How to Get Cited by ChatGPT, Perplexity and AI Overviews

To get cited by ChatGPT and AI search tools like Perplexity and Google AI Overviews, you need pages that answer a specific question in the first two sentences, back that answer with a named source or a concrete number, and organize the rest of the content into scannable sections an AI system can lift without guessing at meaning. None of the three systems require special markup to cite a page: Google's own documentation states there are "no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary," beyond being indexed and eligible to show a normal search snippet (Google Search Central). ChatGPT and Perplexity run their own retrieval layers on top of the open web, but they reward the same underlying signals: clarity, authority, and freshness.

Photo: kevin dooley · BY

What "Getting Cited by AI Search" Actually Means

"AI search" covers three distinct systems that founders usually lump together, and they don't work the same way. Google AI Overviews is a summary generated above traditional results, pulling from pages already eligible to rank. ChatGPT's search mode and Perplexity are separate products: both run a live retrieval step against the web, then generate an answer with inline citations, independent of whether you rank in Google at all.

Generative Engine Optimization (GEO) is the umbrella term researchers at Princeton, Georgia Tech, and IIT Delhi coined for optimizing content to be selected and cited by these generative answer engines, as distinct from classic SEO's goal of ranking a blue link (arXiv:2311.09735). Their study introduced GEO-bench, a benchmark of real user queries, and found that applying specific content-level changes — the paper reports this can boost visibility in generative engine responses by up to 40% — moves the needle more than backlinks or domain age, though the paper notes effectiveness varies a lot by topic, so no single tactic works everywhere.

The practical takeaway: you're not optimizing for one algorithm anymore. You're writing pages that a retrieval system can quote cleanly, in three different products, each with its own selection logic.

How Google AI Overviews Choose What to Cite

The biggest misconception is that ranking #1 guarantees an AI Overview citation. It doesn't, and the gap between ranking and citation is widening. A large-scale study by Ahrefs, analyzing 863,000 keyword SERPs and roughly 4 million AI Overview URLs in early 2026, found that only about 38% of pages cited in AI Overviews also ranked in the traditional top 10 for that query — down from 76% just seven months earlier (Ahrefs). The rest split almost evenly between pages ranking 11–100 and pages that didn't appear in the top 100 at all.

Ahrefs attributes the shift to Google increasingly generating "fan-out" sub-queries behind the scenes rather than summarizing only the exact query's top results, which means AI Overviews now draw citations from a much wider slice of the index than the page you're used to competing on. In practice, that means a well-targeted long-tail page with a direct, quotable answer can get cited even if it never cracks the top 10 for the head term.

Google's guidance for site owners is deliberately unglamorous. Structured data isn't required, but the company does say it "uses structured data that it finds on the web to understand the content of the page," and calls out Article, FAQ, and HowTo markup specifically as machine-readable signals any AI system can use to classify a page's content, even though they don't guarantee inclusion (Google Search Central). Markup helps a system understand your page faster; it doesn't replace having the answer on the page.

How ChatGPT and Perplexity Choose Sources

ChatGPT and Perplexity don't consult Google's index directly — they run their own web retrieval and ranking passes, then generate an answer constrained to what they pulled back. That means a page can be invisible to Google's AI Overviews and still get cited in a ChatGPT or Perplexity answer, or vice versa.

Both systems' citation behavior converges on the same handful of signals: does the page directly answer the query, is the claim traceable to a specific source or number, is the content recent enough to trust, and does the domain carry some independent signal of authority (an author, an about page, other citations of that domain elsewhere on the web). Perplexity goes further and visibly labels some cited domains with shield icons for "Government," "Academic," or "Trusted" sources, signaling that its own source-quality review process sits upstream of citation, not just relevance matching.

The practical implication is the same one Google's guidance points to: write the specific, sourced answer once, structure it so a retrieval system can isolate it as a standalone claim, and let each engine's own crawler and ranker do the rest. You cannot submit a page to Perplexity or ChatGPT for indexing — you can only make it easy to retrieve when the moment comes.

Comparing the Three Systems

The GEO Playbook: What Actually Moves the Needle

Based on how all three systems select sources, the changes worth making to an existing page are structural, not cosmetic:

Lead with the answer, not the setup. Every system above rewards a page that states its core claim as a self-contained sentence in the first two or three lines. That's also the sentence that gets lifted verbatim into a summary — write it so it reads correctly out of context.

