Answer Engine Optimisation (AEO)

The phrase has two halves and both matter. Answer engine is an editorial classification: a search surface whose primary output is a generated answer with sources, not a ranked list of blue links. Google AI Overviews is an answer engine. ChatGPT with web search is an answer engine. Perplexity is an answer engine. Classical Google's ten blue links is a search engine. Optimisation is the deliberate restructuring of content to fit how an answer engine retrieves and quotes — not how a classical ranker scores. The two overlap heavily in 2026 (rankings still feed retrieval), but the optimisation targets differ.

AEO is sometimes spelled "Answer Engine Optimization" in US-English sources and "Answer Engine Optimisation" in UK-English sources¹. The acronym is identical. We use the UK spelling on this site because the team and the founder are based in Germany, where UK English is the default editorial convention.

The mechanical pattern that does most of the work: every H2 on the page reads as a question or a question-shaped statement, and the first 134-167 words after that H2 answer it directly in plain English. AI engines extract these passages as quotable units. The remaining word count expands the answer, links out, and provides the longer-form context a human reader needs — but the quotable unit is the lead. Without it, the engine extracts nothing or extracts a less-useful fragment from later in the page.

Beyond the passage discipline, AEO leans on classical-SEO foundations that AI engines still respect: a clean robots.txt that allows retrieval bots, a valid XML sitemap, an indexed canonical URL per topic, and meta tags that match the page's actual content. AI engines are not magic readers; they are crawlers plus retrieval-augmented generation, and the crawler half still cares about the basics.

The 2026 reframing is mostly substitution. Where the 2018-era AEO playbook said "structure your page to win the featured snippet," the 2026 playbook says "structure your page so AI Overviews and Perplexity cite you." The underlying tactics — concise lead passages, question-shaped headings, structured data, entity clarity — are largely the same. The audience of consumers (the engines) has multiplied from one or two to more than eight.

A practical sequence: open Settings → Crawlers and confirm the 26 AI bots are toggled appropriately for your goal (allow retrieval, decide on training); add the Article and Person JSON-LD via Site Header Code Injection so every editorial page inherits author and publisher entities; on each pillar/cluster page, paste the page-specific structured data (FAQPage on the FAQ hub, LocalBusiness on local pages, Service on the install page) into Page Settings → Advanced → Page Header Code Injection; finally, audit each page's heading hierarchy in DevTools and Code-Inject a small DOM-rewrite snippet where 7.1 emits H1s in section headers that should be H2s.

Practically: a page can be optimised for all three and frequently should be. The same Article schema feeds Google's classical ranker, the AI Overviews retrieval pipeline, and ChatGPT's web-search index. The same lead-passage discipline that wins a featured snippet also wins a Perplexity citation. The reason to keep the terms distinct is diagnostic, not strategic — when a page ranks but does not get cited, the gap is usually AEO; when a page gets cited but for the wrong claim, the gap is usually GEO entity disambiguation. See GEO for the broader treatment.

• GEO — Generative Engine Optimisation: the broader discipline that includes AEO.
• AI Overviews: the Google surface where most public AEO outcomes are measured.
• Passage ranking: the Google retrieval pattern that AI engines also use to extract from longer pages.
• Entity SEO: the disambiguation work behind GEO that AEO inherits.
• E-E-A-T: the experience/expertise framework that AI engines weight when deciding whether to quote.
• Pillar 1 — Squarespace × AI Search: the install-level playbook.