How to rank in Google AI Overviews – get cited in the AI answers, and see whether they have taken your traffic

Since Google rolled out AI Overviews globally, new clients keep asking us the same question: why are we not showing up in the AI answers, and what can we do about it? The answer is more manageable than you might think – but it means understanding how Google chooses its sources and which actions actually affect the outcome. That is the focus of this guide: how to rank in AI Overviews and become a cited source in Google's AI answers.

Being visible in AI Overviews is part of the broader discipline of AEO – sometimes called GEO – but here we focus on Google's own implementation, not on AI answers in general.

What are AI Overviews?

AI Overviews are Google's AI-generated summaries at the top of the search results for selected queries. Google pulls information from several sources, summarises the answer and links to the pages that contributed. According to SparkToro's analysis, AI Overviews appear in just over 20 per cent of all searches, with large variation between search types. Informational searches and "how does X work" questions trigger them considerably more often than commercial or navigational searches.

If you are one of the cited sources, you can get exposure for more searches than you rank for organically – often to visitors with a higher intent to convert.

What decides whether Google cites you?

Google publishes no official checklist. But the pattern among pages that are cited repeatedly points to a few factors:

• Topical authority – you already rank organically for related questions. AI Overviews mostly choose sources that are already established in the topic.
• Standalone answers – content that gives a direct answer at the start of each section, not buried in long introductions.
• Factual precision and sourcing – data claims should be verifiable and backed by sources.
• Technical performance – fast pages with correct schema markup are easier for Google to interpret and extract content from.
• E-E-A-T signals – experience, expertise, authoritativeness and trustworthiness. A clearly identified publisher, knowledgeable writers and external references to your website.

Concrete steps to rank in AI Overviews

Structure the content for extraction

AI models do not read pages linearly. They extract meaningful blocks of text based on the HTML structure. In practice:

• Start every H2 section with a direct answer, not background.
• Use H2s and H3s that mirror actual search queries: "What is X?", "How does Y work?".
• Write so that every paragraph can stand on its own and give a complete answer – semantic chunking in practice.
• Map out the follow-up questions Google may break a search down into – known as query fan-out.

Implement schema markup

Article schema and other structured data help Google identify and interpret specific content types. It does not guarantee a place in an AI Overview, but it makes the content easier for machines to read.

Build topical authority, not individual pages

A single well-written article is rarely enough. Google values websites that consistently cover a topic area in depth. That is why it pays to work systematically with topic clusters and a knowledge base – covering a whole topic area rather than making one-off efforts.

Strengthen the E-E-A-T signals

In practice, this means clear About pages with expert bios, external sources for factual claims and – most importantly – mentions of your website by industry publications, partners or the media. Mentions outside your own website carry more weight than any amount of on-page optimisation.

Does AI Overviews optimisation replace ordinary SEO?

No. The most common misconception is that optimising for AI answers replaces traditional SEO. You cannot ignore organic rankings and hope that AI Overviews will make up for it – the two build on each other. Google largely cites pages that already rank organically for related questions, even though the correlation with the top 10 specifically has weakened.

What changes is how we measure success. Clicks are still relevant, but exposure and citations in AI answers are a new dimension. Tools such as Ahrefs and Semrush have started to build in AI Overviews tracking, but measurement is still maturing. We go deeper into how AI is changing search traffic in How is SEO affected by AI?

What AI Overviews do to your traffic

The most common fear is that AI Overviews eat up all organic traffic. According to an analysis by Contently (2026), the traffic loss is greatest for simple factual questions with unambiguous answers, while comparison, decision and purchase-related pages tend to keep or increase their traffic.

That is the traffic you should prioritise. A visitor who clicks through from an AI answer is often further along in their decision – and so has a higher intent to convert than general organic traffic.

Have AI Overviews taken your traffic? How to find out

If your website is losing clicks even though its positions have not fallen, AI Overviews are a likely explanation – but not the only one. The pattern that points to the AI Overview effect is stable positions and stable or rising impressions combined with falling CTR: the page is seen as much as before, but fewer people click through because the answer is already in the AI summary.

The data points the same way. According to Pew Research (68,879 searches, 2025), users click on a regular result in 8 per cent of searches when an AI summary appears, compared with 15 per cent without one. Ahrefs' measurement of 300,000 keywords (December 2025) shows a 58 per cent lower CTR for the top-ranking page when an AI Overview is present in the search results.

Here is how to diagnose it in GSC before drawing conclusions:

• Compare clicks with impressions per page over the last 16 weeks. Falling clicks with stable impressions and a stable position point to the AI Overview effect.
• Check which search queries are losing clicks. Informational queries ("what is", "how does") trigger AI Overviews most often – if purchase-related queries are losing just as much, the cause is probably something else.
• Rule out algorithm updates. If positions have dropped at the same time, the problem is rankings, not AI Overviews – compare the date of the drop with Google's confirmed updates.
• Rule out seasonality and technical issues. Compare with the same period last year, and check the indexing report to make sure the drop is not caused by a technical regression.

Be aware of the limits of this measurement, though: GSC does not report AI Overviews separately, so the diagnosis rests on patterns, not an exact figure. And the right response is not panic but prioritisation – put your effort into the comparison and purchase-related pages that keep their traffic, and into becoming the source that is cited in the answers.

The technical foundation matters

Fast pages, correctly structured data and a clear content hierarchy decide how well Google can interpret your website. That is why choosing a platform for a new website is not just a design decision – it is also a visibility decision. Read more about choosing the right technical foundation in CMS for large and growing companies.

Want to know where your company is visible – and not visible – in AI-generated answers today? Get in touch – analysing AEO and visibility in AI Overviews is part of our work as an AI SEO agency.