[Agentic Search](atomicglue.co/glossary/agentic-search) — Agentic search is search performed by AI agents on behalf of users. Rather than the user searching, the AI agent independently researches, compares, and executes tasks across multiple steps and sources. It represents the third wave of search evolution.
[Agentic Workflow](atomicglue.co/glossary/agentic-workflow) — An agentic workflow is a structured process where AI agents autonomously plan, execute, and iterate on multi-step tasks, using tools and decision-making at each stage to achieve a goal.
[AI Agent / Autonomous Agent](atomicglue.co/glossary/ai-agent) — An AI agent is an LLM-powered system that can independently plan, use tools, execute multi-step tasks, and make decisions to achieve a goal with minimal human intervention.
[AI Crawler (GPTBot, ClaudeBot, PerplexityBot)](atomicglue.co/glossary/ai-crawler) — AI crawlers are bots operated by AI companies (OpenAI, Anthropic, Perplexity) that index web content for LLM training and/or real-time search retrieval. GPTBot, ClaudeBot, and PerplexityBot are the three you need to know, and you should configure your robots.txt to allow search bots while optionally blocking training bots.
[AI Gateway](atomicglue.co/glossary/ai-gateway) — An AI gateway is a middleware layer that sits between your applications and AI model providers, handling routing, security, rate limiting, cost tracking, and governance for all AI traffic.
[AI Mode](atomicglue.co/glossary/ai-mode) — Google AI Mode is an advanced search interface within Google Search that provides Gemini-powered conversational, multi-step reasoning with inline citations. It goes beyond AI Overviews by handling follow-ups, comparisons, and complex research tasks in a single session.
[AI Observability](atomicglue.co/glossary/ai-observability) — AI observability is the practice of monitoring, tracing, and analyzing AI system behavior in production to understand why models produce the outputs they do, track quality, and debug failures.
[AI Overview](atomicglue.co/glossary/ai-overview) — Google's AI Overview is a generative answer block displayed above organic search results, synthesizing information from multiple sources into a conversational response with inline citations. It replaced the Search Generative Experience (SGE) and became Google's default answer format for a significant share of queries.
[AI-Assisted Development (Cursor / Copilot)](atomicglue.co/glossary/ai-assisted-development) — AI-assisted development uses LLMs to help developers write, debug, and understand code in real time. GitHub Copilot, Cursor, Claude Code, and Codex are the leading tools in this rapidly evolving category.
[Answer Engine](atomicglue.co/glossary/answer-engine) — An answer engine is a search platform that delivers direct answers instead of links. ChatGPT, Perplexity, and Google AI Overviews are answer engines. They extract, synthesize, or generate a response so users don't need to click through to a website.
[Artificial Intelligence](atomicglue.co/glossary/artificial-intelligence) — Artificial Intelligence is the simulation of human intelligence by machines, especially computer systems. It enables software to reason, learn, perceive, and make decisions at scale.
[ChatGPT Search](atomicglue.co/glossary/chatgpt-search) — ChatGPT Search is OpenAI's AI-powered search engine that combines web retrieval with GPT generation. It provides cited, conversational answers to queries. It is one of the three major AI search platforms (with Google AI Overviews and Perplexity).
[Chunking](atomicglue.co/glossary/chunking) — Chunking is the process of breaking content into smaller, semantically coherent segments that AI retrieval systems can index and retrieve independently. It is a core technique in RAG architecture and directly impacts whether your content gets cited.
[Conversational AI](atomicglue.co/glossary/conversational-ai) — Conversational AI is the technology that powers natural, human-like conversations between computers and people. It includes chatbots, voice assistants, and customer support automation, now driven primarily by LLMs.
[Conversational Search](atomicglue.co/glossary/conversational-search) — Conversational search is a search experience where users interact through natural language dialogue with follow-up questions, rather than typing isolated keywords. ChatGPT and Perplexity are conversational search interfaces; traditional Google is not.
[E-E-A-T in AI Search](atomicglue.co/glossary/eeat-in-ai-search) — E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) is Google's quality framework for evaluating content. In AI search, E-E-A-T signals are even more critical because AI systems use them to decide which sources to cite with confidence.
[Edge AI](atomicglue.co/glossary/edge-ai) — Edge AI runs machine learning models on local devices (phones, cameras, IoT hardware) instead of in the cloud. It enables real-time AI inference with no internet dependency, lower latency, and better privacy.
[Embeddings](atomicglue.co/glossary/embeddings) — Embeddings are numerical vector representations of text that capture semantic meaning, enabling AI systems to match concepts by similarity rather than exact keywords. They are the foundation of semantic search and RAG retrieval.
[Entity](atomicglue.co/glossary/entity) — In AI and search, an entity is a distinct, identifiable thing or concept (a person, brand, place, product, idea) that a knowledge graph can recognize and relate to other entities. Entities are the atomic units of understanding for AI systems.
[Fine-Tuning](atomicglue.co/glossary/fine-tuning) — Fine-tuning is the process of taking a pre-trained model and training it further on a smaller, task-specific dataset to improve its performance on a particular use case or change its behavior.
