What you actually do to get recommended by answer engines

For two decades, online visibility meant one thing: ranking high in Google. I have lived inside that world and chased it for years. Today a growing share of users never click a link at all. They ask an answer engine (AI) directly and receive a finished answer, with two or three sources cited inside it. Be honest: when did you last open ten tabs instead?

GEO (Generative Engine Optimization) is the practice of making sure you are one of those sources. And no, this is not another impressive acronym invented by an agency that needed something new to sell you: the term was established in an academic paper presented at KDD 2024 by researchers from Princeton, Georgia Tech and IIT Delhi.

So what actually changes here

A conventional search engine hands you a list of links and then steps aside; you decide where to click. A generative engine does something different. It retrieves sources, reads them, synthesizes them, and produces a single answer with citations embedded in the text. Think of a librarian who no longer points at the shelf, but puts the summary on the table in front of you.

The difference matters, and not only to the technically minded. In a list of links, position is everything. Inside a synthesized answer, the question changes at the root: are you cited at all, and if so, how much of the answer comes from you?

The GEO researchers put it precisely. Average ranking is a fine measure of visibility for an engine that presents a linear list of websites, but it says nothing about an answer where sources are interleaved, at different lengths and different positions. So they proposed new metrics: how many words of the answer come from your source, and how early in the answer you appear.

Two words on how it relates to SEO, then we move on

This is the point that gets misunderstood more than any other, so let us say it plainly: GEO does not replace SEO, it stands on top of it. Generative engines do not read the web from scratch for every question, they are not that romantic. They retrieve sources from an index, usually a conventional search engine’s, and then synthesise. If your site is not fast, reachable by crawlers (the programs that read pages on behalf of an engine) and indexed, there is nothing there to be cited. Simple as that.

The comparison of the two, with a table and with what changes in goal, measurement and competition, we keep separately: what the difference between SEO and GEO actually is. Here we stay on the practical side.

And what does the research actually say?

The paper “GEO: Generative Engine Optimization” (Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan, Deshpande) was presented at KDD 2024 and is the first systematic study of the subject. The researchers built GEO-bench, a benchmark of 10,000 queries across diverse domains, and tested optimization methods against it. So this is not a hunch dressed up as a method.

Their finding: the proposed methods can boost visibility by up to 40%. Specifically, they found that including citations, quotations from relevant sources, and statistics significantly boosts source visibility across various queries.

And they did not stop at the lab, that comfortable place where everything always works. They applied the same techniques to Perplexity.ai, a real, deployed generative engine, and measured visibility improvements of up to 37%.

Two further points from the paper deserve attention, arguably more than the percentages do. First, the effectiveness of each technique varies by domain, so there is no single recipe. Second, the researchers emphasize that generative engines are black boxes: content creators have little control over when and how their material is displayed. I hold on to that second one tightly, because it is exactly why you will never hear me promise you a guaranteed slot in anybody’s answer.

Seven things you can do starting tomorrow

1. Answer the question in the first two sentences. Models split content into chunks and evaluate each one separately. If your answer is buried in the sixth paragraph, the chunk that gets read may not contain it at all. Define the thing first, then elaborate as much as you like.

2. Use numbers. Not “a significant improvement” but “from 19 seconds to 4.4 seconds”. “It depends” is the worst answer you can give a model, because it leaves absolutely nothing to quote. Specific, verifiable figures get cited.

3. Cite your sources. When you claim something, show where you learned it. Academic papers, government data, institutional reports. The model recognizes the chain of credibility.

4. Structure the text. Clear headings that contain the question, short paragraphs, tables, lists. A passage that stands on its own is a passage that can be quoted on its own.

5. Ship structured data, code that describes the page to the engines. In Schema.org format, with types such as Organization, Article, FAQPage, BreadcrumbList. You hand the engine a machine-readable description of what the page is, instead of leaving it to guess. And anything that guesses will, sooner or later, guess wrong.

6. Keep the site crawlable without JavaScript. The crawlers behind these tools generally do not execute JavaScript, the code that builds part of the page in the browser. If your content only appears after that code runs, the page is empty as far as they are concerned. Pages that arrive ready-made from the server (static HTML), clean URLs, fast responses.

7. Refresh. A 2024 guide that was never updated loses ground to a 2026 article on the same subject. Review your pillar pages quarterly, with real changes and a visible update date.

Three misconceptions that cost real money

“I will stuff the page with keywords and artificial intelligence will read it.” Exactly the opposite. Keyword stuffing worked on engines that matched words. A language model understands meaning, and padded text with no substance gives it nothing to quote. Information density matters, not keyword density.

“Google finds me, so ChatGPT will find me.” Necessary, not sufficient. Being indexed makes you a candidate. Whether the model picks you from among ten available sources depends on how easily a clear, evidenced answer can be extracted from your text.

“GEO is a one-off job.” Generative engines weigh freshness. An article that never changed, with a modified date identical to its publication date, signals that nobody is tending it. And you know exactly what a neglected page feels like when you land on one…

Doing this from a small market

Two things are true at once for a business outside the English-speaking web. First, competition in your own language is far thinner: the sources a model has available for a question asked in Greek, Portuguese or Finnish are few, and often poor. There is room. Not many people say that out loud.

Second, your presence off your own site matters more than you think. Models weigh corroboration from independent sources. A flawless site with no mentions anywhere else only answers people who already know your name. A Google Business Profile, consistent details across directories, reviews, one article in a trade publication: this is the credibility infrastructure that on-site GEO stands on.

A site that is already sound starts ahead

The fundamentals stay exactly where they were: technical soundness, speed, quality content, authority. What changes is the shape the content has to take so that a machine can read it, trust it and quote it.

A site that is not well built now has two reasons to be fixed instead of one. And the second reason, to put it plainly, is not going away.

How will you know it is working?

There is no Search Console for generative engines yet. I know, it would be convenient. The practical method is manual and, for all its apparent naivety, it works: build a fixed set of 10 to 15 questions a prospective customer would actually ask, put them to ChatGPT, Perplexity and Gemini every month, and record whether you are cited, in what position, and in what words.

The curve you see after a few months tells you more than any tool.

Let us finish where we started

Search is not disappearing. It is changing shape. From a list where the user chooses, into an answer where a model chooses. The librarian from the opening no longer points at the shelf, but there are still two or three books in his hands, and the whole game is making sure one of them is yours. GEO is the work that makes your content worth choosing: specific, structured, evidenced and accessible.

And as the research shows, this is not theory. It is a measurable difference, on the order of 40%, for anyone willing to do the work. That last part, willing to do the work, is as always the hard word in the sentence.

If you want your own content to be one of the books in the librarian’s hands, that is a project we take on: we measure where you appear today, fix what is needed in structure and content, and measure again so the change is visible. What you get is concrete deliverables and repeated measurements, not a guaranteed position, because nobody serious offers one. Send me through the form your site address and the questions you would like to be recommended for. This work is part of our SEO and GEO service.

The whole picture: this article covers one part of a larger subject. If you are after the whole of it, from beginning to end, it is in the guide to answer engines.