# AI Glossary for Designers: Key Terms Explained in One Line

> An AI glossary for designers with 53 terms explained in one line: models, prompts, agents, evaluation and UX. Each one links to its article in the wiki.

*Tags: ai, glossary, reference, design*

---


> [!info] Quick Definition
> This **AI glossary for designers** gathers 53 terms that appear in the wiki, grouped by topic with a one-line definition. Each term links to its full article. If a term is not here, it probably has its own article: look in the section index.


## Which terms explain how a model works?

- [[what-is-an-llm|LLM]]: a large language model, trained on lots of text to predict the next token.
- [[tokens|Token]]: the unit of text a model processes: a word, a fragment or a symbol. Size and cost are measured in tokens.
- [[context-window|Context window]]: the maximum number of tokens a model handles in one request, including its own answer.
- [[embeddings|Embedding]]: a list of numbers representing a text or image, so that similar things end up close together.
- [[knowledge-cutoff|Knowledge cutoff]]: the date up to which training data reaches; anything later is unknown to the model on its own.
- [[hallucinations|Hallucination]]: a false answer delivered with confidence, an effect of how text is generated and not an intentional lie.
- [[reasoning-models|Reasoning model]]: a model that produces intermediate steps before answering, at the price of more time and cost.
- [[multimodal-models|Multimodal model]]: a model that processes more than one kind of data, such as text, images, audio or video.
- [[temperature-and-top-p|Temperature and top-p]]: parameters that control how varied or conservative the choice of the next token is.
- [[open-vs-closed-models|Open vs. closed model]]: weights available to host, as opposed to a model reachable only through an API.

## How do you give a model instructions and data?

- [[prompt-engineering|Prompt engineering]]: designing instructions, context, examples and format to get useful, consistent results.
- [[system-prompt|System prompt]]: permanent instructions that define the model's role, tone and limits.
- [[few-shot-prompting|Few-shot prompting]]: including a few input and output examples so the model follows the pattern.
- [[chain-of-thought|Chain of thought]]: getting the model to write intermediate steps before the final answer.
- [[structured-outputs|Structured output]]: a response in a fixed format, usually JSON, that an application can read without interpreting free text.
- [[context-engineering|Context engineering]]: deciding what goes into the model at each step, not only the prompt.
- [[rag|RAG]]: retrieving relevant passages from documents and adding them to the request so the answer is grounded in them.
- [[fine-tuning-vs-prompting|Fine-tuning vs. prompting]]: further training the model on your data, as opposed to guiding it with instructions in each request.
- [[prompt-caching|Prompt caching]]: reusing parts of the prompt already processed to lower cost and latency.

## Which terms describe an agent?

- [[what-is-an-ai-agent|AI agent]]: a system where a model receives a goal, decides the steps and uses tools until it finishes.
- [[agent-vs-assistant-vs-copilot|Assistant, copilot and agent]]: the first responds, the second suggests inside your task, the third runs several steps.
- [[tool-use|Tool use]]: the model asks to run an external function; the application runs it and returns the result.
- [[agent-loop|Agent loop]]: the cycle repeated until done: the model decides, a tool runs and the result comes back.
- [[mcp|MCP]]: an open standard for connecting AI applications to files, databases and tools.
- [[autonomy-levels|Autonomy levels]]: a five-level scale (Feng, McDonald and Zhang) based on the role the person takes toward the agent.
- [[human-in-the-loop|Human-in-the-loop]]: a person approves or corrects at the points where an error costs the most.
- [[plan-mode|Plan mode]]: the agent investigates and proposes changes, but does not act until you approve the plan.
- [[subagents|Subagent]]: a specialized agent with its own context and permissions that returns only the final result.
- [[agent-skills|Skill]]: a folder with a `SKILL.md` that teaches the agent a procedure and loads only when needed.
- [[sandboxing-and-permissions|Sandboxing and permissions]]: operating-system isolation plus rules on which tools need approval.
- [[computer-use|Computer use]]: the model observes the screen and operates mouse and keyboard as a person would.

