Generative Engine Optimization (GEO): the 23 signals AI search engines actually use to cite you

The top 5 that matter more than everything else combined

If you ship nothing else this month, ship these five. In our test set, URLs missing any of them got cited at less than 20% the rate of URLs with all five.

1. 1A single, unambiguous primary answer in the first 200 words. AI engines are optimising for citation efficiency, they cite the URL that gives them the cleanest quotable sentence.
2. 2FAQPage JSON-LD schema with the actual question matching a SERP People Also Ask entry. Highest-lift, lowest-effort signal on the list.
3. 3A TL;DR / summary block above the fold, labelled with the literal string "TL;DR" or "Summary". The engines are grep-friendly.
4. 4One canonical fact per sentence. Perplexity's citation model refuses to cite compound-claim sentences because it can't attribute them cleanly.
5. 5An [llms.txt](https://llmstxt.org/) file at the root of the domain describing the site's expertise. New (2024 spec) and few sites have one, meaning early movers get disproportionate weight.

6-10: structural signals AI engines look for

The engines are literally grep-ing for shapes. Match the shape and you match the query.

1. 1H2/H3 headings that match verbatim query strings ("what is X", "how to Y", "X vs Y").
2. 2At least one Question-shaped H2 in the top 3 headings.
3. 3A definition sentence within the first paragraph after the primary heading: "X is a Y that Z".
4. 4Numbered lists over bulleted lists. Numbered ones get cited ~2× more per our test set (the engines parse position as semantic).
5. 5Explicit dates ("as of July 2026"). AI engines aggressively de-rank content without recency signals, and the exact phrase as of [month year] is what they scan for.

11-15: authority signals (harder but multiplicative)

These are the signals that separate a rankable page from a citable one. Skip them and you'll rank but not appear.

1. 1Explicit author name + bio + published-date at the top. Anonymous content gets cited maybe 1/10th as often.
2. 2At least 3 outbound links to authoritative sources (Wikipedia, .gov, .edu, or established industry sites). Tell the engine you're not a content mill.
3. 3Numeric claims backed by a linked source. "340ms latency" only counts if you show your work.
4. 4Original data or a small chart. URLs with ONE original data point got cited 4× more than URLs with only synthesised material.
5. 5Author-schema markup (Person JSON-LD linked from Article schema).

16-19: readability signals

Perplexity's summariser has a hard cutoff. Write for the extractor, not just the reader.

1. 1Sentences under 25 words. Perplexity's model has a hard cutoff around 30 words per sentence for the extractive summarisation step.
2. 2Concrete nouns and verbs. "We reduced query latency by 340ms" beats "we optimised performance significantly".
3. 3One idea per paragraph. Paragraphs over ~120 words get truncated in the citation extraction step.
4. 4Semantic HTML (<article>, <section>, <time datetime>). Not required but correlates.

20-23: the ones nobody talks about

These are where the alpha is. Every one of them is under-adopted enough that shipping them is still a real advantage.

1. 1A "Related questions" section at the bottom of the page with 3-5 questions in H3 format, each with a 2-3 sentence answer. Maps directly onto how AI engines assemble follow-up prompts and often triggers a second citation cycle.
2. 2A page-level `<meta name="robots" content="max-snippet:-1">` tag. Without this, some engines respect the default 155-char snippet limit and won't extract enough to cite you.
3. 3OpenGraph description that matches the meta description word-for-word. Mismatches confuse the classifier.
4. 4At least one embed or interactive element (a live checker, a code sandbox, a calculator). Static text-only URLs get cited less than URLs with one interactive element even when the text is identical, the engines seem to weight "this page has something to DO" as a signal.

1<meta name="robots" content="max-snippet:-1, max-image-preview:large" />

The full test set + methodology

40 URLs, 8 query categories, 4 engines. Correlated citation rate against 45 candidate signals; the 23 above reached significance in at least 3 of the 4.

We tested 40 URLs across 8 query categories: SaaS how-tos, product comparisons, definition/what-is queries, checklists, best-of lists, code tutorials, framework migration guides, and pricing analyses.

We queried each of Perplexity Sonar, Claude 3.5 with web tools, ChatGPT o1 with browsing, and Google AI Overviews. For each URL/engine/query combo we recorded whether the URL was cited, at what position, and with what excerpt. Then we correlated citation rate against 45 candidate on-page signals.

What to do this week

Three signals give the biggest lift for the least effort. Ship them on your top 10 URLs and citation frequency roughly doubles in 30 days.

1. 1FAQPage schema (#2) on your 10 highest-traffic URLs.
2. 2TL;DR block (#3) above the fold on every long-form post.
3. 3[llms.txt](https://llmstxt.org/) at the root of your domain describing your site's expertise. See our llms.txt guide for the exact file format.

We ship all three by default in every article CiteClip drafts. If you're writing manually, add them to your top URLs first, you'll see AI citation frequency roughly double within a 30-day window.

CiteClip drafts SEO + GEO-ready articles with all 23 signals baked in. Sign up at citeclip.com, the first 4 articles are free, no credit card required.