Glossary

AEO and GEO terms explained

The vocabulary of the new marketing discipline — definitions, relationships, and why each concept matters.

This page is kept live. We update it monthly as new AI features emerge or industry terminology evolves. If a term is missing, email us and we will add it.

Terms A–Z(68 entries)

AEO
AEOlab score
AI citation
AI mention
AI Share of Voice
AISO
Answer engine
Anthropic Claude
Apple Intelligence
Applebot-Extended
Article / BlogPosting Schema
Brand entity recognition
Brand position in AI response
BreadcrumbList Schema
ChatGPT
ChatGPT Search
ChatGPT-User
Citation
Citation rate
ClaudeBot
DefinedTerm + DefinedTermSet
E-E-A-T
Embedding
FAQPage schema
Fine-tuning
GBP attributes
GBP categories
Generative ad
GEO
Google AI Overview
Google Business Profile
Google Gemini
Google-Extended
GPTBot
Hallucination
HowTo Schema
Industry aggregator
JSON-LD
Knowledge cutoff
Knowledge Graph
Knowledge Panel
LLM
llms.txt
Local SEO
LocalBusiness schema
Microsoft Copilot
NAP consistency
OAI-SearchBot
OfferCatalog
Perplexity
PerplexityBot
Person schema
ProfessionalService schema
Prompt engineering
RAG
Review velocity
Rich Results
robots.txt
Schema.org
SEO
Service Schema
SGE
Sponsored answers
Structured Snippet
Transformer architecture
Vector database
Wikidata
Zero-click search

Core concepts

AEO (Answer Engine Optimization): The emerging marketing discipline targeting a business's visibility in the answers of generative AI systems — ChatGPT, Gemini, Perplexity, Microsoft Copilot. It is built alongside classical SEO, not as a replacement.
GEO (Generative Engine Optimization): A synonym for AEO. Specifically refers to optimization for generative LLM systems (ChatGPT, Claude, Gemini). The two terms are used interchangeably in practice.
SEO (Search Engine Optimization): Classical search engine optimization, targeting Google's blue-link results. Still necessary alongside AEO.
Local SEO: The local-search branch of classical SEO (e.g. "dentist 5th district"). Overlaps heavily with AEO: Google Business Profile, NAP consistency, local reviews.
AISO (AI Search Optimization): A third label for the same discipline as AEO and GEO. Used by some English-speaking agencies. AEO is the most common term in the Hungarian market.

Structured data

Schema.org: A collaborative initiative by Google, Microsoft, Yahoo and Yandex for structured content markup on the web. Schema.org types (Organization, Person, FAQPage, etc.) help AI systems categorise a website's content more accurately.
JSON-LD: JavaScript Object Notation for Linked Data — the recommended format for Schema.org structured data on websites. It sits in a `<script type="application/ld+json">` block inside the HTML.
Rich Results: Special search results on Google (star ratings, FAQ accordions, etc.) generated from Schema.org structured data. Verifiable with Google's Rich Results Test tool.
ProfessionalService schema: A Schema.org type for professional service providers (lawyers, accountants, consultants). An AEO fundamental — combinable with LocalBusiness for local service providers.
LocalBusiness schema: The Schema.org type for a local business with a physical address, opening hours and phone number. Strongly recommended for any business serving local clients.
FAQPage schema: The Schema.org type for marking up frequently asked questions. A strong AI signal: ChatGPT and Perplexity frequently cite FAQPage content because the Q&A structure is directly reusable.
Person schema: Schema.org type for a natural person (business owner, head physician, attorney). A key AI authority signal: a business associated with a named, identifiable expert gains credibility.
DefinedTerm + DefinedTermSet: Schema.org types for glossary entries. This page uses them: each definition is a DefinedTerm, the whole page is a DefinedTermSet — directly indexable by AI systems.
HowTo Schema: Schema.org type for step-by-step guides. AI Overview particularly likes citing HowTo content for complex processes (e.g. "How to prepare for implant surgery"). Includes: step count, optional duration and tool requirements.
BreadcrumbList Schema: Schema.org markup for breadcrumb navigation. Helps AI systems understand website hierarchy. Recommended on every subpage.
Service Schema: Schema.org's service type — for describing a specific service (e.g. "Dental Implant", "Company Formation"). Can include price offers, provider, and area served.
OfferCatalog: A Schema.org representation of a business's full service catalogue with prices. AEOlab's homepage uses it: the 3 packages are delivered as an OfferCatalog in structured data.
Article / BlogPosting Schema: Schema.org markup for articles and blog posts. Includes: author, publish date, modified date, headline, keywords. Gives AI systems context about the content.

