AI Business Glossary

Agent

Techniques

An AI system that can take multi-step actions on its own toward a goal, such as browsing the web, using tools, or completing a task without step-by-step instructions.

API (Application Programming Interface)

Business & Tools

A way for software to talk to an AI model directly, letting businesses build AI features into their own products rather than using a chat interface.

Chain of Thought

Techniques

A prompting technique where an AI is encouraged to reason through a problem step by step before giving a final answer, often improving accuracy on complex tasks.

Context Window

Core Concepts

The amount of text an AI model can 'see' and remember at once during a conversation, including your prompt and its own prior responses.

Copilot

Business & Tools

An AI assistant built into existing software (like Microsoft Copilot in Word or Excel) to help with tasks inside that tool.

Embedding

Techniques

A numerical representation of text that captures its meaning, allowing AI systems to compare and search content based on meaning rather than exact words.

Few-shot Prompting

Techniques

Giving an AI a few examples of the input/output you want within your prompt, to guide it toward the style or format you're after.

Fine-tuning

Core Concepts

Further training an existing AI model on a specific, narrower dataset so it performs better on a particular task or domain.

Guardrails

Business & Tools

Rules and safeguards put in place to keep an AI tool's behavior within acceptable, safe, or on-brand boundaries.

Hallucination

Core Concepts

When an AI confidently states something false or made-up as if it were fact. A key reason human review of AI output matters.

LLM (Large Language Model)

Core Concepts

An AI model trained on huge amounts of text to predict and generate human-like language. ChatGPT, Claude, and Gemini are all built on LLMs.

Model Weights

Business & Tools

The internal numerical parameters an AI model learns during training. Not something most business users interact with directly, but often referenced in AI news.

Multimodal

Business & Tools

An AI model that can work with more than one type of input or output, such as text, images, audio, or video, rather than text alone.

Prompt

Core Concepts

The instruction or question you give an AI tool. The quality of a prompt has a big impact on the quality of the response.

RAG (Retrieval-Augmented Generation)

Techniques

A technique where an AI model looks up relevant information from an external source (like your documents) before generating an answer, improving accuracy.

Shadow AI

Business & Tools

Unapproved or unsanctioned use of AI tools by employees without company oversight, often a driver behind the need for an AI usage policy.

System Prompt

Techniques

Background instructions given to an AI model that set its role, tone, or rules for a conversation, usually invisible to the end user.

Temperature

Core Concepts

A setting that controls how random or creative an AI's output is. Low temperature gives more predictable answers; high temperature gives more varied, creative ones.

Token

Core Concepts

A chunk of text (roughly a word or part of a word) that AI models use to read and generate language. Usage and pricing are often measured in tokens.

Zero-shot Prompting

Techniques

Asking an AI to complete a task with no prior examples given, relying entirely on its existing training.