More and more searches never reach a list of blue links. People ask ChatGPT, Perplexity, or Google's AI Overviews and get one synthesized answer that quotes a handful of sources. Answer Engine Optimization (AEO, sometimes called GEO) is the work of becoming one of those sources. The short version: nobody can buy or guarantee a citation, but you can make a page markedly easier to retrieve, quote and trust, and that part is a concrete list of fixes. Here is the list, and the honesty most AEO content skips.
What nobody can promise
No one can guarantee a citation in ChatGPT, and anyone who says otherwise is selling something. These systems are non-deterministic — the same question can return different sources on different days — and none of them publish a ranking algorithm. You also cannot reliably track whether you were cited, because there is no public, complete log of AI answers to check against.
What you can do is make your pages measurably more likely to be retrieved and quoted, and check your readiness against the factors that demonstrably matter. That is the honest goal: readiness, not a guaranteed placement.
How a page gets into an AI answer
Simplified, there are two steps. First retrieval: the engine finds candidate sources for the question (often via a search index, so classic SEO still matters). Then synthesis: it reads those candidates and pulls the clearest, most quotable, most corroborated claims into its answer. AEO is mostly about winning the second step — being the easiest source to quote correctly.
The checklist
1. Answer the question in the first two sentences
Lead with a direct, self-contained answer before the context and the story. AI answers lift clean, standalone statements; a claim buried under three paragraphs of preamble rarely gets pulled.
2. Make claims extractable
Short declarative sentences, clear definitions, and specific numbers travel well. Structure helps too — headings phrased as the actual questions people ask, plus lists and tables the model can lift cleanly.
3. Be clear about who you are (entities and schema)
Use structured data (Organization, Article, FAQ, Product) so machines can resolve what your page is about and who stands behind it. Consistent naming across your site, your profiles, and the wider web helps engines trust that the entity is real.
4. Get corroborated elsewhere
AI answers favour claims that more than one credible source agrees on. Mentions, references, and consistent facts across the web make your version the safe one to quote. This is the slowest lever and the strongest.
5. Stay fresh and factual
Current dates, correct figures, and recently updated pages read as more reliable. Wrong or stale facts are the fastest way to get skipped — engines route around sources they cannot trust. This is content decay seen from the AI side: the same drift that costs you rankings also costs you citations.
6. Let the AI crawlers in
Open your robots.txt and check that you are not blocking the crawlers these engines actually use: GPTBot and OAI-SearchBot (OpenAI), PerplexityBot, ClaudeBot, Google-Extended. One thing people get wrong here: Google-Extended does not control AI Overviews. Those follow your ordinary Googlebot rules, so blocking Google-Extended will not keep you out of them.
Then do the five-second version of the same check on a page that matters: open view-source, search for a sentence from your main answer, and see whether it is there. If the text only appears once JavaScript has run, a crawler that does not render will never see it.
How to measure it honestly
Since you cannot reliably track live citations, measure readiness instead: does the page answer up top, are the claims extractable, is the schema there, is it current, can a crawler read it? That turns a vague ambition into a list you can work through. It is also how Tedrix scores your pages — a readiness check bundled with the monthly decay work, not the live citation tracker nobody can honestly build. Check your own pages for free.