What Is LLM SEO? LLMO, GEO, AEO and AI SEO, With the Source of Each Term

Key takeaways

The short version of this page, in the order the sections below develop it. Each point is stated here in full so it can be read without the rest of the page, and the section that argues it gives the measurement and the source behind it.

• Jina AI says it coined LLMO in an opinion post dated December 15, 2022. GEO was introduced in a paper submitted to arXiv on 16 Nov 2023 and accepted to KDD 2024. AEO was in trade press by 7 February 2018, about voice search.
• Of the documents we opened, only GEO's is a research paper, later accepted to KDD 2024. Its headline figure, up to 40%, is a relative gain in the authors' own gpt3.5-turbo test engine, fed the top 5 Google sources.
• Google's guide names AEO and GEO and says optimizing for generative AI search is still SEO. That statement covers Google Search only.
• We found no first documented use for LLM SEO or AI SEO. Both read as descriptive labels.

What is SEO for LLMs called?

SEO for LLMs has no single agreed name: the same work is called LLM SEO, LLMO, GEO, AEO or AI SEO, depending on who is selling or describing it. All five point at one goal: getting a company named or cited in an answer generated by ChatGPT, Claude, Gemini, Perplexity or Google's AI features.

The labels do not share an origin. Three have an early use we could open and date. Two do not.

"Earliest dated use we opened" is not "first use ever". We did not search print archives or conference recordings, so an older document may exist.

Where does each term come from?

In the documents we opened, GEO comes from an academic paper, LLMO from a software company's blog post, and AEO from the voice-search years before ChatGPT existed. The three have different kinds of origin, which is part of why they are used so loosely: one was defined to be tested, one to be sold, and one described a different technology altogether. Among the LLM-era terms, LLMO has the oldest dated document we could open.

LLMO: a company blog post, December 2022

Among the LLM-era terms, LLMO has the oldest dated document we opened. Jina AI, a company that sells search models, published an opinion post titled "SEO is Dead, Long Live LLMO" dated December 15, 2022. It claims the coinage outright:

The post adds that the term "doesn't have a pre-existing definition and has been created for the purposes of this discussion." It lists three techniques: giving information directly to the creators of ChatGPT, fine-tuning the model, and in-context learning, meaning examples supplied in the prompt. None of the three is editing a web page.

GEO: a research paper, November 2023

GEO is the only one of the five whose earliest document here is a research paper, and that paper later passed peer review. The arXiv record for "GEO: Generative Engine Optimization", by Pranjal Aggarwal and five co-authors, shows it was submitted on 16 Nov 2023 and carries the comment "Accepted to KDD 2024". Its abstract says the authors "introduce Generative Engine Optimization (GEO), the first novel paradigm to aid content creators in improving their content visibility in generative engine responses".

"First" is the authors' own word for their framework. The LLMO post is older.

AEO: voice search, 2018 or earlier

AEO predates large language models as a marketing subject. A Search Engine Watch article by Rebecca Sentance, dated 07 Feb 18 on the page (7 February 2018), reports on a webinar given by Chee Lo, Head of SEO at Trustpilot, and consultant Jason Barnard. We read it in an Internet Archive capture from 13 March 2019.

It says the strategy "has come to be known as AEO", which means the term was already in use. Barnard's own site states that he "named Answer Engine Optimization in public at BrightonSEO 2017." That is his account, on his own site. We have not opened an independent document from 2017 that confirms or contradicts it.

The answer engines in that article were voice assistants and Google's single-answer results, not chat assistants.

LLM SEO and AI SEO: no coinage found

We found no document that claims to have introduced either phrase. Both are plain compounds of "SEO" and the thing being optimized for. "AI SEO" also has a second meaning: using AI tools to do ordinary SEO work.

Do LLMO, GEO and AEO mean the same thing today?

In the guidance we read, yes: Google and Forrester treat them as overlapping names, and none of the three pages below defines one term against another. What differs is the name each organization picks. Other publishers may draw lines we have not read.

• Google names two. Its guide to generative AI features, last updated 2026-07-10, says AEO and GEO are "both terms you may see used to describe work specifically focused on improving visibility in AI search experiences."
• Microsoft's Bing Webmaster team uses GEO. A February 2026 post calls its AI Performance report "an early step toward Generative Engine Optimization (GEO) tooling in Bing Webmaster Tools".
• Forrester, an analyst firm that sells research on the topic, uses AEO as the umbrella. A blog post dated Nov 14 2025 describes AEO as "also known as generative engine optimization, AI optimization, and large language model optimization".

None of the three pages uses "LLM SEO".

Is LLM SEO a real discipline?

