What an answer engine is
An answer engine is any search surface that returns the answer instead of, or in front of, the list of pages. The category is older than generative AI. Featured snippets lifted a paragraph from a page into a box; voice assistants read one answer aloud; knowledge panels assembled facts about an entity. Each created the same pressure on content: make the answer extractable.
DOCUMENTED Google’s AI Overviews are the current version of that surface on the results page. Google describes them as a “core Google Search feature, like knowledge panels”, which cannot be turned off, shown “when our systems determine that generative AI can be especially helpful”. AI Mode is the conversational version. Chat assistants are answer engines by construction.
What AEO optimises
The object of AEO is the passage. An answer engine does not lift a page; it lifts a section, a sentence or a table, and the sections that get lifted are the ones that stand on their own. The working rules are simple and have not changed since featured snippets:
- Put the question in the heading and the answer in the first sentences beneath it.
- Keep the answer complete without the surrounding text. A reader who sees only that paragraph should still get the answer.
- Use the structure the answer has: a table for a comparison, a numbered list for a sequence, a short paragraph for a definition.
- Name the subject explicitly inside the passage rather than relying on the page title to supply it.
INFERRED These rules are inference from how extraction works and from long observation of featured snippets. No platform publishes a rule that a passage must be shaped a particular way.
What Google documents
DOCUMENTED Google’s May 2025 guidance for its AI experiences lists the things a site owner can rely on: meet the technical requirements so pages can be crawled and indexed with a 200 status and indexable content; use the preview controls (nosnippet, data-nosnippet, max-snippet, noindex) to set display preferences, noting that “more restrictive permissions will limit how your content is featured in our AI experiences”; make sure structured data matches the visible content; and provide images and video where they help. The same page says the underpinnings of Google’s long-standing advice carry across: “focus on your visitors and provide them with unique, satisfying content”.
DOCUMENTED The site-owner documentation adds that no new machine-readable files, AI text files or markup are needed, and that there is no special schema.org structured data for AI features. For AEO on Google this means the preview controls are the only documented levers over what is shown; everything else is content.
The click question
TESTED An answer engine that answers on the page changes what a click is worth. Pew Research Center’s browsing-data analysis found that US Google users who encountered an AI summary clicked a traditional result in 8% of visits versus 15% when no summary appeared, and clicked a link inside the summary in 1% of visits. The same study found 88% of summaries cited three or more sources. Being one of those sources is the AEO outcome; expecting the click volume of a classic result from it is not supported by the data. The guide on AI Overviews and organic traffic goes through what to expect.
AEO, GEO and AI SEO
The terms overlap. AEO is the right word when the goal is to be the extracted answer, on any surface that extracts. GEO is the right word when the goal is to be composed into a synthesised answer with citations. AI SEO is the umbrella this site uses for both, plus the access and identity work beneath them. A page built to AEO rules is usually the page GEO wants too, because a passage that answers a question cleanly is easier to synthesise from.
How AEO applies in Perth
For a Perth business the practical AEO question is which questions to answer. Local prompts carry a place, so the passages that get lifted for “in Perth” questions are the ones that answer with Perth in them: the suburb served, the regulator, the local condition. A national page that answers the generic question competes for the generic passage; it does not answer the Perth one.
OBSERVED Whether Perth business pages are the passages Google extracts, or whether directories and national publishers are, has now been counted for one sample. In the Perth AI Search Study 2026, AI Overviews drew 38.2% of their citations from businesses the study identified, and 70.9% counting every site that looks like a business website, while Google AI Mode drew 85.8% of its citations from Google-hosted business panels on the same questions. The two Google surfaces extract from different material, so “what Google extracts” is not one answer.
What is documented and what is inferred
- Documented: Google’s preview controls and their effect on AI formats; no special markup; AI Overviews as a core feature that cannot be turned off.
- Tested: Pew’s US click figures.
- Inferred: the passage-level rules; the emphasis distinction between AEO, GEO and AI SEO.
- Observed: the citation mix of each Google surface for 200 commercial Perth questions, in the Perth AI Search Study 2026.
- Not established: that the pattern holds beyond that sample, or any causal account of why a passage is chosen.