Answer Engine Optimisation: how to be quoted by AI, not just ranked
Search used to end with a list of blue links. Increasingly it ends with an answer. ChatGPT, Perplexity, Google AI Overviews and Microsoft Copilot read pages, decide what they mean, and hand the user a summary — often quoting a handful of sources directly. Answer Engine Optimisation (AEO) is the practice of making your pages easy for those systems to read, understand and cite. It sits alongside traditional SEO rather than replacing it: SEO helps you rank, AEO helps you get quoted.
Why AEO is not the same as SEO
Classic SEO optimises for a ranking position in a list. AEO optimises for something different: being the source an AI model paraphrases or names when it answers a question. That changes what matters. An answer engine does not need your page to win a keyword race. It needs to parse your page confidently, extract a clear answer, and trust the source enough to repeat it. Pages that are cleanly structured, factually specific and unambiguous tend to be favoured, because they are cheaper for a model to interpret correctly.
What makes a page quotable
Three things do most of the work. First, structured data. JSON-LD schema tells an engine exactly what a page is — a product, a local business, an FAQ, a how-to — instead of leaving it to guess from raw HTML. Second, answer-first writing: lead each section with a direct, self-contained answer of roughly 40 to 60 words, then expand. Models extract those leading statements well. Third, trust signals: a named author with real expertise, publication and updated dates, cited sources, and accurate claims. Answer engines are conservative about which sources they name, and these signals reduce their risk.
The role of schema markup
Schema markup is the layer that removes ambiguity. Without it, an engine infers meaning from headings, layout and text, and it can get that wrong. With valid JSON-LD in place, the page states its own type and key facts — who published it, what it sells, what it costs, what questions it answers. That is why schema is the highest-leverage single change for most pages: it converts a page an engine has to interpret into one that describes itself. The catch is that the markup must match the visible content. Declaring an FAQ or an article that is not actually on the page erodes trust rather than building it.
llms.txt and the emerging AI discovery files
A newer set of conventions guides AI crawlers at the site level. An llms.txt file sits at your domain root and gives models a clean, plain-text index of your site and its most important pages. Think of it as a table of contents written for a machine with a small context window. Related files such as ai.txt and identity.json declare usage preferences and identity. None of these are guaranteed to be read by every AI provider today, so they are best treated as low-cost, forward-looking additions rather than a ranking factor. The honest position is to publish accurate, consistent files now, in case they are used, without overselling what they do.
A practical AEO checklist
- Confirm AI crawlers can reach your pages — check robots.txt allows GPTBot, PerplexityBot, ClaudeBot and Bingbot.
- Add valid JSON-LD schema for the real type of each page, and keep it matched to visible content.
- Lead sections with a direct answer, then provide detail and cite named sources.
- Show authorship, publication and updated dates, and keep content current.
- Publish an llms.txt at your root and keep it in step with your site as it changes.
Done together, these steps make a page that an answer engine can read without effort, trust without hesitation, and cite by name. That is the whole goal of AEO: not simply to appear in results, but to become the answer.