AI systems

ChatGPT: OpenAI's generative AI assistant. The most important AEO target in the Hungarian market: a significant share of the 25–50 age group now turns to ChatGPT first when searching for professionals.
Google Gemini: Google's generative AI assistant (formerly Bard). It relies on Google search results, so classical SEO directly influences Gemini visibility.
Perplexity: An answer-focused AI search engine that cites sources for every response. The most transparent AEO system: you can see exactly which pages it used for a given answer.
Microsoft Copilot: Microsoft's generative AI assistant, built on Bing and embedded in the Office suite. Uses the GPT model at its core, but with its own web-search index.
Anthropic Claude: Anthropic's generative AI assistant. Smaller consumer market share in Hungary, but growing in B2B (especially consulting and legal).
Google AI Overview: The AI-generated summary appearing above Google's classic search results (formerly SGE — Search Generative Experience). Live in Hungary since early 2026. Around 50–60% of users stop here without clicking through to classic results.
SGE (Search Generative Experience): The original name for AI Overview. Google rebranded it as "AI Overview" in 2025.
Apple Intelligence: Apple's own generative AI platform, enhancing Siri, iMessage, Mail and Safari with AI features. Rolled out in the Hungarian language region during 2025. Estimated user base: ~30% of the Apple device market.
ChatGPT Search: ChatGPT's real-time web search feature (formerly SearchGPT). Available by default for Plus users. Integrates fresh results on top of the base training data.

AI crawlers

GPTBot: OpenAI's official crawler, collecting web content for AI model training. AEO-oriented sites should explicitly allow it in robots.txt.
Google-Extended: Google's dedicated AI training crawler, separate from Googlebot. Must be allowed in robots.txt if you want your content used in Gemini training.
ClaudeBot: Anthropic's official crawler. Allowable via robots.txt Allow directive.
PerplexityBot: Perplexity's crawler. Since Perplexity runs real-time web searches, keeping content fresh and indexed is especially important here.
robots.txt: A text file at the website root that tells crawlers what they may and may not index. AEO-oriented sites use explicit Allow rules to permit AI bots (GPTBot, ClaudeBot, PerplexityBot, etc.).
llms.txt: A proposed new standard: a markdown file at the website root summarising a business's key data, services and pages in a structured, AI-friendly format. Not yet an official standard, but an increasing number of AI systems consider it.
Applebot-Extended: Apple's AI training crawler, separate from Applebot (which indexes for Spotlight and Siri Suggestions). Allowable via robots.txt Allow directive.
OAI-SearchBot: OpenAI's dedicated crawler for the ChatGPT Search service. Collects information for live web queries. Separate from GPTBot (which is for training data).
ChatGPT-User: OpenAI's real-time crawler for user interactions (when a user provides a specific URL to ChatGPT). Should be allowed if you want ChatGPT to be able to read your page when linked directly.

Off-site signals

Google Business Profile (GBP): Google's free business profile service (formerly Google My Business). An AEO fundamental: Gemini and Google AI Overview rely heavily on GBP data. Full completion (hours, attributes, photos, categories, Q&A) has a measurable effect.
NAP consistency: Name, Address, Phone — a business's name, address and phone number should appear in identical form across every online platform (website, Google Business, Facebook, industry directories). NAP discrepancies can reduce AI systems' confidence in citing you.
Wikidata: The open knowledge base behind Wikipedia. Major LLM models (GPT, Gemini, Claude) relied heavily on it during training. A well-filled Wikidata entry measurably increases mention probability.
Industry aggregator: Industry-specific listing portals (e.g. Foglaljorvost.hu and Doklist for dentists; the Hungarian Bar Association directory for lawyers; Ingatlan.com for real estate). An important source for Hungarian AI systems because they store professional data in structured form.
Knowledge Panel: The structured info box on the right side of Google results about an entity (company, person, place). Built from the Knowledge Graph. A strong AI signal — AI models use it directly.
GBP categories: The primary and secondary categories selected in Google Business Profile. A strong AEO signal — Gemini and AI Overview use them to classify the business. One primary + 2–3 secondary categories recommended.
GBP attributes: Properties configurable in Google Business Profile (e.g. wheelchair accessible, parking, online booking, card payment). AI systems use them for specific queries ("Is there wheelchair access?").
Review velocity: The rate of new reviews arriving on Google Reviews. Google's algorithm favours fresh, regular new reviews. Stagnation (6+ months without new reviews) is a negative signal.

