seo

SEO, AEO, and GEO: what actually changed

Three acronyms, one page. Here is what answer engines and generative search changed about optimization, and what they did not.

Most of what you already know about SEO still works. Pages need to be crawlable, HTML needs to be clean, pages need to load fast, and the content on them needs to say something real. None of that got old.

What changed is who is reading your page after the crawler finds it, and what they do with what they read. A search engine used to hand a person a list of links and let them pick one. Now a growing share of that work happens before the person ever sees a link — an assistant reads several pages, decides what’s true, and writes an answer. That shift is what SEO, AEO, and GEO are each trying to describe, and it’s worth being precise about which one you mean.

What each acronym actually means

SEO is about ranking a link in a list of links. The unit is the search results page, the goal is a high position on it, and the reader is a person scanning ten blue links and deciding where to click.

AEO — Answer Engine Optimization — is about being the source a direct answer is extracted from. When someone asks a voice assistant or a search engine’s answer box a question, something on your page gets lifted out, verbatim or close to it, and shown without a click. The reader here isn’t a person browsing results; it’s an extraction system deciding which paragraph on your page best answers one specific question.

GEO — Generative Engine Optimization, sometimes called generative visibility — is about being cited, quoted, or recommended inside a longer piece of generated prose. A chatbot answering “what’s the best crawler for AEO scoring” doesn’t lift one sentence from your page the way an answer box does. It reads your page alongside several others, synthesizes a paragraph, and either mentions you by name or doesn’t.

Same page, three different consumers: a person scanning a results list, an extraction system pulling one answer, and a generation system deciding whether you’re worth citing at all. A page can do well on one and badly on another, because the thing being evaluated isn’t the same thing.

The unit of success moved from the click to the citation

For most of SEO’s history, ranking #1 was the goal because ranking #1 got the click. That link was the outcome. If your position on the results page was good, your traffic was good, and you could draw a straight line between the two.

That line isn’t straight anymore. A page can rank #1 for a question and get a fraction of the clicks it would have gotten five years ago, because the answer already sat above it — pulled into an answer box or spoken by an assistant — before anyone needed to click through. This is the “zero-click” search result: the query gets answered on the results page itself, or inside a chat interface, and the underlying page’s job was to supply the answer, not to receive a visit.

That doesn’t mean the page did nothing. It supplied the words that got used. But if you’re measuring success by clicks and rankings alone, a page can be working exactly as intended — getting read, getting extracted, getting cited — while your click-through numbers tell you it’s failing. The click was always a proxy for “this page was useful.” It’s a proxy that’s gotten noisier, not one that’s stopped meaning anything.

What answer engines reward that ranking never did

Classic ranking factors — backlinks, keyword relevance, page speed, domain authority — still matter, because an answer engine still has to find and trust your page before it can extract anything from it. But extraction rewards a few things that ranking algorithms were mostly indifferent to:

Self-contained passages. A paragraph that makes sense on its own, without needing the three paragraphs before it for context, is a paragraph that can be lifted cleanly.

A direct answer placed near the top of a section. Not buried under three paragraphs of preamble — stated plainly, close to the heading that asks the question.

Unambiguous entities. If your page talks about “the platform” for six paragraphs before naming which platform, an extraction system has more work to do than it usually will.

Structured data that agrees with the visible text. A schema block that says one thing while the page says another isn’t neutral — it’s a contradiction, and it gets treated as one.

Facts stated once and clearly, rather than implied across several sentences or hedged into vagueness. A page that says “usually around 4GB” is less useful to lift than one that says “4GB minimum.”

None of these were ever penalized by classic SEO, exactly — but ranking algorithms didn’t need any of them to put a page in position #1. Extraction does.

What did not change

It’s worth being blunt about what’s the same, because a fair amount of the AEO and GEO conversation implies a clean break from everything that came before, and there wasn’t one.

Crawl access still matters — if a bot can’t reach your page, nothing about the page’s quality is relevant. Response speed still matters, for the same reason it always did: slow pages get deprioritized and impatient readers leave. Unique substance still matters — a page that says nothing a hundred other pages don’t already say has no reason to be the one that gets cited or extracted. Internal links still matter, because they’re how both crawlers and readers understand which pages on your site relate to each other. And accurate titles still matter — a title that lies about the content underneath it hurts you the same way it always did, whether the reader is a person or a system extracting an answer.

Anyone selling you a clean break between “old SEO” and “the AEO/GEO era” is selling you something. What actually happened is additive: the old foundation is still load-bearing, and a new layer got built on top of it with its own separate requirements.

How to tell where you actually stand

All of this is straightforward to describe in the abstract and hard to check in practice. You can read one page and judge, by eye, whether it has a direct answer near the top or whether its schema agrees with its text. You cannot do that across a hundred pages on a real site, consistently, by hand — which is the problem GEOBee exists to solve: it crawls a site and scores every URL across SEO, AEO, and GEO as three separate numbers, so you can see where a page is strong on one lens and quietly weak on another. If you want the exact signals behind each score, the three lenses breaks down what’s measured and how it’s weighted.

On this page