GEO & LLM Optimisation Glossary
A beginner-friendly guide to the terms, concepts, and techniques used in Generative Engine Optimisation (GEO) and Large Language Model (LLM) search optimisation.
Use the links below to jump straight to a term.
Jump to a Section
A–C | D–G | H–K | L–O | P–R | S–V
A–C
Answer Engine Optimisation (AEO) / Artificial Intelligence Optimisation (AIO)
What it means:
Optimising content for AI-driven “answer engines” like Perplexity, Bing Chat, or Google’s AI Overviews, so your information is more likely to appear directly in their responses.
Why it matters for SEO:
AI-driven results often skip the traditional “10 blue links” and instead give a summarised answer. If your content is easy for these systems to parse and trust, you can win visibility in answers even if you’re not top in classic search rankings.
Attribution Consistency
What it means:
How reliably an AI cites the same source when answering similar or repeated questions.
Why it matters for SEO:
If an AI replaces your source with a competitor in similar queries, your brand’s share of AI-generated visibility can drop, even if your SEO is strong.
Bias Drift
What it means:
When AI results gradually start favouring certain viewpoints, sources, or topics over time.
Why it matters for SEO:
A model’s bias drift can push your site out of answer sets, affecting referral traffic and brand authority despite stable organic rankings.
Canonical Source Recognition
What it means:
Ensuring AI systems recognise your page as the “main” or most authoritative version of a piece of content.
Why it matters for SEO:
If AI attributes a fact to a duplicate or syndicated version instead of your original, you lose both the citation and the authority boost in AI-driven results.
Chain-of-Thought Prompting
What it means:
A prompting technique that encourages AI to explain its reasoning step-by-step.
Why it matters for SEO:
When AI uses chain-of-thought internally, it can better understand nuanced content. That means well-structured, logically presented content is more likely to be included in multi-step AI answers.
Chunking
What it means:
Splitting large blocks of text into smaller sections so AI can store and retrieve them more effectively.
Why it matters for SEO:
Content that’s logically segmented is more likely to match the smaller “sub-queries” AI generates during retrieval, increasing your chances of being pulled into answers.
Chunk Overlap
What it means:
Repeating a small amount of content between chunks to preserve context.
Why it matters for SEO:
Maintains continuity in AI retrieval, so your brand’s key facts don’t get lost when AI pulls only part of a page.
Chunk Size
What it means:
How much text is in each chunk.
Why it matters for SEO:
The right chunk size can ensure your content remains relevant to a query without being too diluted or too broad for AI matching.
Citation Coverage
What it means:
The percentage of AI answers on a given topic that reference your brand or content.
Why it matters for SEO:
High coverage means your brand consistently appears in generative results, which is a direct measure of GEO success.
Citation Probability
What it means:
The likelihood that an AI will mention your brand in a single relevant answer.
Why it matters for SEO:
Improving this metric means increasing the odds of your content being the chosen reference in competitive answer sets.
Context Embedding Rank
What it means:
How well your content scores in an AI’s internal search for relevant context.
Why it matters for SEO:
Higher ranking embeddings increase your visibility in AI outputs, regardless of your traditional SERP position.
D–G
Document Store / Corpus
What it means:
A collection of documents or data sources stored in a way that AI systems can search and retrieve them.
Why it matters for SEO:
If your content isn’t part of an AI’s accessible “document store” — whether via open web crawling or structured feeds — it’s invisible to retrieval, meaning no chance of inclusion in generated answers.
Edge Model Sync
What it means:
The process of keeping smaller, local versions of AI models updated so they match the main cloud-based version.
Why it matters for SEO:
If models running locally (e.g., on devices or in enterprise systems) are out of date, they might use older versions of your content or miss recent updates, reducing accuracy and visibility.
Embeddings
What it means:
A numerical representation of text that captures its meaning, allowing AI to compare and find similar content.
Why it matters for SEO:
Strong, context-rich content produces embeddings that match more queries, increasing the likelihood of retrieval in GEO contexts.
Embedding Model
What it means:
The AI system or algorithm that turns your content into embeddings.
Why it matters for SEO:
Different embedding models interpret text differently. If your content aligns well with a widely used model’s interpretation, it’s more likely to be retrieved for relevant questions.
Generative Engine Optimisation (GEO)
What it means:
The practice of improving content so AI tools like ChatGPT or Google’s AI Overviews can find, understand, and use it in answers.
Why it matters for SEO:
GEO is the natural extension of SEO into the world of AI search, it’s about securing brand visibility inside AI-generated answers, not just ranking in the traditional SERPs.
Grounding
What it means:
The process of connecting an AI’s answer to verifiable, trusted sources.
Why it matters for SEO:
When AI systems “ground” their outputs, they’re more likely to cite sources. If your site is considered trustworthy, you’re more likely to be chosen for grounding references.
Grounding Depth Index
What it means:
A measure of how much an AI’s answer is based on actual sourced content versus model “guesswork”.
Why it matters for SEO:
Deeper grounding often means more citations. If your content is detailed and well-structured, it has a better chance of forming the foundation for grounded AI answers.
H–K
Hallucination
What it means:
When an AI produces an answer that sounds correct but is factually wrong or unsupported by real sources.
