The Quiet Truth About Getting Cited by AI
A consulting industry has sprung up around "Generative Engine Optimization." The only controlled study to measure it found that three old-fashioned habits move the needle 30-40% — and that the buzziest advice does nothing.
TL;DR — "GEO" (Generative Engine Optimization) is mostly hype layered over a little real evidence. The one controlled study — GEO, KDD 2024 — found that adding statistics, quotations, and citations to reputable sources improved a page's visibility in AI answers by 30–40%. Keyword stuffing did nothing. Most of the rest ("Claude prefers long-form," "+71% from expert quotes") is vendor folklore. Write well, cite sources, and don't overthink it.
There is a familiar rhythm to the internet's anxieties. Every couple of years a fresh three-letter acronym appears, promising to decode whatever black box currently decides who gets seen. This season's letters are GEO — Generative Engine Optimization — and the box in question is the AI answer. Ask ChatGPT, Perplexity, or Google's AI Overviews a question, and a handful of sources get named in the reply. GEO is the art and emerging business of being one of them. Consultants have already materialized, decks in hand. So before anyone signs a retainer, it is worth asking the unfashionable question: what here is actually proven?
One paper against a sea of vibes
Almost all GEO advice shares a methodology, and the methodology is guessing. Someone watches what the machines spit out, notices a pattern, and turns it into a rule. The lone exception — the only piece of work that ran a real experiment — is a peer-reviewed paper called "GEO: Generative Engine Optimization" (Aggarwal et al., KDD 2024), out of Princeton, Georgia Tech, and the Allen Institute for AI. The authors built a 10,000-query benchmark, generated AI answers over genuine search results, tested nine different ways of editing a page, and measured how prominently each source surfaced.
The results were not ambiguous.
| Edit | Effect |
|---|---|
| Add quotations from credible sources | ~+41% visibility |
| Add statistics (concrete numbers vs. vague claims) | ~+34% |
| Cite sources (inline links to reputable references) | ~+30% |
| Improve fluency / readability | ~+30% |
| Keyword stuffing | negative — it hurt |
Three levers carried real, causal weight: quotations, statistics, citations. And the single most cherished trick of the old SEO underworld — packing a page with keywords — did nothing for AI engines, and if anything dragged a page down. That is the whole sturdy finding. The piece you are reading is dense with numbers, quotes, and links, and that is not an accident.
The reasonable bets
Past that one experiment, the advice thins out into observation — patterns people have spotted in the wild rather than anything they isolated and tested. Useful, but hold them loosely.
- Front-load the answer. A Search Engine Land analysis found roughly 44% of ChatGPT citations came from the first third of a page. A crisp opening answer plausibly earns its keep.
- Write headings as questions. Self-contained FAQ-style sections correlate with being cited — though the writing does the work, not the FAQ schema Google says you can skip.
- Make sentences stand alone. A claim plucked into a retrieval chunk should make sense by itself. Name the subject; don't lean on an "it" from three paragraphs back.
- Stay fresh. Recently updated pages tend to get cited more than ones gathering dust.
- Show your face. Bylines, bios, a named publisher. Google frames quality around E-E-A-T — Experience, Expertise, Authoritativeness, Trust — and AI answers lean on the same authority cues.
Where the money disappears
And then there is the folklore, sold with the confidence of fact.
- Suspiciously precise multipliers — "+71% from expert quotes," "19 statistics doubles your citations." The direction echoes the research; the exact figures trace back to no primary method anywhere. That is marketing dressed as data.
- Platform horoscopes. "Claude prefers long-form, Perplexity rewards freshness." Vendor-blog astrology, not controlled study.
- The invented-brand stunts. The much-shared test that conjured a fake brand ("Xarumei") and watched AI parrot its claims was criticized for testing content in an authority vacuum — it mostly demonstrated that when no real authority exists, answer-shaped detail wins by default.
- "GEO is a brand-new discipline." Google is blunt: optimizing for AI features is still just SEO. No
llms.txt, no mandatory "chunking," no AI-specific rewrite required.
The unglamorous to-do list
What actually works reads like advice from twenty years ago:
- Be genuinely useful and specific — original data, real numbers, firsthand experience.
- Quote experts and cite sources inline — the proven lever, and it makes the writing better regardless.
- Structure for the skimmer — headings, a TL;DR, lists and tables, self-contained sentences.
- Earn real authority — a named author, a real publisher, an actual reputation off your own site.
- Keep the plumbing clean — crawlable, fast, mostly-static HTML, since AI fetchers often see only the initial markup. Don't block the bots.
Strip away the frameworks and "GEO" turns out to be good writing plus genuine authority, with exactly one research-backed nudge worth remembering: stack your stats, quotes, and citations. The rest is someone selling you a map to a place you already know how to reach.
FAQ
Does GEO actually work?
Partly. A controlled 2024 study found that adding statistics, quotations, and citations improved AI-answer visibility by ~30–40%. Most other GEO advice is correlational or unproven, and keyword stuffing was shown to not work.
What's the difference between GEO and SEO?
Contested. Google says optimizing for AI features is still just SEO. GEO emphasizes being cited by AI answers (via stats, quotes, citations, structure) rather than ranked in a list, but the fundamentals overlap heavily.
Do I need llms.txt or special AI formatting?
No, not strictly. Google says you don't need llms.txt, content chunking, or AI-specific rewrites. They're low-cost extras some tools use — see our AI-readable blog guide — but good, authoritative, well-structured content does the heavy lifting.
Sources: GEO paper, KDD 2024 (arXiv), Google AI optimization guide, Search Engine Land, Search Engine Journal.
Image: Muhammad Rafizeldi, CC BY-SA 4.0, via Wikimedia Commons.
← Back to all posts