AI glossary for business: the jargon and the plain word
This AI glossary collects the twelve terms you hear most in meetings, proposals and articles, and sets each one next to the everyday word that explains it. It’s written for the person who runs a business, or its marketing, sales or customer support. Not for the person who writes the code.
From jargon to plain English
The owner of Crumb & Kettle, a bakery in Boise, came back from a talk on AI with a napkin full of acronyms. This is the napkin, cleaned up. Search for the odd word or the ordinary one, then click to translate it.
Generative AI
The kind that makes thingsIt writes, draws or speaks from what it has learned, instead of only sorting what already exists. Read more about generative ai
LLM
The engine that writesA program that has read an enormous amount and keeps guessing the next word. Read more about LLM
AI assistant
The program you ask for thingsYou write in your own words and it answers, or has the thing ready when you come back. Read more about AI assistant
Chatbot
A program that chats in writingSome follow a script of buttons. Others understand what you type. Read more about chatbot
AI agent
An assistant with a jobThe same assistant, with a standing job and its own instructions. Read more about AI agent
Prompt
The requestWhat you type to ask for something. The fuller the brief, the better the answer. Read more about prompt
Knowledge base
What the company knows, sortedYour documents and your rules, filed by topic so they can be looked up. Read more about knowledge base
Hallucinations
Making up a factThe AI states something false with total confidence, because it sounded right. Read more about hallucinations
RAG
Looking it up in your documents firstIt checks the paperwork, then writes with it in view. Read more about RAG
Fine-tuning
Retraining with your own examplesChanging how an AI responds by showing it cases of yours. Rarely needed. Read more about fine-tuning
Embeddings
A map where similar things sit closeIt turns each text into numbers so it can find the ones that mean something similar. Read more about embeddings
Tokens
The pieces it counts text inAI doesn’t count words. It counts pieces, and they measure how much it reads and what it costs. Read more about tokens
AI terms and definitions, from the general to the specific
Three groups of four. In yellow, the terms that give you your bearings. In pink, the ones you’ll use at work every day. In blue, the ones that describe what happens under the hood.
Generative AI
AI that produces something that didn’t exist before (text, an image, a voice) instead of classifying what’s already there.
LLM (large language model)
The part that does the writing inside an assistant: it continues a text by picking, again and again, a likely next word.
AI assistant
The program you talk to when you want an email, a summary or an answer drafted.
Chatbot
Any program that holds a written conversation, from the button menu on a website to one that understands full sentences.
AI agent
An assistant that has been given one specific job in writing, along with how to do it.
Prompt
The text that tells the AI what you want; the better you explain it, the better the result.
Knowledge base
Your business’s information, gathered by topic so anyone can look it up.
AI hallucinations
False facts an AI delivers with the same confidence as the true ones.
RAG
The AI checks your documents before it answers, instead of going from memory.
Fine-tuning
Retraining an AI on your own examples to change the way it responds.
Embeddings
A way of turning text into numbers so a machine can find passages with a similar meaning.
Tokens
The word pieces an AI splits text into, and the unit its use is measured and billed in.
AI terms explained, in order of urgency
The four AI terms to know first
AI agent. The word of the year, used for nearly everything. Remember it as an assistant with a job.
AI assistant. The starting point: what you’ve already tried if you’ve ever opened ChatGPT.
AI hallucinations. The reason nobody should copy and paste without reading.
Knowledge base. It decides whether an AI talks about your company or about any company.
The ones that can wait
Embeddings, tokens and fine-tuning explain the mechanics. You’ll want them the day you compare vendor quotes or someone starts talking about costs. RAG and LLM sit in the middle: they show up in every proposal. Prompt, chatbot and generative AI take a minute each.
An AI glossary for beginners, not for engineers
Every definition here opens with a sentence you could repeat in a meeting, then shows the idea at work in a small business. There’s no math. Where a definition simplifies, it says so.
If you learn better from cases than from definitions, there are the guides; the one on AI agent vs. chatbot is a good way in. If you’re already choosing a tool, head to the comparisons. And to see these words at work in a product, there’s how Flipa works.
AI vocabulary for business: common questions
Which AI terms should a business owner know?
Four will get you through almost any conversation: assistant, agent, hallucinations and knowledge base. The first two tell you what you’re being offered. The third is the risk. The fourth decides whether it will work with your business.
Do I need to understand how AI works to use it at work?
No, just as you don’t need to understand engines to drive. It does help to keep two ideas in mind: it writes what sounds likely, and it doesn’t know your company unless someone has told it. Those two prevent most unpleasant surprises.
What’s the difference between AI and generative AI?
AI is the whole field: any program that does tasks we associate with intelligence, like recognizing a face or recommending a movie. Generative AI is the part that creates new content. Nearly everything called “AI” in an office today is the generative kind.
How often is this glossary updated?
When a new word starts turning up in meetings and proposals, or when a definition falls short. The last review was in November 2026. Today there are twelve terms, each with its own page.
Fewer acronyms in your way?
Flipa is an assistant that knows your company from its documents and its website. If you’d rather keep reading first, there are the guides.
Start free