01Why this list exists
Semantic SEO borrows most of its vocabulary from linguistics and knowledge representation, not from marketing. That’s useful once you know the words, and genuinely confusing before you do — “entity,” “predicate,” and “central search intent” aren’t self-explanatory the way “backlink” or “meta description” are. This isn’t a list of buzzwords stacked for SEO padding. It’s the working vocabulary used across every guide in this knowledge base, and getting it straight matters practically: if “central entity” and “topic” mean the same thing to you, you’ll misread half of what a proper topical map is trying to do.
The terms below are grouped the way they’re actually used in practice — starting with the two or three you need before anything else makes sense, then building outward into structure, retrieval mechanics, and the newer AI-search-specific vocabulary.
02The two terms everything else depends on
Before a topical map, before a content brief, before a single page is outlined, two decisions get made: what is this about, and why are people actually looking for it. Those two decisions have names.
Semantic SEO is the practice of structuring content around entities and meaning rather than exact-match keywords and keyword density — full definition in What Is Semantic SEO? It’s the umbrella term for everything else on this page.
The central entity is the one thing a page, or an entire site, is fundamentally about. It sounds obvious until you try to name it precisely: “CRM software” is too broad to cover completely, “CRM pricing for three-person real estate teams” is too narrow to sustain a real cluster around, and “CRM for real estate agents” sits at the right altitude — specific enough to map exhaustively, broad enough to justify a dozen supporting pages. Every subsequent decision about what to write, what to link, and what to leave out gets tested against this one choice.
The central search intent is the dominant real reason people search for that entity — not the reason you’d prefer them to have. An entity can plausibly carry several intents (informational, comparison, transactional), and picking the wrong dominant one is why a well-written, commercially-focused page sometimes loses to a purely educational one for the same term: the content answers a question the searcher wasn’t actually asking. Both terms are covered together, with worked examples, in Central Entity & Central Search Intent.
03How a site gets structured around that entity
Once the central entity is fixed, the next layer of vocabulary describes how content gets organized around it — and at what scale a given decision actually operates.
A topical map is the planned structure of pages needed to cover that entity completely: a pillar page that introduces the topic broadly, and a set of spoke or cluster pages that each go deep on one sub-topic and link back up to the pillar. See What Is a Topical Map? for the mapping process itself. Within that map, pages usually split into core content — the commercial, money-page material — and outer content, the supporting and educational material that builds depth and earns trust around the commercial core without trying to sell anything directly.
The accumulated result of doing this well across an entire cluster, rather than on one page in isolation, is topical authority: the trust a site earns by comprehensively covering a subject, which search and AI systems treat very differently from a single well-optimized page with no support structure around it.
Two more terms describe the scale at which you’re working rather than the structure itself. Macro semantic SEO is the site-level question — which entities a domain should own, and which it should deliberately avoid. Micro semantic SEO is the page-level question — whether one specific page actually proves its coverage of the entity it was assigned. Confusing the two is one of the most common structural mistakes in practice; see Macro vs Micro Semantic SEO for how to tell which one is actually failing when a site underperforms.
04The mechanics underneath: entities, extraction, and the graph
This is the layer of vocabulary that explains why the structural advice above works, rather than just what to do.
An entity is a uniquely identifiable thing or concept — a person, place, brand, product, or idea — that a search or AI system can recognize consistently across different pages and different wordings. Entities matter because modern retrieval doesn’t match strings; it matches things. That recognition process is called Named Entity Recognition (NER): the step where a system scans your text and extracts the specific entities it refers to, before it ever judges relevance. Ambiguous writing — vague pronouns, inconsistent naming of the same thing across a page — makes this extraction harder and less confident, which is a page-level (micro) problem with real consequences. See NER and SEO.
Once entities are extracted, they get resolved against Google’s Knowledge Graph — a structured database of entities and the verified relationships between them, separate from the crawled web index. A query that resolves cleanly to a known entity can be answered from graph facts directly. Full mechanics in The Knowledge Graph and SEO. Entity SEO is the practice built around making sure your brand, product, or core topic resolves as a distinct, unambiguous entity rather than staying an unresolved string — see Entity SEO Explained.
05AI search: newer terms, same underlying logic
The newest layer of vocabulary describes how AI-driven search and answer systems consume the same entity-and-coverage model, with one added wrinkle: they don’t process a query as a single unit.
Query fan-out is what happens before you see an AI-generated answer: the system splits one question into several smaller sub-queries, retrieves sources for each separately, then synthesizes the result. A page doesn’t need to rank #1 for the head term to get cited under this model — it needs to be the best available source for just one of the sub-queries the fan-out generates, which is why anticipating those sub-questions in the content brief matters. Full breakdown, including a worked example, in What Is Query Fan-Out?
Three closely related disciplines describe the broader goal of being visible inside AI-generated answers rather than just the traditional results list: AEO (Answer Engine Optimization), optimizing to be the direct answer an answer engine surfaces (full guide); GEO (Generative Engine Optimization), optimizing for visibility inside generative AI answers more broadly (full guide); and LLM visibility, the narrower question of whether you get cited by name inside an LLM’s response at all (full guide). All three sit on top of the same entity and coverage fundamentals as the rest of this glossary — they’re a retrieval context, not a separate discipline requiring separate content.
Full index → Semantic SEO · Topical Authority · AI Search · SEO Systems