Most of these words mean the same thing: getting your pages mentioned when someone asks an AI assistant a question. One of them came out of an academic paper and has a specific meaning. The rest were coined by companies that needed a category to sell into, which does not make them useless, but does mean you should not pay extra for the vocabulary.

This is the sister piece to why AI visibility is not SEO, which covers what actually differs. This one is only about the words, so you can read the rest of this subject without being sold to by its glossary.

If you want the thing itself rather than its names, start here instead.

The words, and where each came from

Term	Where it came from	Does it mean something different?
GEO, generative engine optimisation	A 2024 academic paper	Yes, originally
AEO, answer engine optimisation	Marketing, reusing an older phrase	No
AI visibility	Tool vendors, including us	No
LLMO, large language model optimisation	Marketing	No
AI SEO / SEO for AI	SEO agencies	No
AIO, GAIO, and others	Marketing	No

Six names, one subject. Anyone using them as if they were different services is either confused or counting on you being.

GEO is the one with a real origin

Generative engine optimisation comes from a paper called exactly that, published at KDD 2024 by researchers at Princeton, Georgia Tech, IIT Delhi and the Allen Institute for AI. That is unusual in this field and worth knowing, because it means there is something to read rather than a vendor blog to trust.

The paper built a benchmark of queries, tried various changes to source pages, and measured which changes got those pages cited more often in generated answers. It reports visibility gains of up to 40% from the strategies it tested, with the effect varying by domain.

Two caveats before you go and act on it. The tactics it found effective, such as adding statistics, quotations and citations, were tested on general informational queries rather than on shop product pages, and a hiking boot is not an encyclopaedia article. And the work dates from 2023 and 2024, which in this field is a long time ago.

It is also, plausibly, part of why so much writing published since then arrives stuffed with statistics and authoritative-sounding quotations. A paper found that those got cited more, and an industry read the summary.

AEO and the rest are the same thing with a different badge

Answer engine optimisation reuses "answer engine", a phrase that predates this entire wave and used to describe things like a calculator widget in a search result. Applied here it means what GEO means.

AI visibility is the phrase most tools use, and it is the one we use too, so we should be straight about it. We mean how often your shop gets named when a buyer's question is put to an assistant. That is a frequency, not a ranking, and it moves on its own between runs. We picked the phrase because a shop owner can guess what it means, which is more than can be said for the acronyms.

The rest, LLMO, AIO, GAIO and whatever launches next month, are coinages looking for a category. You can safely not learn them.

What the companies themselves call it

Nothing. This is the part worth noticing.

Google calls them "AI features" in Search and writes about AI Overviews and AI Mode as parts of Search. OpenAI writes about ChatGPT's "search features". Neither has a name for the discipline of optimising for them, because from where they sit there is no separate discipline to name.

That is not proof the field is fake. It is a useful check on anyone claiming their new methodology is what these systems require, when the companies building them describe no such requirement.

What Google actually says you need to do

This is the most useful quotation in this whole subject, and it is from Google's own documentation on AI features:

There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.

And, in the same place:

You don't need to create new machine readable files, AI text files, or markup to appear in these features.

Read that carefully, because it is narrower than "none of this matters". Google is saying that appearing in its AI features needs the things appearing in Search already needs: be crawlable, be indexable, have your structured data in order, have pages worth citing. What it rules out is the idea that there is a separate checklist, a special file, or a new markup you are missing.

It also speaks only for Google. It is not a statement about ChatGPT or Perplexity, and the parts of this subject those systems make different, such as whether a page needs JavaScript to render its price, are not covered by it.

The llms.txt question

Since Google mentioned "AI text files", this is the one everybody asks about.

llms.txt is a proposed file you would put at the root of your site, summarising your content for language models. It is a sensible-looking idea with an obvious analogy to robots.txt, and it has been adopted enthusiastically by people publishing files and much less so by the systems that were supposed to read them.

Where it stands: Google's guidance above rules it out for AI Overviews and AI Mode in as many words. OpenAI's crawler documentation describes robots.txt and does not mention llms.txt anywhere. Some tools in the wider ecosystem do use it, mostly for developer documentation rather than shopping.

So the honest position is that it costs almost nothing to publish and there is currently no evidence it will get your products recommended. Publish one if you like. The mistake is treating it as done work while your product pages are still missing a currency.

Does any of this vocabulary matter?

In two places, and neither is the writing.

When you are buying. If an agency sells "GEO" as a separate service to SEO at a separate price, ask what they do that their SEO service does not. There are real answers to that question. There are also agencies who will not have one.

When you are reading. Knowing GEO points at a real paper lets you go and read it instead of the fourteenth blog summarising it. Knowing the other five terms are interchangeable lets you skip five articles that say the same thing.

Otherwise it does not matter. A shop owner never has to use any of these words, and no assistant has ever recommended a shop because its owner knew what AEO stood for.

Questions people ask

Which term should I search for when I want to learn more?

"GEO" and "AI visibility" will find you the most, in English. In Polish the field is much emptier, which cuts both ways: less noise to wade through and fewer people who have done the work.

Is this a real discipline or a fad?

The problem is real: people do ask assistants what to buy, and shops do get left out. The discipline around it is young, partly faddish, and currently sold with more confidence than the evidence supports. Both things are true at once, and anyone telling you only one of them is selling something.

Should my job title change?

No. If you run a shop, this is a thing your product pages either do or do not do. It is not a department.

Someone quoted me a 40% improvement figure. Is that real?

It is from the GEO paper, so it is a real result from real research. It is also about general informational content rather than product pages, and it is now around two years old. Treat it as evidence that content changes can move citations, not as a number you should expect on your shop.

What about "AI SEO"?

The same subject again. The phrase is at least honest about the overlap, which the piece on SEO goes through properly.

Where we come into it

We call what we do an AI visibility audit, which puts us in the vocabulary above, and the only defence of that is being specific about what we mean by it. We check whether crawlers can reach your products, whether your structured data says what it should, and whether your pages answer the questions your buyers ask. Then we put those questions to ChatGPT and record who gets named. The number we report is how often you appear, which is a frequency and not a rank.

We explain how to fix what we find. We do not touch your shop, we do not test every assistant, and we have not invented a methodology.

We are not open yet. Join the waitlist and we will tell you when we are.

AI visibility, GEO, AEO: which of these words matter