The vocabulary, defined plainly.
AI search has collected a lot of acronyms in a short time, several of which mean the same thing. These are the terms that appear on this site, with the definitions Lantad uses when it uses them.
What the words mean here
Ordered by how central they are to a Lantad report rather than alphabetically.
| Term | Definition | Why it matters |
|---|---|---|
| Prose Parity | The share of your rendered, human-visible content that ships in the initial HTML an AI crawler reads. | Half the AI Visibility Score. No other tool reports it, and it is the measure that explains most absences from AI answers. |
| AI Visibility Score | One number from 0 to 100: 50% parity, 25% crawler access, 15% structure, 10% structured data. | A single figure you can track across deploys, with the working published on the methodology page. |
| GEO | Generative engine optimization: making a site readable and citable by AI answer engines. | Overlaps with SEO without being the same work, because Google renders JavaScript and most AI crawlers do not. |
| AEO | Answer engine optimization: the same job as GEO, under a different acronym. | Used interchangeably in the industry. Nothing turns on which one you say. |
| AI crawler | A bot that fetches pages on behalf of an AI system, for training, for search, or on a user's request. | Most run no JavaScript, which is the single fact the whole product is built around. |
| robots.txt | A file at your site root stating which crawlers may fetch which paths. | Honoured today by every crawler that behaves, and the only access control most sites have. |
| llms.txt | A proposed Markdown convention: a map at your site root pointing AI systems at your best pages. | No major vendor has committed to reading it. Cheap to serve, and this site will not claim more than that. |
| JSON-LD | Structured data embedded in a page describing what it is, in a machine-readable form. | Scored separately from parity, at 10%, and capped hard when it is injected after render rather than shipped in the source. |
| Rendered vs raw | The page after JavaScript has run, versus the HTML the server returned. | The gap between the two is the whole measurement. Google reads the first; most AI crawlers read the second. |
| Rule-only token | A robots.txt name a vendor reads as a switch, with no crawler user agent behind it. | 3 of the 15 tokens work this way, so they can be honoured but never probed. |
Every term
The full list ships in this page's HTML. There is no filter to click and nothing to expand: a crawler reads the same catalogue a visitor does.
-
Generative Engine Optimization (GEO)
Generative Engine Optimization (GEO) is the practice of structuring and publishing content so that generative AI engines can fetch it, read it in full, and cite it when they answer a question.
/glossary/geo -
Answer Engine Optimization (AEO)
Answer Engine Optimization (AEO) is the practice of making a page readable, allowed, and structured enough that AI answer engines can use it as a source when they compose an answer.
/glossary/aeo -
AI visibility
AI visibility is whether AI systems can access, read, and faithfully represent your web pages, the prerequisite for being cited in an AI answer.
/glossary/ai-visibility -
llms.txt
llms.txt is a proposed plain-text markdown file placed at a site's root that gives AI systems a curated map of its most important content, so models can find and read what matters without parsing every page.
/glossary/llms-txt -
Prose Parity
Prose Parity is Lantad's metric for the share of a page's rendered, human-visible text that is present in the raw HTML an AI crawler receives, where 100% means a crawler sees everything a human visitor sees.
/glossary/prose-parity -
AI crawler
An AI crawler is an automated bot, such as GPTBot, ClaudeBot, or PerplexityBot, that fetches web pages so an AI system can read, train on, or cite what those pages contain.
/glossary/ai-crawler -
Structured data for AI (schema markup)
Structured data is schema.org markup, usually written as JSON-LD, that labels a page's facts in a machine-readable format so AI systems can extract and attribute them without having to interpret the surrounding prose.
/glossary/structured-data -
Entity confidence
Entity confidence is a free diagnostic on every graded Lantad report that scores, out of 100, how completely the JSON-LD in your crawler-visible HTML declares one identity and how far that identity is anchored by identifiers which resolve outside the page.
/glossary/entity-confidence
How to read these definitions
Each entry's first sentence is written to stand alone, with no pronoun that needs the sentence before it. That is deliberate: these are the lines an assistant lifts, and a definition that only makes sense in place is a definition that gets quoted wrong.
Where a term names something Lantad measures, the entry says how it is measured and links to the methodology, so a reader can check the definition against the code that implements it rather than taking it on trust.
Pairs people mix up
Four confusions that come up in almost every conversation about this subject.
Indexed is not readable
Google can index a page it renders. An AI crawler that runs no JavaScript may receive almost nothing from that same page. Ranking well is not evidence of being readable.
Allowed is not reachable
robots.txt allowing a crawler does not mean your CDN lets it through. The two disagree often enough that Lantad probes rather than trusting the file.
Mentioned is not cited
A model naming your brand from training data is different from an answer engine fetching and quoting your page. Only the second depends on readability.
Training is not search
A training crawler collects content for future models. A search crawler fetches a page to answer a question now. A site can reasonably allow one and block the other.
GEO and AEO are the same job under two acronyms. Generative engine optimization and answer engine optimization are used interchangeably across the industry, and both describe the work of being readable and therefore citable by AI answer engines rather than only by search engines.
Where a term is contested or a convention is unratified, this page says so rather than picking a side quietly.
Common questions
Is GEO just SEO with a new name?
No, though they overlap. Classic SEO optimises for engines that render JavaScript and rank links. GEO starts one step earlier, with whether a non-rendering crawler receives your content at all, which is a technical question SEO tools do not ask.
Should I say GEO or AEO?
Either. They are used interchangeably across the industry and nothing practical turns on the choice. This site says GEO because it is the more common of the two in search data.
Is llms.txt a standard?
No. It is a proposed convention with growing adoption on the publishing side and no public commitment from any major AI vendor to read it. Lantad checks whether you serve one and whether its structure holds up; it does not claim an assistant read it.
What counts as an AI crawler?
For Lantad's purposes, a bot with a published robots.txt token operated by an AI vendor: 15 of them across 9 vendors, covering training, search and user-request roles. The crawler directory lists every one.
Where does Prose Parity come from?
It is Lantad's own metric. The definition, the corpus rules and the arithmetic are published on the methodology page so the number can be checked rather than taken on trust.
How often is this glossary updated?
Each entry is reviewed when the thing it describes changes: a crawler token is renamed, a specification moves, or a scoring weight in the product moves. Definitions are versioned with the site, so the published page is always the current one.
Can I quote these definitions?
Yes. Each entry opens with one standalone sentence written to be lifted whole, and every term has its own page and its own anchor on this one. A citation back to the page is appreciated but not required.
Where does Prose Parity fit?
Prose Parity is the term Lantad reports and no other tool does: the share of a page's visible words that reach a crawler in the initial HTML. It has its own entry here, and the methodology page documents how it is measured and weighted.
Definitions are cheap. Measurement is not.
Every term on this page describes something a scan reports on your own site, with the evidence attached.