How to Appear in Google AI Overviews

What are Google AI Overviews and where do they appear?

AI Overviews are AI-generated summaries that Google places above the traditional search results on a subset of queries. They synthesise an answer from pages Google has retrieved, then cite a small number of sources beneath it. They are built on the same core ranking and quality systems as normal search, and Google states explicitly that the SEO fundamentals for appearing in them are unchanged.

They do not appear on every query. Google has said it deploys them where they add value over a list of links, and the presence rate varies by topic, query phrasing and user location. For many commercial and informational queries they are now common enough that they materially change the economics of the results page.

Why do AI Overviews reduce click-through?

Because they answer the query on the results page. When a user gets what they needed without clicking, there is no reason to click — and studies of zero-click search have consistently found substantial click-through loss on queries that trigger an AI answer.

The commercial effect is uneven, and the businesses that lose most are the ones whose traffic came from informational queries rather than from people already ready to buy. If your traffic converts, the loss matters less than the raw click numbers suggest. If it does not, an AI Overview above your listing can remove most of a page's traffic overnight.

How does Google select sources for an AI Overview?

Google retrieves pages relevant to the query, then synthesises an answer grounded in what it retrieved. Pages must be indexed and eligible to appear in search with a snippet — a page that is not indexed cannot be cited regardless of how good it is. Beyond that, the research on generative visibility gives us the clearest picture of what correlates with citation.

Four consequences follow, and each one changes how you should write.

• Passages must stand alone. A retriever lifts a chunk of text without the surrounding page. If your answer depends on context three paragraphs earlier, the lifted version is wrong or incomplete.
• Specificity beats comprehensiveness. "AI Overviews launched in the United States on 14 May 2024" is a retrievable fact. "AI Overviews launched recently" is not.
• Freshness is weighted, not decisive. Current questions favor current sources, and a page cited last month that has since gone stale tends to be replaced by a fresher competitor.
• Heading phrasing matters. People ask assistants full questions. A heading phrased as the literal question is more likely to be matched than "Optimization Strategies".

Which tactics actually work, and which don't?

This is where most GEO advice goes wrong, so it is worth separating what has evidence behind it from what is merely repeated. The 2024 GEO paper by Aggarwal and colleagues at Princeton and IIT Delhi, published at KDD 2024, tested nine content modifications and measured the effect on generative visibility.

Two results in that table are counterintuitive and worth internalising. The largest gains went to pages that were not already ranked first — the paper reported roughly +115% visibility from source citations for a rank-five page, while the top-ranked page could lose ground. Adding evidence is a leveling play, not a winner's prize. And keyword stuffing was actively harmful.

The strongest predictor is not technical

Ahrefs analyzed 75,000 brands and reported a correlation of 0.664 between branded web mentions and appearing in Google AI Overviews, against 0.326 for Domain Rating. How often and how visibly your brand is discussed elsewhere on the web correlates roughly twice as strongly as your link profile. That is an off-site authority problem dressed up as an on-site one, and it is why GEO work that stops at the website has a ceiling.

How do you get cited rather than merely ranked?

Ranking first is the precondition. Citation is the goal, and it comes from being the most useful and most extractable source available on a specific question. Four practical moves, in order of impact.

1. Write sections that answer their own heading completely. Read any single section in isolation. If it fully answers its heading without referring to anything above or below it, it is citable.
2. Put evidence in the sentence. Named sources, numbers, dates. A claim with a source attached is more likely to be repeated accurately than one without.
3. Answer the question in the first 40 words. Retrievers favor passages that resolve the question immediately. Do not build up to it.
4. Own something specific. Original data, a benchmark you ran, a case from your own work. This is the one thing a language model cannot generate, which is precisely why it gets cited.

How do you measure AI Overview visibility?

You cannot measure it with the tools you already have. Search Console does not report it, GA4 does not know about it, and rank trackers do not model it. Measuring it means sampling the answers themselves.

The method that works is a citation ledger. For a defined set of priority queries — typically the 30 to 100 questions your customers actually ask — sample answers across Google AI Overviews, ChatGPT Search, Perplexity and Gemini on a recurring schedule, and record two numbers separately.

Track them on a rolling window rather than a single day, because answers are non-deterministic and the cited set rotates. Tools such as Profound, Otterly, Athena and Goodie automate the sampling; the discipline matters more than the tool, and a spreadsheet updated weekly beats an expensive dashboard nobody opens.