Glossary

The AI glossary for product people

Every term here is one you will actually hear in a product review, a vendor call or a roadmap debate. Definitions are short, opinionated and linked to the module that goes deeper.

• Large language model (LLM)A model trained on vast amounts of text that generates language by predicting the most likely continuation of its input.
• TokenThe unit an LLM reads and writes - roughly three quarters of a word in English.
• Context windowThe maximum amount of text a model can hold in mind at once, including your prompt, the history and any attached documents.
• Prompt engineeringDesigning the instructions you give a model so it reliably produces the output you need.
• Context engineeringDesigning everything the model can see besides your prompt: files, retrieved documents, memory, tool outputs and history.
• RAG (retrieval-augmented generation)A pattern where relevant documents are retrieved from your own data and inserted into the prompt before the model answers.
• MCP (Model Context Protocol)An open standard that lets any AI client discover and call your tools and data sources through one interface.
• AI agentA system where a model plans a sequence of steps and calls tools to accomplish a goal, rather than just answering once.
• HallucinationOutput that is fluent and plausible but factually wrong or invented.
• EvalA repeatable test set that scores an AI workflow against examples of what 'good' looks like.
• LLM-as-judgeUsing a second model call to grade the output of the first against a written rubric.
• System promptStanding instructions that apply to every message in a conversation or project.
• Chain-of-thoughtAsking a model to reason step by step before answering, which improves accuracy on multi-step problems.
• TemperatureA setting that controls how random a model's output is - low for consistency, high for variety.
• EmbeddingA numeric representation of text that lets software compare meaning rather than keywords.
• Fine-tuningFurther training a model on your own examples so it internalises a format or style.
• MultimodalA model that handles more than text - images, audio, video or screenshots - in the same conversation.
• Vibe codingBuilding working software by describing what you want in natural language and iterating on what the AI produces.
• GuardrailsThe checks around a model that constrain what can go in and what can come out.
• EU AI ActEuropean regulation that classifies AI systems by risk and attaches obligations to each tier.