Twenty-seven concepts explained precisely and without promises: what they mean, how they relate to GEO and where to go deeper. Where a term has no standard definition, we say so.
To use the same vocabulary when auditing, measuring and communicating GEO. The definitions are deliberately conservative; Majotek's own working terms (Owned Semantic Capture, Publisher Capture) are flagged as such.
Work to improve how a brand is understood, retrieved, cited and recommended by AI search engines and assistants.
It combines technical SEO, citable content, entity signals, external authority and measurement. There is no public ranking equivalent to Google's, so it is assessed by trend.
Relation to GEO: It is the term that names the discipline. It builds on SEO and extends it; it does not replace it.
Optimisation so that a brand or a piece of content becomes the answer given by answer engines.
Authors use the term with different nuances: sometimes as a synonym for GEO, sometimes focused on direct answers and featured snippets.
Relation to GEO: It overlaps with GEO. Majotek uses GEO as the umbrella term and treats direct answers as one technique within citable content.
The degree to which a brand appears, is described correctly and is cited in answers from AI search engines.
It groups several signals that are best measured separately: prompt coverage, mentions, citations and accuracy.
Relation to GEO: It is the outcome GEO work pursues. Majotek uses it as the name of its service alongside GEO.
The share of prompts in a defined set in which the brand appears.
It is only comparable if the prompt set stays the same between measurements, which is why it is frozen and documented.
Relation to GEO: It is one of the baseline tracking metrics. It depends on the chosen set; it is not an absolute value.
The share of answers in which the brand is named in the text.
A mention does not imply a recommendation or a link. It is recorded separately from citations.
Relation to GEO: It measures the brand's presence in the answer, which is different from its website being cited.
The share of answers that link to one of the brand's URLs as a source.
It can only be measured when the engine's interface shows sources, and the sources shown can change between runs.
Relation to GEO: It shows whether the brand's own content is used as a source, complementing mentions.
A brand's share of presence compared with others in its category across a set of answers.
There is no standard definition in GEO: each tool or study calculates it its own way (mentions, citations or positions). Check the formula before comparing figures.
Relation to GEO: It places a brand against competitors, as long as the calculation is made explicit.
The share of answers to a set of prompts in which the brand is part of the answer.
Like Share of Voice, its definition varies with whoever uses it and it is not a standard.
Relation to GEO: It is one way to summarise presence in answers. Majotek prefers to publish the separate metrics and their formula.
An identifiable thing that is distinct from others, such as a company, a person, a product or a place.
Search and AI systems reason about entities and their relationships, not only about keywords.
Relation to GEO: A well-defined entity (what it is, what it does, who it is for) reduces confusion in answers. It is the ENTITY layer of the framework.
Agreement of an entity's name, category and details between its own website and external sources.
Inconsistencies (a different name, contradictory categories, outdated details) can lead to wrong descriptions.
Relation to GEO: It is reviewed in the audit by comparing the website, profiles and third-party mentions.
A knowledge base that organises entities and their relationships; the best known in search is Google's.
Being included in a knowledge graph is not something a company can request directly. It depends on reliable sources and consistent data.
Relation to GEO: It explains why data consistency and independent sources matter, although it does not guarantee appearing in a graph.
A technique in which a language model retrieves external documents and uses them as context to generate its answer.
It allows answering with information newer than the model's training and citing sources. How each product implements it is not public in detail.
Relation to GEO: It is why current, retrievable pages can influence the answers of AI search engines.
The stage in which the system searches for and selects the sources that go into the answer's context.
It is not always observable from outside. A source can be retrieved without ending up cited.
Relation to GEO: It is the first condition to influence an answer: if the page is not retrieved, it cannot be cited.
A link to a source that the engine shows alongside the answer.
Not every engine shows sources, nor does it always do so. A citation is different from a mention in the text.
Relation to GEO: It is the most direct and measurable result of a brand's own content in generative answers.
Presenting the brand as an option to consider or choose within an answer.
It requires interpreting the context of the mention, so it needs an explicit coding criterion to be measured.
Relation to GEO: It is the RECOMMENDATION layer of the framework. No methodology can guarantee it.
A brand presented as the first option or the most recommended one in an answer that ranks options.
It only makes sense when the answer compares or prioritises alternatives. It can vary between runs.
Relation to GEO: It is the most demanding case of recommendation and is recorded separately.
Information in a standard, machine-readable format that describes the content of a page.
Structured data must match what the page shows. It helps interpret the content but does not replace clear content.
Relation to GEO: It is reviewed in the ACCESS/ENTITY layer. There is no public evidence that it guarantees citations on its own.
A collaborative vocabulary of types and properties used to mark up structured data, for example with JSON-LD.
Common types are Organization, Person, Service, Article, FAQPage and BreadcrumbList. Only what is visible on the page should be marked up.
Relation to GEO: It is the usual way to express the entity and its relationships in the page's code.
The preferred URL of a page when several versions with identical or very similar content exist.
It is declared with the rel=canonical tag. It is a signal, not a directive: search engines may choose another URL.
Relation to GEO: It concentrates signals on a single URL and stops duplicate versions from competing. It is part of the ACCESS layer.
OpenAI's crawler used to show websites in ChatGPT search results.
According to OpenAI's documentation, it is independent of GPTBot. Allowing OAI-SearchBot in robots.txt is a requirement to be eligible for ChatGPT search, but it does not guarantee appearing or being cited.
Relation to GEO: It is checked in the ACCESS layer: robots.txt and the server must not block it by mistake.
OpenAI's crawler associated with collecting content that may be used to train its models.
It is controlled independently of OAI-SearchBot. Blocking GPTBot does not block ChatGPT search, and vice versa. Deciding on it is each company's policy.
Relation to GEO: It should not be confused with OAI-SearchBot when reviewing robots.txt.
An AI answer engine that shows the cited sources next to each answer.
It uses its own crawlers (for example PerplexityBot) and shows sources explicitly, which makes citations easier to observe.
Relation to GEO: It is one of the engines analysed separately when measuring citations and mentions.
AI-generated summaries that Google shows in some search results, with links to sources.
They appear depending on the query. To be shown as a supporting link, a page must be indexed and eligible to show snippets.
Relation to GEO: It connects classic SEO with GEO: the foundation of indexing and quality stays the same.
Google's AI-based conversational search experience, with answers and links to sources.
It allows longer and follow-up questions. Its results can differ from those of AI Overviews.
Relation to GEO: It is treated separately in measurement because its behaviour can differ from other engines.
Confirmation of what a company says by independent third-party sources.
It includes press, directories, comparisons, profiles and mentions. The more independent and consistent the source, the more it is worth.
Relation to GEO: It is layer D of the framework: it helps the entity be credible and keeps answers from relying only on the brand's own website.
Presence in third-party content (media, comparisons, listings) that engines use as a source to answer.
A Majotek working term, not a standard. It refers to being represented accurately on third-party pages that are already retrieved for a query.
Relation to GEO: It is part of External Corroboration. It must rely on verifiable information and genuine editorial sources.
Covering a category's questions and concepts with clear, citable pages of your own.
A Majotek working term. It includes landing pages, definitions, case studies and comparisons, connected by internal linking and without duplicating intents.
Relation to GEO: It is layer C of the framework: it gives engines your own pages to retrieve and cite.
