How to Get Cited in AI Overviews: A Design Guide

Strong Google rankings help, but clear, useful pages give AI systems better material to understand and cite.

A business owner reviews a clear service page beside an AI answer with a source citation, illustrating how to get cited in AI Overviews.
On this page
  1. What does it mean to get cited in AI Overviews?
  2. Why can a page rank well but not appear in an AI Overview?
  3. How does query fan-out affect the pages AI may cite?
  4. What makes a page easier for AI to understand and cite?
  5. Does schema markup help you get cited in AI Overviews?
  6. How should you redesign a service page for AI visibility?
  7. How do you measure AI visibility if Google Search Console does not show a citation report?
  8. What should small businesses do first?
  9. Common mistakes when trying to get cited in AI Overviews
  10. The practical takeaway

Ranking on Google and getting cited in AI Overviews are related, but they are not the same thing. To get cited in AI Overviews, make your pages easy to understand, specific enough to answer real questions, and supported by clear evidence. Good design helps by putting the answer, structure and proof where people and search systems can find them.

If your site ranks well but enquiries or organic clicks are slipping, the next step is not automatically to publish more. First, review what your pages say and how they present it. A free AI search audit can help identify where your business appears in AI answers and where useful information is hard to find. If the problem is page structure or an outdated experience, website design and redesign can address those issues as part of the build, not as a cosmetic layer afterwards.

This guide explains how AI citations differ from rankings, what makes a page easier to cite, and which design changes are worth testing. It also explains what not to assume: no layout, schema markup or writing trick can guarantee a citation.

What does it mean to get cited in AI Overviews?

An AI Overview is a Google search feature that uses generative AI to produce a summary in response to some searches. It may include links to webpages that support or expand on parts of that summary. When your page is shown as one of those links, it has been cited in that AI Overview.

A citation is not the same as a ranking. A traditional organic result is a webpage displayed in a ranked list. A citation is a source linked from an AI-generated response. A page can appear in one place, both places, or neither. The exact results can vary by search, location, language, device and changes to Google’s systems.

Google’s documentation on AI features in Search says the usual foundations of SEO still matter. There is no special technical requirement or separate markup that guarantees inclusion in an AI Overview. That is an important starting point: you are improving the usefulness and clarity of your content, not trying to unlock a secret ranking switch.

Ranking versus citation: what is the difference?

Ranking describes where a page appears in the conventional search results for a query. Citation describes whether a source is selected to support a generated answer. Ranking depends on many signals, including relevance and quality. Citation also depends on whether the page offers information the system can use for the specific answer it is composing.

Think of a search for “how long does a commercial solar installation take?” A page may rank for that broad topic, but a useful answer may need to distinguish planning, approvals, installation and connection to the grid. A different page that explains one of those steps clearly could be a more useful source for that part of the response, even if it does not outrank every competitor for the broad phrase.

That does not mean rankings are irrelevant. Strong, relevant pages can still be discovered through search and can earn organic visits. It means that a single position is not a complete measure of how visible your business is in search anymore.

Traditional organic result AI Overview citation
Displays a page in a ranked list Links a source within or beside a generated response
Often competes for a click to the page May support an answer before the user decides whether to visit
Ranking is usually tracked by query and position Citation visibility needs to be checked across relevant questions and contexts
A broad page can rank for a broad phrase A specific page or section may be useful for one part of a larger answer

An Ahrefs analysis of AI Overview citations explores the relationship between citations and organic rankings. Treat any study of this kind as a snapshot of the queries and dates it measured, not a promise about what Google will do for your site. The practical lesson is to assess both rankings and whether your useful pages appear as sources.

Why can a page rank well but not appear in an AI Overview?

A strong ranking tells you that Google considers your page relevant to a search. It does not mean the page directly answers every related question an AI Overview might address. A result can be relevant but still hard to quote, too general, out of date, or missing the specific detail needed for the answer.

A common cause is a mismatch between the way a business describes its services and the way a customer asks for help. A page titled “Solutions for modern businesses” might look polished, but the words do not tell a visitor whether the business handles commercial solar installation, residential repairs or ongoing maintenance. A clear service page that names the audience, problem and service is more useful to a person trying to choose a provider, and easier to interpret.