## How do you measure and control quality?

- [[what-are-evals|Evals]]: tests for an AI system; the equivalent of automated tests for behavior that varies.
- [[golden-datasets|Golden dataset]]: test cases with reference answers, used to compare versions in a repeatable way.
- [[llm-as-judge|LLM as judge]]: a model grades answers against a written criterion; it is only reliable if you validate it with people.
- [[ai-benchmarks|Benchmark]]: a standardized test for comparing models; the score only describes what the test measures.
- [[prompt-regressions|Prompt regression]]: a change that makes cases that used to work fail.
- [[guardrails|Guardrails]]: controls that check inputs, outputs or actions and can block them; they do not guarantee total safety.
- [[red-teaming|Red teaming]]: attacking the system on purpose to find failures before someone else does.
- [[model-bias|Model bias]]: a repeatable tendency that gives worse or unfair results for certain groups.

## Which terms matter when designing the experience and the product?

- [[response-streaming|Streaming]]: showing the answer in pieces as it is generated; the model is not faster, the text just appears sooner.
- [[trust-calibration|Trust calibration]]: making the person's trust match the system's real reliability.
- [[showing-uncertainty|Uncertainty]]: communicating how sure the system is and how to verify what it says.
- [[explainability|Explainability]]: giving the user what they need to know where a result comes from and when to check it.
- [[human-handoff|Human handoff]]: passing the conversation to a person with enough context that nobody repeats everything.
- [[ai-dark-patterns|AI dark patterns]]: designs or dialogue that push the user to believe or do something they would not have chosen.
- [[generative-ui|Generative UI]]: an interface where the model decides which components to show, instead of them being fixed in advance.
- [[ai-first-vs-ai-added|AI-first vs. AI-added]]: a product designed around AI from the start, as opposed to AI added to an existing one.
- [[what-is-aeo|AEO]]: a marketing term for preparing a site so AI answer engines cite it; it is not a standard.
- [[llms-txt|llms.txt]]: a proposed Markdown file with a site summary and key links for AI agents.
- [[what-is-vibe-coding|Vibe coding]]: building software by describing it to an AI and accepting the code without reviewing it closely.
- [[design-engineer|Design engineer]]: a profile between design and frontend that prototypes and builds in code; the role has no standard definition.
- [[design-tokens-for-agents|Design tokens]]: named design values in a standard format, which an agent can use without guessing a hex.
- [[code-connect|Code Connect]]: a Figma tool that links a Figma component to its code component.

## Example

Illustrative example: in a meeting someone says "the agent hallucinated because it ran out of context window". With this glossary you can separate the terms. The **context window** is the token limit per request. If the agent lost its initial instructions, the problem is **context engineering**. If it invented a fact, that is a **hallucination**, and the defenses are **RAG** with **citations**, or an **eval** that catches it. Three different problems that sounded like one in the meeting.

## Common mistakes

- **Using "agent" for any AI chat.** An assistant that responds and an agent that runs several steps with tools call for very different permission and supervision designs.
- **Believing temperature 0 guarantees the same answer.** Anthropic's documentation states that even then results are not fully deterministic.
- **Confusing RAG with fine-tuning.** RAG adds documents to the request; fine-tuning changes the behavior of the model itself.
- **Treating a benchmark as a verdict.** The score describes performance on that test, not on your product; for that you need your own evals.
- **Using AEO as a guarantee.** It is a marketing term, not a feature with its own settings.

## Sources

- [Glossary (Claude documentation)](https://platform.claude.com/docs/en/about-claude/glossary)
- [What is the Model Context Protocol (MCP)? (modelcontextprotocol.io)](https://modelcontextprotocol.io/docs/getting-started/intro)
- [Building effective agents (Anthropic)](https://www.anthropic.com/engineering/building-effective-agents)
- [Levels of Autonomy for AI Agents (Feng, McDonald and Zhang, arXiv:2506.12469)](https://arxiv.org/abs/2506.12469)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks (Lewis et al., arXiv:2005.11401)](https://arxiv.org/abs/2005.11401)
- [The /llms.txt file (llmstxt.org)](https://llmstxt.org/)


---

Source: https://www.fernandoux.com/en/wiki/ai/ai-glossary-for-designers/