already rank in the top ten for the query or one of its sub-questions. Eligibility starts with conventional visibility.
- They answer within the first hundred words. The answer is stated, then explained. Not built up to across six paragraphs of context.
- They cover one question properly rather than fifteen questions loosely. A page about one thing is easier to extract cleanly.
- They contain extractable structures. Real tables, real ordered lists, clear subheadings phrased as questions.
- They carry visible expertise signals. A named author, a real bio, a date, evidence that a person with relevant experience wrote it.
- They are current where currency matters. Anything with pricing, tools or regulations attached gets checked hard on recency.
Notice what is missing from that list. Nothing about word count. Nothing about keyword density. The unit of selection is a clean, attributable answer to a specific question.
The work, in the order that pays
1. Find pages already ranking that answer too slowly
Open Search Console, filter to queries in positions four to fifteen with meaningful impressions, and read the landing pages as if you were in a hurry. If the answer is not visible in the first screen, you have found your first job. This is the highest return work available, because eligibility is already established. A free SEO audit will surface these pages if you would rather not comb Search Console yourself.
2. Rewrite the opening, not the whole page
Put a direct two to three sentence answer immediately under the H1. Then continue with the depth you already had. You are not shortening the article. You are moving the conclusion to the front, which is the opposite of how most people were taught to write.
3. Convert prose comparisons into tables
If a section compares options, prices, or approaches, it belongs in a real HTML table with proper header cells. A div grid styled to look like a table loses the row and column relationship that makes it extractable. This is small work with a disproportionate payoff.
4. Add the sub-questions as subheadings
AI Overviews are assembled from sub-questions. Take the People Also Ask box for your target query, phrase each one as an H2 or H3, and answer it in a short self-contained paragraph directly underneath. You are matching the shape of the machine’s query fan-out.
5. Mark it up
Article, FAQPage and BreadcrumbList markup makes attribution and structure explicit rather than inferred. It is not a magic entry ticket, but it removes ambiguity. I covered which types are worth the effort in schema markup for AI search.
6. Then, and only then, worry about authority
If your pages are not in the top ten for anything relevant, steps one through five have nothing to work on. That is a different problem, and it is solved off the website. The same logic applies across every answer engine, which I unpacked in how ChatGPT decides which sources to cite.
How to tell whether it is working
There is no report for this. Search Console does not separate AI Overview impressions, so you are reading indirect signals.
| Signal | Where | What it suggests |
|---|---|---|
| Impressions steady, clicks falling | Search Console, query level | An Overview is likely absorbing the click on that query |
| Manual spot checks | Incognito, logged out, Malaysian location | The only direct way to see whether you are cited |
| Branded search volume rising | Search Console, brand queries | People are seeing you upstream and searching you by name later |
| Direct and organic sessions of higher quality | GA4, by landing page | Fewer visitors, further along in their decision |
Pick twenty queries that matter commercially, check them by hand on the same day each month, and record what you find in a sheet. That is unglamorous and it is the only measurement that survives contact with reality. There is more on building that habit in how to track AI search visibility, and a comparison of where each engine sends real traffic in which AI search engines actually send business traffic.
Frequently asked questions
How do I get my website featured in Google AI Overviews?
Rank in the top ten for the query first, then rewrite the page so a direct answer appears within the first hundred words. Add the related People Also Ask questions as subheadings with short self-contained answers, use real HTML tables for comparisons, and mark the page up with Article and FAQPage schema.
Can I opt out of AI Overviews?
Partly. The nosnippet and max-snippet robots directives limit how much of your content can be shown, and applying them will usually keep you out of Overviews for those pages. It also suppresses regular snippets, which normally costs more traffic than it saves. Most sites should not do this.
Why did my traffic drop when rankings stayed the same?
This is the signature pattern of an AI Overview appearing on your queries. Impressions hold steady because you still rank, but clicks fall because the answer is delivered above the results. Check Search Console query by query and compare click-through rate before and after the drop began.
Do AI Overviews use the same sources as ChatGPT?
Not usually. AI Overviews draw from Google's own index and favour pages already ranking there. ChatGPT runs its own retrieval, which weights third-party corroboration and community sources more heavily. Overlap exists, but a page cited in one is not automatically cited in the other.
Does schema markup guarantee inclusion in AI Overviews?
No. Schema makes your structure and authorship explicit rather than inferred, which helps a system extract and attribute your content correctly. It does not create eligibility. A page with perfect markup that ranks on page four will not be summarised, because it was never retrieved.
How long does it take to appear in an AI Overview after making changes?
Once the page is recrawled and reindexed, usually a few weeks. The changes that work fastest are on pages already ranking in the top ten, because eligibility is established and only the answer format is being fixed. Pages needing authority work take considerably longer.
Written by Robin Ooi
Robin is a Malaysian AI SEO and reputation specialist with more than fifteen years in search. He is the Amazon bestselling author of Your SEO Sucks! and works with MNC and listed company clients across Kuala Lumpur, Penang, Johor and Singapore. Read more about Robin Ooi.
