How to rank in AI Overviews
Eli McGarvie · · 8 min read Google generates an AI Overview for around half of all search queries. The AI Overview has had an enormous impact on website traffic and search competition. Since Google is answering queries directly, less traffic is being directed to websites on the first page, and for the traffic that does remain, competition is fierce.
Regardless of how we all might feel about AI Overviews, they are here to stay and will continue to show up in search. The only thing your brand can do is to work hard to make sure your site is listed as a source in the overview. This is now the most valuable position on Google, and what’s great for smaller sites is that you can get cited in the AI Overview without ever making it to the first page, and you can get cited multiple times, which drastically increases your visibility.
In this article, I’m going to teach you how to rank in Google’s AI Overviews, with pages built to get cited within your niche or industry.
What Google says
Google will always tell you what it looks for in its documentation, and while it only provides a short section for AI Overviews, it’s worth noting. Google’s advice is to follow its SEO best practices and make sure that your site meets its technical requirements — if your site already appears in Google search, then you most likely meet these requirements. The only thing for you to do is create helpful, reliable, and people-first content.
How to rank in Google’s AI Overviews
The quickest way to get your brand mentioned in AI Overviews is to create a page that targets the search term and answers the query as directly and specifically as possible (exactly how you would with traditional SEO). If you offer new data or a unique insight that backs up your answer, you have a very high chance of getting mentioned in the overview.
Let me give you an example. If the search query is a question about market price, say the user wants to know the average house price in X city, you should write a page that talks about the house prices in X city, providing the most up-to-date figures possible. First, you should crunch the numbers and arrive at a single figure. Then from there, you should provide detailed price breakdowns for different areas of the city, and past and future figures. The more data you can provide, the better. You won’t really know what Google will choose, so give it everything. And the more complete the data, the more opportunities to get cited multiple times. It’s always best to offer proprietary data where you can.
That’s the short answer. Now I’ll give a detailed breakdown of the three fundamental areas of SEO and explain how to approach each of them when you optimise for AI Overviews.
On-page SEO best practices
On-page SEO is when we optimise web pages to rank for search queries. We do this by matching the content on the web page to the specific query. A web page needs to align with Google’s guidelines so that Google knows where and who to show your web pages to. The structure of the page, keywords, metadata, images, internal linking — these all help Google understand your site and web pages.
Each of your web pages should target a keyword, and the keyword should be placed in the title, header, body, meta title and description, and URL. The content on the page should also match the search query; it should provide answers and solutions and match the user intent. If a user is looking for an answer to a question, the page should provide that answer. If the user is looking for a product or service, your page should reflect that.
On-page is very important, but by itself it won’t get you into the AI Overview or even get you ranking; it’s just the requirement to be in the race.
Technical SEO
Technical SEO is all the stuff we do behind the scenes to make sure Google can crawl and index a site and its pages. Most of it has to do with cleaning up a website that hasn’t been designed or managed correctly (most have not!). If a site is not structured correctly, Google will not rank your site the way it deserves to be ranked, which means lower visibility and a loss in traffic.
It’s always important to remember that if your site is confusing to Google, then it will be confusing to users. Google always optimises for the best user experience.
Most big technical issues can be found through an audit, with Claude or ChatGPT, and in Google Search Console. Here’s what you want to look for:
- Make sure Googlebot isn’t blocked.
- The page works, meaning that Google receives an HTTP 200 (success) status code.
- The page is being indexed (Google Search Console).
- Create and submit a sitemap to Google Search Console.
From here, we want to look at site structure and hierarchy:
- URL slugs are short, keyword optimised, and organised in the correct hierarchy (example.com/category/sub-category/product/).
- Pages are interlinked and organised in the correct hierarchy. The homepage should link to service or category pages. Category pages should link to product pages or sub-category pages. Pages that relate to a certain topic should all link to each other.
- Make sure your sitemap is clean and up to date.
A few other things to consider:
- Duplicate pages — Google might deindex pages that target the same keyword. Go through and delete duplicate pages.
- Schema markup — Schema is code on the page which helps search engines understand page content and show rich results.
- Page speed — page speed isn’t a major thing unless you have a really slow site with long waiting times, which will create a negative user experience.
- Spammy backlinks — If you have a ton of spammy backlinks pointing to your site, it can be worth submitting a disavow list to Google to let them know those sites are unrelated to yours. Google for the most part filters spammy links, so it’s only necessary if you think they’re causing issues.
Content quality
Technical and on-page SEO cover the requirements and make you eligible for AI Overviews, but they won’t get you mentioned by themselves. The thing that will move the needle is the actual content on the page.
Google says it values helpful, reliable, people-first content… what do they mean by that?
I’m going to give you a list of tips that you can fold into your strategy that will help create content that meets these requirements, but first, I want to mention that understanding user intent and content types is incredibly important. Let me explain.
If a user is looking to buy, create a page about buying. If a user is looking for an answer, provide that answer and all the additional information they need. You are writing for the person reading first and foremost, not the search engine. Anticipate what the user needs and help them get to their destination. Think about how a user would like that information displayed (Google gives you the clues) and format it accordingly. Listicle, comparison, guide, data figure, table, chart — whatever makes sense for that search query.
Now, you don’t have to apply every tip below (I’ll highlight the most important ones), but the more you apply to your content, the higher your chances are of hitting the top of Google and jumping ahead of your competition.
1. Use proprietary data – Google and other AI search engines feed on new and fresh data. Provide that to them and your content will do well. Sometimes all it takes is collecting the data that already exists on the web and putting it into a readable format.
2. Answer-first content – with all queries, you should make it a rule to provide the answer upfront and as soon as you can. This is great for both users and bots. A bot doesn’t want to spend time crawling a full page to get the answer, and neither does a user.
3. Always be specific – when you are answering questions, try to be as specific as possible. If you can’t be specific, do more research. For example, instead of providing a figure range, give an average or median. Experts give specifics; AI and amateurs speak in generalisations. This is good for AI search and for user experience.
4. Cut the fluff – no one wants to read your long intro paragraphs. Cut them out of your content and get to the point. Every sentence should exist for a reason, and that reason shouldn’t be word count.
5. Context and explanation – the format you should follow with every section is answer, then context. Answer the question and then provide context and additional information to support the answer. Add more data, add different data, or simply explain the logic behind the answer.
6. Offer new insights – always try to offer at least three new insights that haven’t been mentioned elsewhere. Look at the page one results, read their content, and try to add new info to the conversation. If you can’t add new insights, then your content will not stand apart, and you’ve given Google no reason to consider your content.
7. Avoid AI-written content – AI-written content has done more harm to websites than good. Google will penalise mass-published AI content and rank it lower on average. Human-written content has a higher chance of performing better. There are ways to produce AI content that isn’t terrible, but it requires a lot of work and is best left to writers who know what they are doing.
8. Graphics and data visualisation – Google likes rich content. It likes to see images, videos and graphics within your content. This indicates that you took time to craft your web page and gave it elements that it can display elsewhere, like in the Google Images tab.
9. Cite sources – use external sources to back up your claims. A well-researched and well-written piece will pull from multiple sources, and external links and a reference list will help Google know that you have done your research.
