Content Optimization for AI Overviews: Visibility to AI Citations

Sponsored Post by Adsy

TL;DR: If you already work on your website’s SEO, you don’t need to start over for AI Overviews optimization. But you do need to look beyond rankings, since a top position doesn’t guarantee a citation. Check whether your content answers the questions customers actually ask and where your experience could add useful details. Keep your business profiles up to date too. There’s no special file or writing formula that will get you cited.

Key takeaways

• Rankings and citations are different results. A page can do well in regular search and still be missing from an AI answer. It’s useful to check both when reviewing your visibility.

• Traditional optimization remains part of the process. Your pages need to be accessible to Google and useful to visitors. The work you’ve already done on these things still serves a purpose.

• Your daily work can give you material to publish. A project, a survey, or a problem your team solved may help readers understand something better. You don’t always need to come up with a completely new subject.

• Readers may need more than the first answer. Give them the details they need to use your advice, and organize those details so they’re easy to find. Let the subject determine how much explanation is necessary.

• Some AI advice adds unnecessary work. You don’t need an llms.txt file for Google or a fixed paragraph length for AI citations. A change should have a clear purpose before you apply it across your website.

Are AI Overviews different from traditional rankings?

AI Overviews are part of Google Search, but the way they present information is different from regular results. With traditional search, people get a list of pages and choose which ones to visit. An AI Overview gives them an answer directly, with links to websites where they can read more about the topic.

But those links don’t necessarily come from the pages ranking at the top.

According to Ahrefs’ research, only around 38% of AI Overview citations appeared within the first 10 results for the same query. This includes different search features. When looking at regular organic links alone, the figure was slightly lower, at 37.1%.

So, your article could rank well and still have limited AI search visibility. Besides that, a page outside the top 10 could get cited.

The main difference is that ranking a page and selecting sources for an AI answer aren’t the same thing. An Overview may cover several parts of a question, using different pages to support its response. One of the mechanisms that makes this possible is query fan-out.

What does query fan-out have to do with this?

Query fan-out is a technique that allows Google to run several related searches when preparing an AI response. Google explains that both AI Overviews and AI Mode may use it to explore subtopics and find supporting information.

For example, imagine a business owner searching for an email marketing platform for their online store. There are several things they may need to consider before choosing one. Related searches could include:

• How much different platforms cost.

• Which platforms work with their store.

• Whether they can automate abandoned-cart emails.

These are possible examples of related searches. The actual queries Google uses may differ.

Here, an article about abandoned-cart emails could help answer one part of the original question. It may not appear among the top results for email marketing platforms, but it could still provide relevant information for the Overview.

This helps explain why checking your main keyword’s rankings won’t show you all the pages you could be competing with for a citation.

Does traditional SEO still matter?

SEO covers more than getting a page into Google’s top 10. It also includes helping search engines find your content and making the website useful for people who visit it. These things also form the foundation of AI search optimization.

According to Google’s guidance, its AI features rely on the core Search ranking and quality systems. So, businesses that already follow SEO best practices don’t need to start everything from the beginning.

They still need to pay attention to:

• Whether their important pages are accessible to Google.

• Whether the content helps people understand a topic or complete a task.

• Whether visitors can find their way around the website.

• Whether pages work well on the devices people use.

For example, a web design agency may have a useful article explaining website redesign costs. Keeping it accessible and easy to read supports visitors coming from both traditional results and AI answers.

Of course, doing these things won’t make Google cite every article you publish. But the difference between rankings and citations doesn’t make traditional optimization unnecessary. It means businesses have another type of SEO visibility to consider.

For more information on how to approach this work, our AI content optimization guide brings together advice from SEO specialists.

How to get cited in Google AI Overviews

1. Build a strong SEO foundation

A useful article can still have problems that make it harder to discover. It may have no internal links pointing to it, or your website settings may prevent Google from accessing it.

That’s why the first step is to check the basics. Among its SEO recommendations for AI features, Google includes:

• Making content accessible to its crawler.

• Connecting pages through internal links.

• Providing important information in text.

• Keeping pages easy to use.

• Updating Business Profile and Merchant Center details where relevant.

For example, an agency could link a relevant case study from its service page. Visitors would have somewhere to explore the agency’s experience, while Google would have another route to discover the study.

2. Focus on topical authority and deep expertise

Your business probably knows some subjects much better than others. These are a useful place to start when deciding what to publish as part of your AI search optimization work. Google’s guidance also encourages businesses to bring their own knowledge and experience into their content.

An email marketing agency may know why subscribers stop opening emails. Its team can explain what they check first, which problems they often find, and how the solution depends on the client.

To build on that knowledge, consider covering:

• Questions that come up during client calls.

• Problems your team has solved more than once.

• Situations where the usual advice doesn’t apply.

You’ll be surprised how quickly these articles can give readers more reasons to return to your website when they need help with that subject.

3. Publish ‘non-commodity’ content

AI tools can help write an article, but they haven’t worked with your customers or tested your services. Those experiences can give your content something that a general summary of the topic would miss.

For instance, a web design agency could write about a project where visitors struggled to find the booking form. It could show the original layout, explain the changes, and share the results.

You can use the same approach with:

• Case studies from completed projects.

• Research based on your own data.

• Practical examples from your daily work.

Google encourages this kind of contribution in its guidance on non-commodity content. The example doesn’t have to be unusual. Even a familiar problem can be worth discussing when you have useful details to share.

4. Cover your topics in-depth

Your readers may have several questions about the same subject. Answering the first one can help them get started, but they may still need more information before making a decision.

