Fan Out (Query Fan-Out): An AI search technique that breaks a single user query into multiple sub-queries to cover different facets or intents, allowing deeper and more diverse results.
Knowledge Retrieval: The process of extracting relevant information from data sources to support AI-generated answers or tasks.
RAG (Retrieval-Augmented Generation): An AI method that improves accuracy by retrieving external content during response generation, effectively combining search and synthesis.
Generative AI: AI that can create new content (text, images, video, etc.) based on patterns it has learned, rather than just analyzing or classifying existing data.
Large Language Model (LLM): A powerful AI model trained on vast text data to understand and generate human-like language.
Hallucination (AI Hallucination): When an AI confidently provides made-up or incorrect information. These errors highlight the need for fact-checking AI outputs.
Conversational AI (Conversational Search): AI that communicates through natural dialogue, allowing for multi-turn interactions instead of one-off search queries.
Google AI Mode: Google’s upcoming AI-powered search interface that replaces traditional results with conversational, personalized responses.
SGE (Search Generative Experience): A trial phase where Google tested AI-generated summaries at the top of search results to inform AI Mode.
AI Overview: Google’s AI-generated summary of a query, offering concise answers sourced from multiple websites, typically shown above regular search results.
Search Journey (AI-Driven Search Journey): The recognition that users follow multi-step processes when searching. AI links past behavior and context to guide future queries.
Hyper-Personalization: Ultra-tailored content or search results generated using personal data, context, preferences, and behaviors.
Google Discover: A predictive feed of personalised content that shows users articles and updates before they even search.
Multimodal AI: AI that can interpret and respond to different data types (e.g. text, images, voice) simultaneously. Enables richer and more natural user interactions.
Gemini: Google’s advanced foundation model powering AI Mode and other services. It’s multimodal, context-aware, and integrated across the Google ecosystem.
Semantic SEO: An SEO strategy that optimizes content for meaning and intent, rather than just specific keywords.
Entity Optimization (Entity SEO): Making your brand or topic clearly recognisable to AI and search engines by aligning with known concepts (entities) in the Knowledge Graph.
Knowledge Graph: Google’s database of entities and their relationships. It supports Knowledge Panels and helps AI understand concepts and context.
Knowledge Panel: A box on Google Search that displays key facts about a known entity, pulled from the Knowledge Graph.
E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness): Quality signals used by Google to assess whether content and creators are credible.
Zero-Click Search: A search where the user’s query is answered directly on the results page (via snippets, AI summaries, etc.), so they don’t click to a website.
Augmented Reality (AR) Search: A search mode where real-world surroundings are enhanced with digital information using devices like smartphones or smart glasses.
Vector Search: A way of retrieving content based on semantic meaning rather than keyword matching. Useful in AI-driven search experiences.
Embedding: A numerical representation of text (or other data) that AI uses to measure similarity and understand context.
Token: A basic unit of input/output in LLMs. Tokens include words, punctuation, or parts of words. Token count affects how much content AI can process at once.
Agentic AI (AI Agents): AI that can perform tasks on behalf of users across platforms (e.g., booking, replying, updating tools), not just provide answers.
Prompt Engineering: The craft of designing effective prompts to guide AI responses and improve output quality.
Chain-of-Thought Reasoning: An AI method where answers are explained step-by-step, making complex results easier to understand and trust.
Content Fragmentation (Fraggles): The surfacing of small, high-value content blocks (e.g., an FAQ or stat) in search or AI summaries, rather than full pages.
Entity-First Indexing: Google’s shift to prioritising “things” (entities) over keywords when indexing and ranking content.
Topical Authority: A site’s depth of content and credibility on a specific subject area, which helps AI and search engines trust and surface its material.
Data Leakage: When sensitive or unintended data is used or revealed by AI, especially if training data includes private or confidential sources.
Synthetic Data: Artificially created data used to train or test AI systems without exposing real customer information.