The short answer
This AI glossary explains 30 terms CEOs hear in meetings, grouped into the basics, how models work, how to use them, how they connect to your business, and risk. Each entry is one or two plain sentences. If you only remember a few, learn model, prompt, hallucination, agent and MCP, because those five explain most of what vendors will pitch you.
- Thirty terms in six groups, each in plain English.
- Five worth knowing first: model, prompt, hallucination, agent, MCP.
- Most jargon describes a simple idea with a complicated name.
- When a vendor uses a term you do not know, ask them to explain it simply.
The basics
AI: software that learns patterns from data and produces answers or content. Machine learning: the method of training software on examples instead of writing rules. Generative AI: AI that creates text, images, audio or code. Model: the trained system itself, such as Claude Opus or GPT. Chatbot: an interface where you converse with a model. You will hear all of these used loosely, and often interchangeably, in sales pitches. Knowing the difference between them helps you ask better questions and spot vague claims.
How models work
LLM (large language model): a model trained on huge amounts of text to predict and produce language. Training: the process of building the model from data. Parameters: the internal settings learned in training, often used as a rough size measure. Token: a small chunk of text a model reads and writes, about three quarters of a word. Context window: how much text a model can consider at once.
Using AI day to day
Prompt: the instruction you give a model. Prompt engineering: writing prompts to get better results. System prompt: standing instructions that shape every reply. Output: what the model produces. Multimodal: able to handle more than text, such as images, audio or video. Better prompts are usually specific ones: say who the audience is, what the output should look like and what to avoid. Giving an example of what good looks like helps more than a long description.
Connecting AI to your business
Agent: AI that pursues a goal across steps and uses tools. Agentic AI: the approach of building agents. MCP (Model Context Protocol): an open standard that lets an assistant connect to business systems. API: a way for software to talk to other software. Integration: a link between your AI tool and another system. RAG (retrieval-augmented generation): fetching your own documents so the answer is based on them. Sources: Model Context Protocol
Making it better
Fine-tuning: further training a model on your own examples. Embedding: a numeric representation of text that lets software find similar meaning. Vector database: storage built to search those embeddings. Open-weight model: a model whose trained settings are published so anyone can run it. Inference: the act of running a model to get an answer. Most small businesses never need these. Good prompts and a connection to your own documents solve most problems, and these techniques matter mainly when you build a product or run very large volumes.
Risk and trust
Hallucination: a confident answer that is false or invented. Bias: unfair patterns the model picked up from its data. Guardrails: rules that limit what an AI can do or say. Prompt injection: hidden instructions in content that trick an AI into misbehaving. Data privacy: control over what happens to the information you put in. Human in the loop: a person reviews or approves before the AI acts.
Questions people ask
What is the most important AI term for a CEO?
Hallucination. Knowing that AI can be confidently wrong shapes how you review everything it produces.
Is an LLM the same as AI?
No. An LLM is one kind of AI, the kind behind tools like Claude, ChatGPT and Gemini.
What does MCP stand for?
Model Context Protocol, an open standard for connecting AI assistants to the tools and data a business uses.
Do I need to learn all of these?
No. Learn the handful that come up in your own decisions and ask vendors to explain the rest.
No affiliate links on this page at publication. Product names link to official sites so you can check current pricing. How we work.