Other common problems include:

  • The answer is buried below a long brand introduction, hero video or repeated promotional copy.
  • Headings use slogans rather than describing the questions each section answers.
  • Important details are trapped in tabs, images or interactive widgets.
  • The page makes claims without showing how they are supported.
  • Several pages repeat the same broad information without offering a distinct, useful answer.
  • The business’s services, locations or terms are described differently across the site.
  • A page is difficult to scan on a phone, even if its desktop layout looks good.

A useful diagnostic is to ask: if a customer saw only this page, could they tell what we do, who we do it for, what happens next and why they should trust us? If the answer is no, a search system may also have difficulty identifying which parts of the page are useful for a particular question.

How does query fan-out affect the pages AI may cite?

Query fan-out is a process where a search system explores related sub-questions behind a user’s original query. Instead of treating “commercial solar installation timeline” as one narrow phrase, a system may need information about site checks, approvals, equipment and installation before it can produce a fuller answer.

Google describes the way AI features can use a range of searches and sources in its AI features documentation. The details of how an answer is assembled can change, and Google does not publish a checklist that reveals exactly which pages will be selected for every question. But the idea of related sub-questions is useful when planning content.

Start with the customer’s main question. Then list the smaller questions they need answered to make a decision. For a clinic, “Which physiotherapist should I choose?” could involve areas of treatment, practitioner qualifications, appointment access, location and what a first appointment involves. Each topic might belong on one well-organised service page or on a related page that provides genuinely useful detail.

This is not a reason to create dozens of thin pages, each targeting a slightly different wording. It is a reason to cover the real decision thoroughly, with a clear structure and no unnecessary duplication. A content plan should reflect how customers compare options, not just a list of keywords.

  1. Write down the main customer question. Use the language people use when they call, email or search. Avoid starting with internal service names.
  2. List the decision questions underneath it. What does the customer need to know about fit, cost, timing, process, location or risk?
  3. Group questions that belong together. Put closely related answers on one page. Separate a topic only when it needs enough detail to deserve its own page.
  4. Check the results and the business reality. Look at the questions people ask in search, then confirm that each proposed answer is accurate and useful for your business.
  5. Choose a page for each group. Make sure your navigation and internal links help a visitor move from the general service to the detail they need.

For instance, a law firm should not bury its areas of practice under a generic “How we help” page if prospective clients need to understand a particular legal service. A dedicated page can explain who the service is for, common situations, what a first discussion covers and how to get in touch. The design can still feel consistent with the wider brand while making that specific information much easier to find.

What makes a page easier for AI to understand and cite?

There is no guaranteed citation formula. The strongest approach is to make the page genuinely helpful, accurate, clearly organised and accessible to both people and search systems. Design decisions can support that by making the information easier to locate, scan and interpret.

1. Put the direct answer near the top

A page should not make a visitor work through several paragraphs of brand language before they discover what it is about. Under the main heading, give a short, plain answer to the question the page addresses. Then add the context, caveats and detail that help the reader make a decision.

For example, a page about a clinic’s sports physiotherapy service could open with: “Sports physiotherapy helps people manage sports injuries and return to activity with an individual assessment and treatment plan.” The clinic should adjust that sentence to match its actual service, qualifications and scope. The point is to make the subject unmistakable, not to insert a rehearsed snippet.

This is sometimes called an answer-first structure. It is an editorial choice, not a special Google rule. A good designer makes that first answer easy to see without turning it into an oversized promotional banner. Use readable type, clear spacing and a subtle visual distinction so it stands apart from surrounding material.

2. Write headings that describe the answer underneath

Headings create a map of the page. A heading such as “A better way forward” gives a reader little idea of what comes next. “What happens at your first appointment?” tells them what the section answers.

Use a clear order: one main heading for the page, then section headings, then smaller headings where they add a useful subtopic. Do not choose heading levels solely because a certain size looks attractive. Typography can be styled independently from the underlying heading structure, so the visual design and page outline can both work well.

Question-led headings can be useful when they reflect real questions. They are not required everywhere. A natural heading such as “What our website redesign includes” can be clearer than forcing every section into a question. The test is whether a visitor can scan the headings and understand the page’s structure.

3. Use specific service pages, not just a polished homepage

A homepage has a broad job. It introduces the business, signals its positioning and guides visitors to the right next step. A service page has a narrower job: explain one offer to one audience and help that audience decide what to do.

