How to optimise your website for Google AI Overviews in 2026

How Google picks the sources in an AI Overview

If you want to know how to appear in AI Overviews, start with how the links get chosen. Google names two processes in its guide.

The first is retrieval-augmented generation, which Google also calls grounding. Before the model writes its answer, it pulls current information from Google's search index to back up what it says. The pages it pulls from become the links you see in the Overview.

The second is query fan-out. The model runs a batch of related searches in the background to cover the topic. Google's own example is someone asking how to fix a lawn full of weeds, which also triggers searches like "best herbicides for lawns", "remove weeds without chemicals" and "how to prevent weeds in lawn".

Fan-out is where smaller sites get a look in. Your page can be cited for one of those follow-up questions even if it doesn't rank for the original search. A Newcastle landscaper probably won't outrank the big gardening sites for the broad question. A clear, local page on [specific sub-topic, e.g. the weeds that take over Hunter lawns after a wet summer] has a real shot. 

Write what only your business can write

Most of the work in SEO for AI Overviews happens in the content. Google calls the kind it wants "non-commodity" content: pages built on information or experience a reader can't get from fifty other sites. Its example compares "7 Tips for First-Time Homebuyers", which anyone could write, with "How We Waived the Inspection & Saved Money: A Look Inside the Sewer Line", which only one person could.

An AI Overview is a summary. If your page repeats what every other page says, the model has no reason to cite you over anyone else. The original material is usually already sitting in the business. Look for:

• The questions customers ask on the phone or at the counter, in their own words 
• Real prices, timeframes and job details you're happy to share 
• Photos and short videos from actual jobs 
• The mistakes you see customers make, and what they cost to fix 

Then cover the fan-out questions. It's worth checking other areas of the search, like People Also Ask, the follow-up suggestions in AI Mode and the queries in your Search Console performance report. Then, you can group the related sub-topics on one solid page, and put a direct answer in the first sentence or two under each heading before you go into detail.

It's not ideal spinning up a thin page for every variation of a question. This makes your website look like a collection of mass-produced pages. If you use AI tools to help draft, a subject matter expert should ideally check every fact and add the experience the tool doesn't have. 

Get the technical basics right

Google AI Overviews SEO starts with the same checks as any other SEO. Simply put, the play book is the same, and you start with the basics. A page must be indexed and eligible to appear in Search with a snippet, if it is to be shown as a link. Does that automatically guarantee a citation? No, but it avoids disqualification. 

Run your key pages through the URL Inspection tool in Search Console and check that: 

• Googlebot isn't blocked in your robots.txt file 
• The page is indexed, and the canonical URL Google chose is the one you want 
• Your main content is on the page as text, and anything loaded with JavaScript still renders for Google 
• No nosnippet, max-snippet or data-nosnippet tags are hiding content you want shown, since these preview controls also limit what AI features can use 
• Duplicate URLs are consolidated, so Google isn't wasting its crawl on copies 

Google also says you don't need perfect semantic HTML. Clean, logical markup is enough. Spend the time and energy on page speed and mobile layout instead, because the people who click through from an AI Overview still need a page that's easy to use.

Structured data and local listings

You don't need special schema markup to appear in AI Overviews. Google lists "over-focusing on structured data" as one of its myths. Schema is still worth adding for rich results and to describe your business clearly, as long as it matches what's visible on the 

page.

SEO.com's 2026 AI Overview ranking factors study rates structured data as a "likely" factor with medium impact. Google hasn't confirmed that, so treat it as good housekeeping and keep your expectations modest.

Local listings is worth investing your time. Google points businesses to Google Business Profile, and online stores to Merchant Center product feeds, as ways to get accurate details into its results and AI responses. If you run a cafe in Sydney's south-west or you're a sparky covering the Hunter, check your profile's hours, categories, service areas and photos at least once a quarter. 

AI search myths you can skip

Plenty of agencies now sell AEO (answer engine optimisation) or GEO (generative engine optimisation) as a separate service. Google's view is that AI search engine optimisation for its own results is still SEO. Its guide lists several tactics that don't help:

• llms.txt files: Google says it's fine to have one, but Google Search ignores them, so they won't help or hurt you there. 
• Chunking: Don't chunk your content into tiny pieces just for AI. The length depends on how you want to present the topic, and how you think your users would like to read. 
• Rewriting for AI: Google's systems understand synonyms and related wording, so there's no need to chase every long-tail version of a phrase. 
• Inauthentic mentions: planting your brand name across forums and low-quality sites. Google's ranking systems are built to find quality content and block spam, and its AI features run on those same systems. 
• Tools promising inside data: Google warns that no third-party tool has access to its internal ranking or AI systems. Be wary of anyone who claims otherwise.

On mentions, there's a fair line to draw. Some 2026 guides, including Elementor's, push brands to build a presence on Reddit and Quora, mostly with ChatGPT and Perplexity in mind. Answering questions in forums where your customers already hang out is fine. Manufactured posts are the problem.

How to measure your AI search visibility

Start in Search Console. Google's guide points to its generative AI performance report for tracking how your content shows up in AI features across Search and Discover.

Once a month, look at: 

• Which pages earn impressions and clicks from AI features 
• Whether searches for your business name are climbing, since people who see you in an AI answer may look you up later 
• Enquiries and sales from organic search, which is the number your business runs on 
• A manual spot check: You can do a manual search on Google by making queries about your top 10 customer questions in an incognito or private window and note where you're cited 

Google is also preparing Search for AI agents that complete tasks for people, such as booking a table or comparing products. If that suits your business, read its guidance on agent-friendly websites once the basics above are done.

Where to start this month

Everything can't be done all at once, so start slow. You can start the first month by doing the following:

1. Run your 10 most important pages through URL Inspection and fix anything that isn't indexed or is blocked from showing a snippet. 
2. Find out 20 questions customers ask you most, then check which ones your site actually answers. 
3. Work on your weakest service page by updating content with real prices, timeframes, photos and answers to those questions. 
4. Update your Google Business Profile. 
5. Do your first spot check, so you've got a baseline to measure against.

If you'd rather hand it over, this is the kind of digital marketing services we do every week for businesses across Sydney and Newcastle, from local cafes and tradies to established brands. Give the Stroberri team a call on 02 4032 5661 or email solutions@stroberri.com.au, and we'll show you where your site stands in AI search today.