Why it matters for SEO:
If an AI hallucinates about your brand, it can mislead users and damage trust. Clear, authoritative content reduces the risk of misrepresentation.
Hierarchical Chunking
What it means:
Breaking content into large sections first, then splitting those into smaller, related chunks, preserving the original structure.
Why it matters for SEO:
Maintaining hierarchy helps AI see your content as complete and well-organised, which increases retrieval accuracy.
Indexing (in RAG)
What it means:
The process of converting content into embeddings and storing them for AI retrieval.
Why it matters for SEO:
If your content isn’t indexed into an AI-accessible system, it won’t be found – just like a page not indexed by Google won’t rank.
Knowledge Base
What it means:
An organised set of content designed to answer common questions, often in a structured, fact-focused way.
Why it matters for SEO:
Knowledge base formats are easy for AI to parse and quote from, making them ideal for capturing GEO visibility.
Knowledge Graph
What it means:
A database of interconnected entities (people, places, things) and the relationships between them.
Why it matters for SEO:
If your brand, products, or topics are represented accurately in knowledge graphs, AI systems are more likely to recognise and surface them in answers.
L–O
Large Language Model (LLM)
What it means:
A type of AI trained on massive amounts of text to understand and generate human-like language. Examples include ChatGPT, Claude, and Gemini.
Why it matters for SEO:
LLMs increasingly power search experiences. Understanding how they process and select content is now critical for SEO strategy.
Model Explainability
What it means:
How easy it is to understand why an AI produced a specific answer.
Why it matters for SEO:
If an AI system can explain why it chose your content, it’s easier to demonstrate value to stakeholders and identify optimisation opportunities.
P–R
Prompt
What it means:
The text or question given to an AI to generate a response.
Why it matters for SEO:
The way users phrase prompts affects which content AI retrieves. Understanding prompt patterns can help you create content that matches them.
Prompt A/B Testing
What it means:
Testing two or more variations of a prompt to see which produces better results from AI.
Why it matters for SEO:
By testing prompts, you can discover which query styles are most likely to trigger your content in AI answers.
Prompt Chaining
What it means:
Using a sequence of prompts where each one builds on the response from the last.
Why it matters for SEO:
Some AI queries evolve over multiple steps – having comprehensive coverage of a topic increases the chances your content appears at each stage.
Prompt Engineering
What it means:
The practice of designing prompts to get accurate, relevant, and useful AI responses.
Why it matters for SEO:
Knowing prompt engineering helps predict how users and search engines might trigger AI responses that include your content.
Prompt Hygiene
What it means:
Keeping prompts clean, clear, and free from unnecessary details.
Why it matters for SEO:
Cleaner prompts make it easier for AI to match relevant content – reducing the risk of unrelated answers where your brand could be replaced.
Query Fan-Out
What it means:
When AI breaks a single user query into multiple sub-queries to gather information.
Why it matters for SEO:
If your content doesn’t answer these sub-queries, you could miss out on being part of the AI’s final answer even if you rank for the main query.
RAG (Retrieval-Augmented Generation)
What it means:
An AI method where relevant information is retrieved from external sources before generating an answer.
Why it matters for SEO:
If your content is accessible and relevant, RAG systems can pull it directly into AI responses, effectively acting as a new search ranking opportunity.
Reference Rate
What it means:
How often an AI cites your content in its answers.
Why it matters for SEO:
A high reference rate shows your content is a trusted go-to source for AI, similar to high backlink acquisition in classic SEO.
Retrieval Freshness
What it means:
How up-to-date the information is that an AI uses to answer a question.
Why it matters for SEO:
If your content is outdated, it’s less likely to be pulled into AI results for time-sensitive queries.
Rule-Based Chunking
What it means:
Splitting text into chunks based on set rules (e.g., every heading becomes a new chunk).
Why it matters for SEO:
Consistent structure makes it easier for AI to locate and reuse specific parts of your content.
S–V
Semantic Chunking
What it means:
Splitting text into chunks based on meaning, keeping related ideas together.
Why it matters for SEO:
Semantic grouping improves AI’s ability to extract complete, coherent answers from your content.
Source Blend Ratio
What it means:
The balance of sources AI uses to form an answer.
Why it matters for SEO:
If your brand is one of only a few sources in a high blend ratio, you have stronger influence over the final AI-generated answer.
Tokenisation
What it means:
Breaking text into small units (tokens) that AI models use to process information.
Why it matters for SEO:
Token limits control how much of your content can fit into an AI’s context window, affecting what’s retrieved.
Vector Database / Vector Store
What it means:
A database that stores embeddings so AI can quickly find and compare content based on meaning.
Why it matters for SEO:
Content in a vector store can be retrieved faster and more accurately by AI, a key factor for GEO performance.
Vector Salience Score
What it means:
A measure of how important your content is within a set of embeddings.
Why it matters for SEO:
High salience means AI considers your content particularly relevant, improving inclusion odds.
Vector Similarity
What it means:
A measure of how close two pieces of content are in meaning based on their embeddings.
Why it matters for SEO:
If your content has high similarity with a user query’s embedding, it’s more likely to be retrieved and cited.
Visual Search Optimisation
What it means:
Making images easy for AI to find, understand, and use in visual search results.
Why it matters for SEO:
As AI integrates visual results into answers, optimised images become a new visibility channel alongside text.