A Search Engine Land report on AI Overview citations and deep pages discusses the reported prominence of pages deeper in a website compared with homepages. The exact proportions in a study should not be treated as universal, but the underlying content principle is sensible: a specific answer is more likely to live on a page written for that specific subject.

A useful service page usually explains:

  • What the service is and who it is for.
  • The situation or problem it helps with.
  • What is included and what is not.
  • How the process works.
  • Relevant experience, qualifications or evidence.
  • What a customer can expect next, including how to enquire.

Do not create a separate page simply because a keyword tool shows a phrase. Make a page when there is a real audience, a meaningful difference in the service or enough distinct information to answer the question properly.

4. Make comparisons and steps easy to scan

A table is useful when readers need to compare options. A numbered list is useful when the order of steps matters. A definition list or clearly labelled set of short sections can help explain technical terms. These formats are good editorial design because they reduce the effort of finding a detail.

Keep tables simple. Give each column a clear heading, keep the information concise and make sure the table still works on a narrow screen. If it becomes too wide on mobile, consider a different layout rather than shrinking the text until it is hard to read.

Do not hide essential information behind tabs or an interaction that only appears after a visitor clicks, unless that interaction has a clear usability reason. Search systems can process many modern webpages, but a visitor should not have to guess where the important content is. A plain, visible explanation is usually a safer design choice than an elaborate widget whose content is difficult to scan or access.

5. Use diagrams to explain, not decorate

A process diagram can clarify a service that is difficult to explain in paragraphs. A labelled diagram of a booking process, product assembly or project timeline can help a person understand the steps quickly. It can also provide a visual layer of information for systems that interpret images.

But an image is not a substitute for the explanation. Add a useful caption and descriptive alternative text where appropriate. Alternative text is a written description that helps people using assistive technology understand an image. A complex chart should also have the important facts available in nearby text. Do not rely on a graphic to carry details that are essential to understanding the service.

Use original diagrams where they add real value. Generic stock photography may make a page look finished, but it rarely explains what makes your process, product or expertise different. A well-labelled diagram can do more for trust than another decorative image.

6. Make claims specific and support them

A page should help a reader distinguish verified information from marketing language. Replace vague claims such as “industry-leading results” with concrete, supportable explanations of what the business does and what the customer can expect. If you publish qualifications, service locations, product specifications or policies, make sure they are accurate and consistent.

Evidence can include relevant credentials, a clear explanation of the process, genuine customer reviews, named case studies published with permission, or references to original research where relevant. Do not invent numbers or imply a result is typical if it is not. Good evidence makes a page more useful to customers as well as easier to assess.

7. Keep important content in the page itself

The website’s page structure matters. Semantic HTML is the underlying code that labels content according to its role, such as a heading, paragraph, list or table. A semantic page gives browsers, assistive technologies and other systems clearer information about how its content is organised.

You do not need to remove all JavaScript or build a plain text website. Interactive features can be valuable when they solve a real user problem. But important explanations should not exist only inside a graphic, depend on an interaction that fails, or appear after several unnecessary steps. Ask your developer to check that the main content is available in the rendered page and works for keyboard and screen-reader users too.

Does schema markup help you get cited in AI Overviews?

Schema markup is structured data added to a webpage to describe information in a format that search systems can interpret. For example, it can identify an organisation, a product, a service or a local business. This is different from the visible text and design that a customer reads, although both should accurately describe the same business.

Schema can help make information more explicit, but adding it does not guarantee a rich result, a higher ranking or an AI Overview citation. Google’s guidance on AI features does not require special AI-specific markup for eligibility. Follow Google’s published structured data policies and use a schema type only when it accurately represents content on the page.

For example, do not add FAQ structured data simply because a page has questions, if the markup does not match the visible content or the current eligibility rules. Do not mark up claims or reviews that are not actually present. And do not add schema as a substitute for a clear service description, helpful content or a sound website structure.

If you want a practical introduction to the visible and technical sides of local business information, see DesignFIT’s local business schema markup checklist. Use it to review what your site communicates, then validate structured data with Google’s own tools and documentation.

How should you redesign a service page for AI visibility?

A redesign for AI visibility should start with the customer’s task, not with a new colour palette. The goal is a page that presents a clear answer, enough detail to support a decision, and an obvious path to act. The visual system should make those elements easy to find without sacrificing the brand.