Metrics

AI mention: An instance where a generative AI system explicitly names a specific business or professional in a response to a user query. The primary KPI for AEO: mentions / queries tested.
AI Share of Voice: The proportion of AI answers that mention a given business across a defined set of queries, relative to competitors. E.g. "mentioned in 5 of 15 queries, top competitor in 8" → 33% Share of Voice.
AEOlab score (0–100): AEOlab's proprietary 0–100 scoring system, evaluating four dimensions: technical signals, content signals, off-site signals, AI visibility. Delivered in the audit report with competitor comparison.
Citation rate: The frequency of AI citations pointing to a given page. E.g. if Perplexity cited the business website in 8 of 15 prompts, that is a 53% citation rate.
Brand position in AI response: The order in which a business is mentioned in an AI answer. The business cited as the first recommendation can achieve 2–3× higher conversion rates.
Knowledge cutoff: The training data end date of an AI model. The model knows data up to this date. E.g. GPT-4's knowledge cutoff is April 2024. The real-time web layer (RAG) supplements this.
Brand entity recognition: An AI system's ability to unambiguously recognise a business name. If a name is ambiguous or a common word, the system may confuse it with other entities. A Wikidata entry and Schema.org Person strongly improve this.

Other

RAG (Retrieval-Augmented Generation): The technique where an LLM searches the web or a knowledge base before generating a response. Perplexity, ChatGPT Search and Microsoft Copilot are all RAG-based. The key AEO insight: fresh web content directly influences the answer.
E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness — Google's quality evaluation principles. In AEO the named expert (Person schema), cited sources and transparency are particularly strong signals.
Knowledge Graph: Google's structured knowledge base storing entities (people, businesses, places, concepts) and their relationships. A business appearing in the Knowledge Graph (the info box on the right side of Google results) is a strong AI signal.
Structured Snippet: Structured information blocks in Google results (opening hours, phone number). A joint result of Schema.org and GBP — signals that Google successfully parsed the business's structured data.
Citation: When an AI system (e.g. Perplexity) references a specific passage from a webpage in its response. Clear, well-structured content packed with factual statements is cited more frequently.
Hallucination: When an AI system confidently but incorrectly states information about a business (e.g. wrong opening hours, discontinued address). AEO aims to minimise hallucinations: structured data and NAP consistency help.
Prompt engineering: The craft of formulating questions (prompts) for AI systems. Relevant to AEO when building the test query set: it should mirror the actual prompts real customers use.
LLM (Large Language Model): Large Language Model — the foundation of generative AI systems. A deep-learning model trained on vast text corpora. Examples: GPT, Gemini, Claude.
Transformer architecture: The neural network architecture behind LLMs. Google's 2017 Transformer paper launched the generative AI revolution. Worth understanding for AEO context — not directly influenceable.
Fine-tuning: Further training a base LLM for a specific purpose. E.g. a base Llama model fine-tuned on Hungarian legal texts. Growing demand in the AEO market for industry-specific models.
Embedding: A numerical representation of a text in a high-dimensional space. AI systems use embeddings to find similarity between pieces of content. For AEO: the "nature" of content matters, not just keywords.
Vector database: A database designed to store embeddings and perform fast similarity search. Examples: Pinecone, Weaviate, Supabase pgvector. The foundation of RAG systems.
AI citation: When an AI system (Perplexity, ChatGPT Search, AI Overview) gives a concrete source reference in its answer. The strongest AEO feedback: the business is not just mentioned, but linked.
Sponsored answers: Paid AI advertising concept — a business pays to be more prominently mentioned in AI answers, labelled as "Sponsored". Already live on Perplexity in the US market; not yet available in Hungary.
Generative ad: Paid advertising content generated by AI systems. Unlike classic PPC, the ad appears embedded in the natural response rather than in a separate block.
Zero-click search: A search where the user gets the answer directly on the results page (e.g. from AI Overview) without clicking any link. One of the main causes of declining traffic for classical SEO.
Answer engine: A search system that returns an answer, not a list. The future evolution of the classic "search engine". AEO stands for Answer Engine Optimization.