Attach a number or a named source to every claim you want quoted. "AI Overviews reduce clicks significantly" is not citable. "Ahrefs measured a 58% CTR drop on the top-ranking page for keywords with an AI Overview in February 2026" is a sentence a generative engine can lift and attribute. If you can't point to where a number came from, don't use the number — state the point qualitatively instead.

Define entities on first use. If your page mentions "GEO," "AI Overviews," or "RAG," define the term the first time it appears in one clause. Retrieval systems assemble answers from fragments pulled out of context; a fragment that assumes prior knowledge doesn't survive extraction as cleanly as one that's self-contained.

Use FAQ and HowTo structure where it's genuinely a Q&A or a process. This isn't about gaming rich snippets — Google explicitly says the markup itself confers no ranking or citation boost — but a clearly labeled question-and-answer block is easier for any retrieval system to isolate and quote than the same content buried in a paragraph.

Go after the long tail, not just the head term. Since AI Overviews increasingly cite pages outside the traditional top 10 via fan-out queries, a page that thoroughly answers a specific sub-question has a real shot at citation even without a head-term ranking. This is the same logic behind building long-tail keyword targets for a new site — narrow, specific, answerable pages compound faster than broad ones.

Keep the surrounding site coherent. A page that's internally linked into a cluster of related, topically consistent content gives every crawler — Google's, OpenAI's, Perplexity's — more confidence in the domain's authority on the subject. A basic internal linking system does this cheaply.

Don't fake expertise you don't have. Both AI Overviews and Perplexity's trust labels reward apparent authority — but the underlying content still has to hold up. This is the same standard covered in what Google actually says about AI-generated content and SEO: the format of authorship isn't the issue, unverified or low-effort claims are.

Where This Overlaps With Losing Clicks

Getting cited is only half the story — Ahrefs' same body of research has separately tracked click-through collapse on pages that do get cited or outranked by an AI Overview, which is worth reading before you decide how much editorial time to spend chasing citations versus traffic-driving formats. The short version, covered in more depth in how AI Overviews are changing blog traffic, is that citation and click-through are now two separate goals, and a page can win one without winning the other.

Full Disclosure: We Build a Tool for This

Rankody is an AI SEO tool that researches a keyword, writes a long-form article with sourced claims and citations attached, and publishes it on a schedule — full disclosure: we build one, and this article was produced by that same process. The reason this matters here specifically: GEO isn't a plugin or a checkbox, it's a habit of writing verifiable, well-sourced, answer-first content on every single post, which is tedious to enforce by hand across dozens of articles. Rankody's $49/month plan runs a fact-check pass with visible corrections and cited sources on every article it drafts, precisely because unsourced claims don't get cited by anything. If you're comparing tools in this category rather than building the habit manually, our comparison of AI SEO writer alternatives covers where the tradeoffs actually are.

FAQ

Does adding FAQ or HowTo schema guarantee my page gets cited by AI Overviews? No. Google's own documentation states there's no special schema required to appear in AI Overviews or AI Mode. Structured data helps machines parse what a section of your page is about, which can make it easier to extract, but it doesn't replace having a direct, well-sourced answer on the page.

Can a page rank outside Google's top 10 and still get cited by an AI Overview? Yes, and increasingly so. Ahrefs found only about 38% of AI Overview citations came from top-10 organic pages in its early-2026 dataset, down from 76% seven months prior, because Google now draws citations from a wider set of fan-out sub-queries rather than only the exact search term's top results.

Do ChatGPT and Perplexity use Google's search index to decide what to cite? No. Both run their own live web retrieval and ranking systems independent of Google's search index, which is why a page can be cited in one system and absent from another for the same query.

Is GEO a replacement for SEO? No — the underlying signals largely overlap (authority, freshness, clear answers, indexability), but GEO specifically optimizes for being selected and quoted inside a generated answer rather than for ranking a clickable link, and the two goals can pull in different directions when a citation satisfies the user without a click.

What's the single highest-leverage change to make first? Rewrite your opening paragraph so the first two sentences state the direct answer with a specific number or named source, and make sure that paragraph reads correctly with zero surrounding context — that's the exact unit most retrieval systems extract first.

Getting cited by AI search isn't a separate discipline from good writing — it's what happens when every claim on the page is specific, sourced, and stated plainly enough to survive being lifted out of context.

Rankody researches your market, writes sourced long-form articles, and publishes on schedule — every piece waits for your approval first. Three articles free, no card. Analyze my site