A useful service page layout might follow this sequence:

  1. Clear page heading. Name the service and, where useful, the audience or location.
  2. Short opening answer. Explain what the service does in plain language.
  3. Primary action. Make the next step obvious, such as booking, requesting a quote or discussing a project.
  4. Fit and scope. Explain who the service suits, what is included and any important limits.
  5. Process. Show what happens from first contact through delivery, with realistic expectations.
  6. Evidence. Include relevant credentials, examples or policies that can be verified.
  7. Decision support. Answer common questions, compare options or explain costs where appropriate.
  8. Final action. Repeat the next step after the visitor has had a chance to review the detail.

This order is a starting point, not a template to apply blindly. A high-trust professional service may need credentials and privacy information earlier. An ecommerce product page may need product specifications, delivery details and returns information near the purchase action. The right hierarchy depends on what the customer needs to know before they commit.

Visual hierarchy is the way a design signals what matters first, second and third. It comes from size, contrast, position, spacing and repetition. If every section has the same visual weight, readers have no clear route through the page. If the call to action is visually dominant but the service details are hard to read, the design may feel pushy rather than trustworthy.

A website redesign should also consider mobile layouts, accessibility, page speed, navigation and brand consistency. These are not separate from AI visibility. They shape whether a person can use the information once they arrive, and whether the site is clear and functional across different devices.

How do you measure AI visibility if Google Search Console does not show a citation report?

Google Search Console is a free Google service that reports how a website performs in Google Search, including impressions, clicks and average position. At the time of writing, Google’s public reporting does not provide a dedicated, complete report that labels each AI Overview citation separately from other search appearances. Check Google’s current documentation as reporting can change.

This means you should not try to infer an AI Overview citation from one metric alone. A lower click-through rate can have several possible causes: changing search demand, a different result layout, seasonality, a change in position or a change in the query mix. It may be worth investigating, but it does not prove that an AI Overview has taken a click from your page.

Use a combination of checks:

  • Search Console: Look at impressions, clicks, position and click-through rate for important pages and queries over time. Compare similar periods and check whether the query mix changed.
  • Manual searches: Search a set of real customer questions in the markets and locations you serve. Record the date, query, device and whether your business appears as a source. Results can vary, so repeat the checks rather than treating one search as definitive.
  • Page and enquiry data: Check whether organic sessions and qualified enquiries are changing together. A traffic decline is not always a business decline, and a traffic increase is not always a better result.
  • Third-party monitoring: Some SEO platforms offer AI visibility or citation tracking. Before buying, check which platforms and queries they cover, how often they update, what they count as a citation and whether the plan suits your needs.

Ahrefs provides an AI Overviews tracking tool and publishes research on how its measurements work. Its guide to appearing in AI Overviews is another useful perspective, but read any recommendations alongside Google’s official guidance. Third-party tools can help compare trends, but their samples are not a complete view of every person’s results.

Make a small, relevant question set before you start tracking. Include the questions customers ask before hiring you, the service terms you care about and local variations if geography matters. Keep a record of your wording and dates, and note whether a citation links to your homepage, a service page or another source. That gives you something practical to compare after a content or design change.

What should you do if clicks drop but impressions stay steady?

First, check which pages and queries changed, whether average position moved, and whether the period is comparable. Then review the current search results for those exact questions. If an AI Overview or another search feature appears, note the change as a possible explanation, not a confirmed cause.

Next, check whether your page still answers the question quickly and accurately. If the result is a service page, make sure the page helps a visitor take the next step rather than only attracting a broad informational visit. If the page is informational, check that it provides something useful beyond a short answer, such as a clear explanation, comparison, process or decision checklist.

Do not block snippets or remove useful content in a rush to control how a search result appears. Settings that restrict snippets can affect what Google can show for your pages. Read Google’s documentation and understand the trade-off before changing directives across your site.

What should small businesses do first?

You do not need to redesign every page or subscribe to several monitoring tools to begin. Start with the pages most closely linked to revenue: the main services, locations, products or booking paths that prospective customers use to decide whether to contact you.