The work is real; whether it is a discipline separate from SEO is disputed, and the party with the most direct knowledge of one product says it is not. Google's guide puts it in one sentence:

That sentence covers Google Search only, a product Google is describing as its maker. It says nothing about how ChatGPT, Claude or Perplexity choose which companies to name. The same guide sends readers considering third-party AEO or GEO services to its guidance on evaluating third-party SEO advice.

An analyst view lands in a similar place. Forrester principal analyst Nikhil Lai wrote in a post dated Nov 24 2025 that "AEO is significantly, but not fundamentally, different from SEO." In the same post he writes that the tool vendors popularizing these acronyms "tend to exaggerate SEO and AEO's differences to carve a startup-sized hole in marketers' tech stacks." This is a blog post, not a study, and Forrester sells research on the subject.

We found no standards body that defines any of the five terms, and the earliest uses we opened are a blog post, coverage of a webinar and one paper. What exists is a set of tasks, and we cover whether they pay off in is GEO worth it for a mid-size company.

Which claims about LLM SEO rest on a study?

Of the earliest documents making claims about LLM SEO, one rests on a study: the GEO paper. The LLMO post and the AEO article are a company's opinion and trade coverage of a webinar, and neither tests a tactic.

The one study needs its conditions next to its number. "Up to 40%" is the abstract's sentence about visibility, and the paper compares methods by "relative improvement in impression" on its own metrics. In the paper's full text, the answer being scored was "generated by the gpt3.5-turbo model" inside a pipeline the authors built, fed "only the top 5 sources" from Google for each query. A separate check ran on Perplexity.ai.

"Up to" marks a ceiling, not a typical result. The paper measures how much of one generated answer a source receives, not whether a current assistant names your company.

Our reading at UPAI is that a claim made under any of these five names should pass four questions before it goes into a plan:

1. Who published it, and what do they sell? A tracking vendor, an agency and a search engine each have a reason to prefer one answer. We sell this work, so ask it of us too.
2. Which engine was tested, and when? A result from a 2023 test pipeline is not a result from the assistants your buyers use today.
3. What was counted? Share of an answer, a citation link, a brand mention and a click are four different measurements.
4. Can you open the source? A figure that links to a round-up, not to the study that produced it, is unverified.

Our own S&P 100 study answers to the same test. It recorded which companies four assistants named on one date in September 2026. It did not test any optimization tactic, so it says nothing on whether LLM SEO works.

When does the choice of term matter?

The choice of term matters when you search for evidence, when you buy a service, and when you read anything written before 2023. In the guidance we read, the names are used interchangeably.

• Searching for research. The paper introduced "generative engine optimization", so that is the phrase to search for academic work.
• Buying a service. A proposal titled AEO, GEO or LLMO does not tell you what will be done. Ask for the task list and how results will be measured, per assistant. Our guide on how to know if ChatGPT mentions your brand describes a manual check to compare against.
• Reading older material. The 2018 AEO article is about voice search, not ChatGPT. LLMO in the Jina AI post meant making information known to the model, not editing a page.

Questions people ask next

These are the follow-ups we are asked most often once the answer above has landed. The same rule applies to each: where a figure appears, its source and year sit in the same sentence, and where we do not know something we say so rather than estimate it.

What is AI SEO called now?

It has no single name. Google's guide for site owners names two terms, AEO (answer engine optimization) and GEO (generative engine optimization), and says both describe work that is still SEO from Google Search's perspective. Forrester uses AEO as the umbrella name, and Microsoft's Bing Webmaster team uses GEO.

Is LLMO the same as GEO?

One analyst firm treats them as the same: a Forrester blog post from Nov 14 2025 lists large language model optimization and generative engine optimization as other names for AEO. That is analyst opinion. The earliest documents we opened differ: a Jina AI post of December 2022 says it coined LLMO for getting business information mentioned within a large language model, and a November 2023 research paper introduced GEO for improving content visibility in generative engine responses.

Who coined the term generative engine optimization?

The earliest use we opened is the paper GEO: Generative Engine Optimization, by Pranjal Aggarwal and five co-authors, whose abstract says the authors introduce Generative Engine Optimization (GEO). arXiv lists it as submitted on 16 Nov 2023 and accepted to KDD 2024.

Is SEO dead now that buyers use LLMs?

Not by Google's account of its own product. Its guide says the best practices for SEO continue to be relevant because its generative AI features are rooted in its core Search ranking and quality systems. That covers Google Search only. The SEO-is-dead line is the title of the December 2022 opinion post in which Jina AI says it coined LLMO; it reported no measurement.

Related

Other questions in this library that cover the ground next to this one. Each is answered the same way — the direct answer first, then the measurement behind it — and each carries its own sources, so you can check any of them on its own page without taking this one on trust.

• GEO vs SEO: Which Work Carries Over and Which Does Not