SEO & performance
AI Overviews
SGE · Search Generative Experience
AI Overviews are the AI-generated summaries Google places above traditional results, synthesised from several sources and citing them. They absorb a large share of informational clicks, which changes the goal of content: being cited in the summary now matters more than ranking beneath it. Pages win citations by answering a question directly, in a self-contained passage, near the top.
Why it matters
AI Overviews changed what a top ranking is worth. A query that once sent most of its clicks to the first organic result now frequently resolves in the summary itself, and the traffic that remains goes to the sources cited within it.
For informational content — how something works, what a term means, which option suits a case — the practical goal has shifted. Ranking first below the summary is worth less than being one of the sources the summary quotes.
For transactional queries the effect is smaller, because people still want to see the product, compare, and buy. Which means the honest response is not panic but segmentation: work out which of your queries are informational and accept that their traffic profile has permanently changed.
How it works on Shopify
The summary is synthesised from several sources and cites them, and the pages that get cited share recognisable traits.
They answer the question directly and early. A passage that states the answer in its first sentence, in complete form, without requiring the surrounding paragraphs, is extractable. Content that builds to its answer over four paragraphs is not.
They are self-contained. Each section stands alone, because the system is quoting a passage rather than reading a page.
They are specific. Numbers, thresholds, named versions, and concrete conditions survive summarisation better than general claims, which get compressed into nothing.
They are corroborated. Claims a business makes about itself carry less weight than claims other credible sources also make, which is why E-E-A-T signals and third-party mentions matter here.
Structured data helps the system resolve what entities a page is talking about, particularly for products and organisations.
Common mistakes
- Burying the answer. An introduction before the answer is what stops a passage being quoted.
- Writing for word count. Padding dilutes the extractable content and helps nothing.
- Vague claims. "Improves performance significantly" survives no summarisation. "Reduces LCP by roughly a second" might.
- Blocking AI crawlers by accident. A robots rule intended for scrapers can remove you from the sources entirely.
- Measuring only clicks. Impressions with falling click-through on informational queries is the signal, and it looks like a ranking problem if you only watch position.
- Chasing citations for transactional queries. Effort is better spent where the summaries actually appear.
When you need help
The useful analysis is diagnostic: which of the site's queries now return AI Overviews, what share of its impressions those represent, and whether the site is cited in them. That reframes a traffic decline from mysterious to specific, and it is measurable with data the business already has.
The other case is restructuring content to be citable — passage-level rewriting, adding the specificity that survives summarisation, and building the corroboration that makes claims credible. That is editorial work with a technical foundation, described further under GEO.
Need this done on your store?
Semantic SEORelated terms
- GEOGEO, or generative engine optimization, is the practice of making content citable by AI search engines — Google's AI Overviews, ChatGPT, Perplexity — rather than merely rankable in a list of blue links. In practice it rewards clear entity definitions, self-contained answers, structured data, factual precision, and third-party corroboration of the claims you make about yourself.
- Structured dataStructured data is machine-readable markup that tells search engines what a page means, not just what it says. On ecommerce sites it is written as JSON-LD following Schema.org vocabulary, and it is what produces rich results — star ratings, prices, and stock status in the SERP. It is also how AI search engines resolve entities, which makes it foundational to GEO.
- E-E-A-TE-E-A-T stands for experience, expertise, authoritativeness, and trustworthiness — the framework Google's human quality raters use to judge content. It is not a direct ranking factor but a description of what Google's algorithms try to approximate. For ecommerce it cashes out as real author bylines, genuine reviews, clear returns and contact information, and claims that other credible sites corroborate.