A business owner reading about email automation might want to know which emails they should automate first, what they need to set up those emails, or how to check whether the automation works.

This is where query fan-out becomes relevant again.

Google’s bots may be dispatched to search related subtopics when preparing an AI Overview, so an article covering these questions could contribute to different parts of the response.

The key point to understand here is that the amount of detail depends on the subject. Sometimes a short explanation is enough. Other questions need examples or instructions that readers can follow.

5. Structure your content logically

Not all topics can be explained on a single page; some need a long explanation. But that doesn’t mean every reader needs to go through the whole thing. A person may come for a definition, a comparison, or one step they don’t understand. In fact, modern users prefer to scan through the content, rather than read it from A to Z.

When deciding how to structure content for AI search, start with how people will use the page. There are several ways to make the information easier to find:

• Use descriptive headings for the main questions.

• Provide a clear table of contents from the start.

• Include a DLTR section on top of your article or an executive summary.

• Include key takeaways or a summary when there’s a lot to remember.

The inverted pyramid structure can help too. It puts the main information first, with further details below. This is an ideal structure for busy visitors who need to grasp the essence quickly, and if they like what they find, they might go deeper, reading other sections of your post/article.

6. Get more external brand mentions

Your website isn’t the only place where people can learn about your business. They may find a customer review, watch an interview, or see your agency included in a list of recommended services.

Original research gives businesses a reason to pay attention to these mentions:

• Ahrefs found a strong relationship between YouTube mentions and AI visibility, including AI Overviews.

• Wix’s research found listicles were common citation sources across ChatGPT, AI Mode, and Perplexity.

• Semrush identified Reddit among those platforms’ frequently cited websites.

You shouldn’t read these findings as proof that every brand mention increases your chances of getting into SEO AI Overviews. They only indicate that some authoritative places may indeed be useful targets for your brand mentions, but only when you approach them with a good reason (e.g., a relevant and high-quality article to post) and on the right occasion (when your brand mention is truly relevant or makes good sense).

7. Make sure your brand information is accurate across channels

When you change something in your business, updating the website is usually the first thing you do. But there may be other pages that need attention too. Your social profiles, business directories, and review platforms may still have the old information.

An agency could add a new service and announce it on LinkedIn, while forgetting to update its profile on a review site. Or it could move to another city and leave the previous address in several directories.

So, it helps to check these pages whenever you make a change. The wording can be different on each platform. The main thing is that the details agree, so customers and AI systems can get a clear understanding of what your business offers.

Most common AI optimization myths

When you start looking into AI optimization, you may come across advice that makes it sound like your website needs a completely different setup. There are files to add, paragraphs to shorten, and special formats to follow. But some of these tasks aren’t necessary for appearing in Google’s AI features.

Google addresses several of these claims in its own guidance:

• You need an llms.txt file. Google says it ignores these files. Adding one won’t improve your visibility in AI Overviews, although other services may have their own uses for it.

• Every article needs to be split into small chunks. Google can understand different topics within a page. There’s no requirement to turn every explanation into a set of short, separate answers.

• Google AI search needs special schema markup. No specific schema is required for AI Overviews. Regular structured content and data can still help your pages qualify for rich results.

• You have to include every possible query variation. Google can understand related meanings and synonyms. You don’t need a separate sentence for every way someone might ask a question.

However, not every short section or detailed heading is automatically, or necessarily, a bad thing. Everything depends on placement, i.e., where you are using them. They can be harmful if misused, but very helpful in situations where the context demands them.

If a paragraph is about a complex topic (e.g., psychology of interpersonal relationships, or about the science of how AI works), more detail may be needed. But if you’re explaining a relatively simple thing, or mentioning a known and well-understood concept, you can keep it relatively brief.

In any case, there’s no need to reshape a useful article just to follow an assumed AI preference.

Frequently asked questions

Can a small business get cited in AI Overviews?

Yes, a small business can appear in AI Overviews. You don’t have to be a large company, but Google must be able to use your page in Search with a snippet. That makes the page eligible for a citation, although Google may still choose another source.

Do you need to rank on the first page to appear in AI Overviews?

No. Even if your page is not on the first SERP in Google, your authoritative content can still qualify for AI Overviews and appear therein. For example, Ahrefs found that most cited URLs in its study fell outside those positions. That’s why checking your keyword rankings alone won’t tell you whether your content appears in AI answers.

Should you create separate content for AI Overviews?

You don’t need a separate version of your website or articles written specifically for AI. Google says its usual SEO practices remain relevant to AI Overviews. Start with the content you already have and check whether it answers your audience’s questions clearly and provides useful information.

How long should an article be to get cited in AI Overviews?

An article doesn’t have to be long to get cited. Google says there’s no ideal page length for its AI search features. If you’ve answered the question fully, there’s no need to keep adding paragraphs.

Does getting mentioned in an AI answer mean your website gets a link?

No, these are two different things. A business can be named in an AI answer while the source link points to a customer review on another platform. When checking your results, record brand mentions and links to your website separately so you can see which you’re getting.

The bottom line

AI Overviews give businesses another place to be found, but appearing there isn’t something you can arrange by following a fixed checklist. Good rankings don’t guarantee a citation, and adding more content won’t necessarily solve the problem either.

A good place to begin improving your AI search visibility is with the questions your customers need answered. Check whether your pages answer their questions and whether your team has experience worth adding.

Besides that, keep up with the ordinary website work: internal links, accurate details, and pages that are easy to use. These improvements serve a purpose even before you see your first AI citation.