AI Glossary for Business Owners
Plain-English explanations of AI terms. No PhD required. Understand what people mean when they talk about AI.
A
Agent
An AI system that can take actions autonomously—not just generate text, but actually do things like browse the web, run code, or use tools to accomplish tasks.
Example: An AI agent that can research a topic, compare prices across websites, and draft a recommendation report.
API (Application Programming Interface)
A way for different software systems to communicate. AI APIs let businesses add AI capabilities to their existing tools without building AI from scratch.
Example: A CRM might use OpenAI's API to automatically summarize customer calls and add notes to records.
Artificial Intelligence (AI)
Software that can perform tasks that typically require human intelligence—like understanding language, recognizing patterns, making decisions, or learning from experience. AI isn't magic or sentient; it's sophisticated pattern-matching at scale.
Example: A chatbot that understands customer questions and provides relevant answers is using AI.
Automation
Using technology to perform tasks without human intervention. AI-powered automation can handle complex tasks that previously required human judgment.
Example: Automatically categorizing incoming emails and routing them to the right department based on content.
C
Chatbot
Software that simulates conversation with users, typically through text. Modern chatbots use AI to understand intent and provide relevant responses. Older chatbots just matched keywords to scripted responses.
Example: The chat widget on a website that answers questions about business hours and can schedule appointments.
Computer Vision
AI that can interpret and understand images and video. Used for everything from quality control in manufacturing to analyzing medical images.
Example: A system that automatically detects defects in products on an assembly line by analyzing camera footage.
Context Window
The amount of text an AI can consider at once—its 'working memory.' Larger context windows mean the AI can handle longer documents or conversations without 'forgetting' earlier content.
Example: GPT-4 Turbo has a 128K token context window, meaning it can process roughly a 300-page book at once.
E
Embedding
A way of representing text (or images, etc.) as numbers that capture meaning. Similar concepts have similar embeddings, which helps AI find relevant information.
Example: A search system might use embeddings to find documents related to a query even if they don't share exact keywords.
F
Fine-tuning
Taking a pre-trained AI model and training it further on specific data to make it better at particular tasks. It's cheaper and faster than training from scratch.
Example: A law firm might fine-tune a general language model on legal documents so it better understands legal terminology.
G
Generative AI
AI that creates new content—text, images, code, audio—rather than just analyzing existing content. The 'generative' part means it produces something new, not just categorizes or predicts.
Example: Midjourney creating an image from a text description, or ChatGPT writing a blog post draft.
GPT (Generative Pre-trained Transformer)
A specific type of language model architecture developed by OpenAI. GPT-4 is the model behind ChatGPT. 'Pre-trained' means it learned from existing text before being fine-tuned for conversations.
Example: ChatGPT Plus uses GPT-4 to have more nuanced, capable conversations than earlier versions.
H
Hallucination
When an AI generates information that sounds plausible but is factually incorrect or made up. LLMs don't 'know' facts—they predict likely text, which can lead to confident-sounding errors.
Example: An AI might cite a research paper that doesn't exist, complete with fake authors and publication details.
L
Large Language Model (LLM)
An AI system trained on massive amounts of text that can understand and generate human-like language. ChatGPT, Claude, and similar tools are LLMs. They're called 'large' because they have billions of parameters.
Example: When you ask ChatGPT a question and it writes a coherent response, that's an LLM at work.
M
Machine Learning (ML)
A type of AI that improves through experience rather than explicit programming. Instead of coding every rule, you show the system examples and it learns patterns. Most modern AI is machine learning.
Example: An email spam filter learns from examples of spam vs. legitimate email, getting better over time.
Model
The trained AI system itself—the result of processing training data through a learning algorithm. When people talk about 'GPT-4' or 'Claude,' they're referring to specific models.
Example: A company might use a pre-trained model and fine-tune it for their specific industry.
N
Natural Language Processing (NLP)
AI technology that enables computers to understand, interpret, and respond to human language. It's what allows you to ask a question in plain English instead of using specific commands.
Example: When you ask Siri 'What's the weather?' and it understands you want a forecast, that's NLP.
O
OCR (Optical Character Recognition)
Technology that converts images of text into actual text data. Modern OCR uses AI to handle handwriting, poor image quality, and complex layouts.
Example: Scanning paper receipts and automatically extracting vendor, date, and amount into expense software.
P
Prompt
The input you give to an AI system—the question, instruction, or context that tells it what you want. Prompt quality heavily influences output quality. 'Prompt engineering' is the skill of crafting effective prompts.
Example: 'Write a professional email declining a meeting' is a prompt. A better prompt would include context, tone, and specific details.
Prompt Engineering
The practice of crafting effective instructions for AI systems to get better results. It's part art, part science—understanding what context helps the AI and how to structure requests clearly.
Example: Instead of 'write about AI,' a prompt engineer might write: 'Write a 200-word explanation of AI for small business owners with no technical background. Use simple analogies.'
R
RAG (Retrieval-Augmented Generation)
A technique where AI retrieves relevant information from a knowledge base before generating a response. This helps ground responses in actual data rather than relying purely on training.
Example: A company chatbot that searches internal documentation to answer employee questions, rather than making up answers.
RPA (Robotic Process Automation)
Software that automates repetitive, rule-based tasks—like copying data between systems or filling out forms. Traditional RPA follows exact rules; AI-enhanced RPA can handle variations.
Example: A bot that automatically transfers invoice data from emails into accounting software.
S
Sentiment Analysis
AI that determines the emotional tone of text—positive, negative, neutral. Commonly used to analyze customer feedback, reviews, and social media mentions.
Example: Automatically flagging negative customer reviews for immediate follow-up by the support team.
T
Token
The basic unit that language models use to process text. Roughly, a token is about 4 characters or ¾ of a word. AI pricing and context limits are often measured in tokens.
Example: The sentence 'AI helps businesses' is about 4-5 tokens. API costs are often fractions of a cent per thousand tokens.
Training Data
The examples used to teach an AI system. The quality and quantity of training data significantly affects how well the AI performs. Garbage in, garbage out applies here.
Example: A customer service AI is trained on thousands of past support conversations to learn how to respond to common questions.
V
Vector Database
A database optimized for storing and searching embeddings. Essential for building AI applications that need to quickly find relevant information from large datasets.
Example: A customer service AI uses a vector database to find relevant help articles based on customer questions.
Still have questions?
Understanding the terminology is step one. Figuring out what actually matters for your business is step two. We can help with both.
Book a Free Discovery Call