[Freshness Signal](atomicglue.co/glossary/freshness-signal) — A freshness signal is any indicator that tells AI systems your content is current and regularly updated. AI search engines, especially Google AI Overviews, increasingly favor fresh content over older but potentially more authoritative sources.
[Generative AI (GenAI)](atomicglue.co/glossary/generative-ai) — Generative AI refers to AI models that create new content text, images, audio, video, or code rather than just analyzing or classifying existing data.
[Generative Engine](atomicglue.co/glossary/generative-engine) — A generative engine is an AI-powered system that produces original text responses by synthesizing information across multiple sources, rather than simply extracting and displaying a single passage. ChatGPT is the prototypical generative engine.
[GPT](atomicglue.co/glossary/gpt) — GPT stands for Generative Pre-trained Transformer. It is a family of LLMs developed by OpenAI that popularized the modern AI boom, starting with GPT-1 in 2018 and evolving through GPT-4o and beyond.
[Grounding (AI)](atomicglue.co/glossary/grounding) — Grounding is the practice of tying an AI's generated response to verifiable, up-to-date sources from the web. It is the primary technique for preventing hallucinations. RAG is the most common grounding architecture for AI search.
[Guardrails](atomicglue.co/glossary/guardrails-ai) — AI guardrails are safety mechanisms that prevent AI systems from producing harmful, biased, or unsafe outputs. They filter inputs, constrain model behavior, and validate outputs in real time.
[Hallucination](atomicglue.co/glossary/hallucination) — A hallucination is an AI-generated response that is confident, plausible, and completely false. LLMs produce hallucinations when they lack grounded, retrieved sources and rely instead on pattern completion from training data.
[Knowledge Graph](atomicglue.co/glossary/knowledge-graph) — A knowledge graph is a structured database of entities and their relationships, used by Google and other AI systems to understand facts about people, places, things, and how they connect. It powers knowledge panels and informs AI Overviews.
[Large Language Model](atomicglue.co/glossary/large-language-model) — A Large Language Model is a neural network trained on massive text data that can generate, summarize, translate, and reason about human language. GPT-4, Claude, Gemini, and Llama are all LLMs.
[Machine Learning (ML)](atomicglue.co/glossary/machine-learning) — Machine Learning is a subset of AI where systems learn from data rather than being explicitly programmed. Instead of writing rules, you show the system examples and it figures out the patterns.
[Multi-Agent System](atomicglue.co/glossary/multi-agent-system) — A multi-agent system coordinates multiple AI agents that specialize in different tasks and collaborate to solve problems that a single agent cannot handle effectively.
[Natural Language Processing](atomicglue.co/glossary/natural-language-processing) — Natural Language Processing is the branch of AI that enables computers to understand, interpret, and generate human language. It powers everything from spell check to ChatGPT.
[Perplexity](atomicglue.co/glossary/perplexity) — Perplexity is an AI-powered answer engine that provides cited, real-time responses to queries with distinctive inline annotations and a dedicated source panel. It is one of the three major AI search platforms and is known for its transparent citation style and research-focused interface.
[Prompt Chaining](atomicglue.co/glossary/prompt-chaining) — Prompt chaining is the technique of breaking a complex task into multiple smaller LLM calls, where the output of one prompt becomes the input to the next. This improves reliability and enables multi-step reasoning.
[Prompt Engineering](atomicglue.co/glossary/prompt-engineering) — Prompt engineering is the practice of designing and refining input prompts to get reliable, accurate, and useful outputs from LLMs. It is the primary interface for controlling model behavior without modifying the model itself.
[RAG (Retrieval-Augmented Generation)](atomicglue.co/glossary/rag-ai) — RAG is a technique that improves LLM outputs by retrieving relevant information from a knowledge base before generating a response. It grounds the model in facts instead of relying solely on its training data.
[Responsible AI / AI Bias](atomicglue.co/glossary/responsible-ai-ai-bias) — Responsible AI is the practice of designing, developing, and deploying AI systems that are fair, transparent, accountable, and aligned with human values. AI bias is a core concern: models can amplify historical and societal biases present in training data.
[Retrieval-Augmented Generation (RAG)](atomicglue.co/glossary/rag-ai-search) — RAG (Retrieval-Augmented Generation) is an AI architecture that retrieves relevant passages from a knowledge base (often the web) and passes them to a language model to generate a grounded, cited response. It is the technical backbone of ChatGPT Search, Perplexity, and many AI Overviews.
[Semantic Search](atomicglue.co/glossary/semantic-search) — Semantic search is search that matches the meaning and intent of a query rather than just the keywords. It is powered by embeddings and entity understanding, and it is how all modern AI search engines work.
[Vector Embeddings](atomicglue.co/glossary/vector-embeddings) — Vector embeddings are numerical representations of data (text, images, audio) as arrays of numbers that capture semantic meaning. Similar items have similar numerical representations, enabling semantic search and pattern matching.
[Zero-Click Search](atomicglue.co/glossary/zero-click-search) — Zero-click search refers to queries where the user finds the information they need directly on the search results page and never clicks through to an external website. It is the structural condition that makes AEO necessary.