Use this checklist for each priority page:

  • Can a new visitor identify the exact service and intended audience from the page heading and opening?
  • Does the page give a direct, accurate answer before the promotional detail?
  • Do the section headings tell the reader what each section contains?
  • Are key facts written as selectable text, not only shown inside an image?
  • Are comparisons and processes presented in a format that is easy to scan on mobile?
  • Does the page include relevant evidence, with claims that can be verified?
  • Are service names, locations and business details consistent with the rest of the site?
  • Can a visitor find the next step without scrolling past repeated, low-value sections?
  • Does the page work with a keyboard and assistive technology?
  • Is there a clear reason for every major block in the layout?

Then choose one page and improve it. Record its current organic performance and the questions you want it to answer. Rewrite the opening if needed, clarify the headings, add missing decision information and adjust the design so the answer and next step are easy to find. Review how users and search results behave afterwards. Make changes based on evidence, not an assumption that any one format will trigger a citation.

A content change can also expose a site-wide design problem. If service information is scattered, the navigation is confusing or every page uses the same generic template, a redesign may be more effective than repeatedly patching individual pages. DesignFIT’s website design service focuses on the structure and design decisions that help websites communicate clearly and support business goals.

Common mistakes when trying to get cited in AI Overviews

Treating a citation as a ranking hack

There is no reliable shortcut that guarantees a citation. Repeating a phrase, adding an arbitrary word count or inserting structured data without a real purpose can make a page worse for people without solving the underlying problem. Build pages around a useful answer and a genuine business need.

Putting the brand story before the customer’s question

Your business story matters, but it should not make the visitor search for the service itself. Lead with what you provide and who it helps. Share the story later, where it supports trust and helps someone understand why your approach is different.

Making every page broad

A page that tries to explain every service, audience and location can become difficult to use. Give important services enough space to describe their scope and answer the questions that affect a customer’s decision. Link related pages so people can explore without forcing everything into one long page.

Publishing a pile of shallow question pages

A list of short pages that repeat the same answer does not automatically demonstrate expertise. Group related questions, write complete answers and create a separate page only when the subject deserves one. Review older content for accuracy and usefulness instead of continuously adding new URLs.

Overstating what a study proves

Studies of AI citations are useful for generating questions and spotting patterns. They may differ by date, search sample, industry and method. Avoid turning a finding from one dataset into a universal rule for every business. Use the evidence as a prompt to test your own important pages.

Designing for a scraper instead of a person

Readable headings, well-structured text and accessible tables help people as well as machines. But a page should not become a collection of answer boxes with no personality or context. Make the information easy to extract because it is well organised, while keeping the experience useful, credible and recognisably yours.

The practical takeaway

Google rankings remain useful, but they are only one part of search visibility. To get cited in AI Overviews, make your site a clear source for the questions your customers actually ask: explain the answer early, organise the supporting detail, use specific service pages, and support claims with trustworthy evidence. Design is part of that work because hierarchy, layout and readable structure determine how easily people can find and understand what you know.

Start with your most important page, not your whole website. If you want a baseline of where your business appears in AI answers and what to improve next, request a free AI search audit.

Key takeaway

Rankings help people find a page, but clear structure, specific answers and trustworthy evidence make it more useful as a source.

Questions people ask

Why does my site rank on Google but not show up in an AI Overview?

A strong ranking does not guarantee that your page contains the clearest source for every part of an AI-generated answer. The page may be too broad, bury the relevant detail or lack the specific information the answer needs.

How do I get cited in Google AI Overviews?

There is no guaranteed method or special markup that ensures a citation. Publish accurate, useful pages with clear headings, direct answers, relevant supporting detail and a structure that works for people and search systems.

Does schema markup improve AI Overview citations?

Schema markup can describe information on a page in a structured format, but it does not guarantee an AI Overview citation. Use it only when it accurately represents visible page content and follows Google’s current guidelines.

Should I create a separate page for every question customers ask?

No. Group closely related questions on a useful, well-organised page, and create a separate page only when the topic has distinct information and a real audience. Avoid thin pages that repeat the same answer.

Can I tell from Google Search Console whether an AI Overview cited my page?

Search Console does not currently offer a complete, dedicated report that identifies every AI Overview citation separately. Use it alongside manual checks of relevant searches and other analytics, and treat click changes as clues rather than proof.

Sources

  1. google.com
  2. ahrefs.com
  3. ahrefs.com
Written by DesignFIT teamDesignFIT, the data-driven design studio
Share

Start a project

Ready to put this into practice?

Tell us about your business in three minutes. We recommend a plan, book your kickoff and get started.