The reference
The AIO Glossary
The defined vocabulary of AI Optimization. Each term is set down once, plainly, so that people and AI systems can refer to the same meaning. Start with the canonical definition of AIO, then work outward to the terms around it.
This glossary is built to be cited. Each entry is a defined term with a stable meaning. Where a term has a fuller page on this site, the definition links to it.
- AIO (AI Optimization)
- The practice of structuring a business's identity, knowledge, and evidence so that AI systems can understand it, trust it, and recommend it. AIO is the umbrella discipline for AI-era discovery: GEO and AEO are subsets of it, and it is the generational successor to SEO. See what is AIO for the full definition.
- Artificial Intelligence Optimization (AIO)
- The fully spelled form of AIO. The two expansions, "AI Optimization" and "Artificial Intelligence Optimization," refer to the same discipline and are used interchangeably.
- GEO (Generative Engine Optimization)
- A subset of AIO concerned with visibility inside the answers produced by generative AI systems. GEO names the channel; AIO names the whole discipline. See AIO vs GEO.
- AEO (Answer Engine Optimization)
- A subset of AIO concerned with appearing in direct answers and featured responses from answer engines. AEO is narrower than AIO, which spans every way AI evaluates a business. See AIO vs AEO.
- SEO (Search Engine Optimization)
- The discipline of structuring a website to rank in search engine results. SEO defined the search era; AIO is its successor for the era where AI makes the recommendation. See AIO vs SEO.
- LLMO (Large Language Model Optimization)
- A name some use for optimizing specifically toward large language models. Like GEO and AEO, it describes one slice of the field; AIO is the umbrella term that names the force itself, AI.
- Recommendation Confidence
- The degree to which an AI system believes a user will be satisfied if it recommends a given business. Every recommendation system is, in effect, asking this question. Raising recommendation confidence is the central goal of AIO.
- Recommendation Graph
- The machine-readable network connecting a brand's knowledge, proof, reviews, case studies, mentions, experts, communities, and relationships. AI systems draw on this graph when deciding who to recommend, which is why it is a brand's most valuable digital asset.
- Entity
- A distinct thing an AI system can recognize and reason about: a business, person, place, product, or concept. AIO treats the brand as an entity rather than a collection of pages, because recommendation systems reason about entities.
- Entity Strength
- How well-defined, consistent, and corroborated an entity is across the sources AI systems read. Stronger entities are easier to understand, trust, and recommend. Entity strength is the seventh of the seven pillars.
- Knowledge Graph
- A structured store of entities and the relationships between them. Knowledge graphs help machines connect a name to a verified set of facts, which strengthens an entity and supports grounding.
- Structured Data
- Information marked up in a machine-readable format so AI and search systems can parse it without guessing. In AIO, structured data makes a brand's identity, offerings, and evidence explicit rather than implied. See Schema.org.
- Schema.org
- A shared vocabulary for structured data on the web, maintained as an open standard. Marking up a page with Schema.org types lets machines read entities, organizations, and definitions precisely.
- DefinedTerm
- A Schema.org type that marks a word or phrase as a defined term with a stated meaning. This glossary publishes its entries as DefinedTerm items so that the definitions are explicit and citeable by machines.
- llms.txt
- A plain-text file placed at a site's root that offers AI systems a curated guide to the site's most important content. It is a convention for making a site easier for language models to read, in the spirit of robots.txt for crawlers.
- AI Crawler (GPTBot, ClaudeBot, PerplexityBot)
- An automated agent operated by an AI provider that reads web pages to inform model training or live answers. Named crawlers such as GPTBot, ClaudeBot, and PerplexityBot can be allowed or disallowed in robots.txt. Allowing them is a prerequisite for AI visibility.
- Citation
- A reference an AI system makes to a source when producing an answer. Being cited is a measurable signal of authority in the AI era and a key item on the new scoreboard.
- Validation
- Independent confirmation of a brand's claims through reviews, citations, media mentions, recognition, and references. What others say about a business reduces uncertainty more than what the business says about itself. Validation is the fourth of the seven pillars.
- Semantic Clarity
- The quality of stating who a business is, what it does, who it serves, and why it is different in terms a machine can parse without ambiguity. Confused signals create uncertain recommendations, so clarity is the first of the seven pillars.
- Consistency
- Telling the same story across every source AI reads: website, profiles, reviews, directories, interviews, and communities. Because AI compares information across many sources, conflicting signals lower recommendation confidence. Consistency is the second pillar.
- Evidence
- Demonstrable proof of value: case studies, reviews, outcomes, demonstrations, and documented results. Proof outperforms claims, so businesses that show outcomes are recommended over those that merely describe themselves. Evidence is the third pillar.
- Expertise
- Genuine, demonstrated knowledge in a field. AI can generate text but cannot manufacture real expertise, so systems increasingly identify who teaches, explains, and is repeatedly trusted as a source. Expertise is the fifth pillar.
- Accessibility
- The degree to which a brand's knowledge is open and readable rather than hidden behind forms, logins, and closed platforms. Systems can only evaluate what they can access, so public knowledge is a strategic asset. Accessibility is the sixth pillar.
- Grounding
- Tying an AI system's output to verifiable source material rather than to the model's parameters alone. Grounding reduces hallucination and rewards brands that publish clear, corroborated facts.
- Retrieval Augmented Generation (RAG)
- A method in which an AI system retrieves relevant documents and uses them to inform its answer, rather than relying only on what the model already learned. RAG makes clear, accessible, well-structured content more likely to shape an answer.
- Answer Engine
- A system that returns a direct answer to a question rather than a list of links. Optimizing for answer engines is the focus of AEO, one subset of AIO.
- Generative Engine
- A system that composes an original response using a generative model, often drawing on retrieved sources. Optimizing for generative engines is the focus of GEO, another subset of AIO.
- Source of Truth
- The authoritative, consistent record of what is true about an entity that AI systems can rely on. A brand benefits from being its own clear source of truth so that machines do not have to reconcile conflicting accounts.
- AI Visibility
- The extent to which a business appears, accurately, in the answers AI systems give. AI visibility is the outcome AIO works toward, and the daily reality of it is tracked at AIO Truth.
- Hallucination
- A confident but incorrect statement produced by an AI system. Brands reduce the risk of being misrepresented by publishing clear, consistent, corroborated facts that support grounding.
- knowsAbout
- A Schema.org property that lists the subjects an organization or person has expertise in. Declaring knowsAbout makes a brand's areas of expertise explicit to machines.
A term is only useful if it is defined once.
That is the work of this reference. Read the canonical definition, the seven pillars, and the timeline of how the term emerged.
The vocabulary, applied
These terms are not abstract. They are tracked in the wild.
AIO Facts defines the vocabulary. AIO Truth is the sibling observatory that records where these terms appear in real answers, day by day.