Resources
AI Glossary
Plain-English definitions of the AI terms that come up in enterprise work — written so they're useful to operators, not just engineers.
34 terms
Agentic AIAI systems that autonomously execute multi-step tasks to reach a goal.Read definition Generative AIAI that creates new content — text, code, images, audio, video.Read definition Large Language Model (LLM)A neural network trained on massive text corpora to predict next tokens.Read definition Retrieval-Augmented Generation (RAG)Injecting retrieved documents into an LLM prompt to ground outputs.Read definition Fine-TuningContinued training of a base model on task-specific data.Read definition Prompt EngineeringDesigning inputs to LLMs that reliably produce desired outputs.Read definition Tool UseAn LLM calling external functions to extend its capabilities.Read definition Model Context Protocol (MCP)An open standard for connecting AI systems to tools and data sources.Read definition Vector DatabaseA database optimized for semantic similarity search over embeddings.Read definition EmbeddingsNumeric vector representations of text or other data that capture meaning.Read definition AI Evaluation (Evals)The practice of systematically measuring AI system quality.Read definition AI GovernancePolicies, controls, and accountability for AI inside an organization.Read definition AI StrategyThe plan for where AI creates value across a business and how it gets there.Read definition AI Maturity ModelA staged framework for assessing how advanced an organization's AI capability is.Read definition Model RiskThe risk that an AI model produces incorrect, unfair, or unsafe outputs.Read definition HallucinationWhen an LLM generates confident-sounding output that is factually incorrect.Read definition Context WindowThe maximum amount of text an LLM can consider in a single input.Read definition TokensThe units LLMs use to process text — roughly 3-4 characters or 0.75 words each.Read definition TemperatureA parameter controlling how deterministic or creative an LLM's output is.Read definition Chain-of-Thought (CoT)A prompting technique where the model explains its reasoning step by step.Read definition Reasoning ModelsLLMs that explicitly plan and reason before producing an answer.Read definition Multi-Agent SystemA system where multiple AI agents cooperate to complete work no single agent could.Read definition GuardrailsControls that prevent an AI system from producing unsafe or off-policy outputs.Read definition Prompt InjectionAn attack where an attacker's text is treated as instructions by an LLM.Read definition Human-in-the-Loop (HITL)AI system designs where a human reviews or approves key AI actions.Read definition Structured OutputWhen an LLM is constrained to produce output in a strict format like JSON.Read definition MLOpsThe engineering discipline for building, deploying, and monitoring ML systems.Read definition LLM-as-JudgeUsing one LLM to evaluate another LLM's outputs at scale.Read definition GroundingAnchoring AI outputs in verified source data to reduce hallucination.Read definition Data & Model DriftWhen the world changes and a deployed model's accuracy degrades.Read definition EU AI ActEuropean Union regulation governing AI systems by risk tier.Read definition SR 11-7 (Model Risk Management)The US Federal Reserve's model risk management guidance for banks.Read definition Foundation ModelA large model trained on broad data that can be adapted to many downstream tasks.Read definition Agent OrchestrationThe layer that coordinates multiple AI agents, tools, and models in one system.Read definition