The AI Glossary
A simple, no-nonsense dictionary for the buzzwords and technical jargon you'll encounter on your AI readiness journey.
🧠Core Concepts (Foundation Layer)
Artificial Intelligence (AI)
Broad idea of machines performing tasks that typically require human intelligence.
Plain English
A broad term for machines doing things that usually need a human brain, like seeing, reasoning, or solving problems.
Practical Example
"Your email automatically filtering spam or sorting primary messages."
Machine Learning (ML)
Systems that learn patterns from data instead of being explicitly programmed.
Plain English
Teaching computers to recognize patterns by showing them lots of examples, rather than giving them rigid step-by-step instructions.
Practical Example
"Netflix recommending movies because it learned what people with similar tastes watch."
Large Language Model (LLM)
AI models trained on massive text data to understand and generate language. Examples: ChatGPT, Claude
Plain English
A giant autocomplete engine that has read millions of books and articles to learn how humans write and speak.
Practical Example
"Asking ChatGPT to draft an email and it writing perfect paragraphs instantly."
Model
The trained system itself. Think of it as the “brain” behind AI.
Plain English
The software 'brain' that was created after the AI finished its training phase.
Practical Example
"GPT-4 is the model powering ChatGPT Plus."
Training Data
The information used to teach a model. Quality here directly impacts output.
Plain English
The textbooks, articles, and websites the AI studied to get smart.
Practical Example
"If an AI is only trained on math books, its training data makes it great at algebra but terrible at poetry."
đź’¬ How You Interact with AI
Prompt
The input or instruction you give to an AI system.
Plain English
The question or command you type into the chatbox.
Practical Example
"Typing 'Write a 5-day diet plan for a diabetic' into an AI chat."
Prompt Engineering
Designing better prompts to get better results. Less about tricks, more about clarity and structure.
Plain English
The skill of giving incredibly clear instructions to an AI so it does exactly what you want.
Practical Example
"Instead of 'write a blog', saying 'Act as a fitness coach and write a 500-word blog for beginners avoiding jargon.'"
Context
Additional information you provide to guide AI responses. Example: your role, data, constraints, goals
Plain English
The background details you give the AI so its answer is customized to your specific situation.
Practical Example
"Pasting your company's refund policy into the chat before asking it to write an email to a loud customer."
Token
A unit of text (word or part of a word) that AI processes. Important because it affects: • cost • speed • output length
Plain English
Chunks of words the AI reads at a time. It's the 'currency' of AI processing.
Practical Example
"The word 'apple' is one token. A giant PDF might be 50,000 tokens."
⚙️ How AI Actually Works (Simplified)
Inference
When the model generates a response based on your prompt.
Plain English
The act of the AI 'thinking' and typing out the answer for you in real-time.
Practical Example
"When you hit enter and see the text streaming in, the AI is performing inference."
Fine-Tuning
Training a model further on specific data to specialize it.
Plain English
Taking a smart general AI and sending it to 'trade school' for a specific job.
Practical Example
"Taking an open source AI and giving it 10,000 legal contracts so it becomes a specialized lawyer bot."
Embedding
Turning text into numerical vectors so machines can “understand” similarity.
Plain English
Translating words into numbers so the computer can mathematically calculate which ideas are related.
Practical Example
"An embedding knows that 'puppy' and 'dog' are mathematically closer than 'puppy' and 'toaster'."
Vector Database
Stores embeddings to enable smart search (semantic search).
Plain English
A special filing cabinet that organizes information by its meaning, rather than just exact keywords.
Practical Example
"Searching for 'pet' and it successfully finds documents only containing the word 'dog'."
RAG (Retrieval-Augmented Generation)
AI pulls relevant data first, then generates answers. → Reduces hallucination, improves accuracy
Plain English
Making the AI do an open-book test. It searches your files first, then reads them to answer your question.
Practical Example
"A chatbot hooked up to your company wiki that cites exactly which policy document it got its answer from."
🤖 Emerging “Agentic” Concepts (Very Important Now)
AI Agent
A system that can take actions, not just generate text. Example: reading data → making decisions → calling tools
Plain English
An AI that doesn't just talk, but has 'hands' to do things for you automatically.
Practical Example
"An AI that reads your email, decides it's a meeting request, checks your calendar, and replies with available times."
Tool Calling
AI using external systems (APIs, databases, apps) to complete tasks.
Plain English
Giving the AI the ability to press buttons in other software.
Practical Example
"The AI using a 'Weather Tool' to fetch live weather data before answering your question."
Workflow / Orchestration
Coordinating multiple steps or agents to complete a task.
Plain English
A multi-step assembly line where different tasks happen in a specific order.
Practical Example
"First an AI summarizes an email, then another AI categorizes it, then a final one pushes it into a spreadsheet."
Multi-Agent System
Multiple AI agents working together, each with a role.
Plain English
A digital team of workers collaborating to finish a big project.
Practical Example
"One AI agent acts as a researcher gathering facts, while another acts as a writer drafting the report."
đź“Š Quality, Risk, and Reality
Hallucination
When AI generates incorrect or made-up information confidently.
Plain English
When the AI doesn't know the answer, so it confidently makes something up.
Practical Example
"The AI inventing a historical event that never happened, but citing fake books to prove it."
Evaluation (Eval)
Measuring how good an AI system is. Examples: • accuracy • usefulness • safety
Plain English
The grading system used to see if a new AI feature is actually better than the old one.
Practical Example
"Running 1,000 test questions to see if the AI bot gives correct answers more than 95% of the time."
Guardrails
Controls to prevent harmful or incorrect outputs.
Plain English
Digital bumpers in the bowling alley to keep the AI from going off track.
Practical Example
"A hidden rule that blocks the AI from ever giving medical advice or using profanity."
Latency
Time it takes for AI to respond.
Plain English
The lag or waiting time before the AI starts answering.
Practical Example
"Waiting 5 seconds for Siri to process what you said."
Cost per Request
How much each AI interaction costs (important for business use).
Plain English
The tiny fee charged by AI providers for every question you ask it.
Practical Example
"Paying $0.01 every time a customer talks to your website chatbot."
đź’Ľ Business & Practical Layer (Where most value is)
Use Case
A specific problem AI is solving. Example: customer support automation
Plain English
The actual real-world job you are hiring the AI to do.
Practical Example
"Using AI to summarize hour-long Zoom meetings into bullet points."
Automation vs Augmentation
• Automation → AI replaces human work • Augmentation → AI enhances human work Most real value today is augmentation.
Plain English
Automation means the robot does the job alone. Augmentation means the robot is a super-suit helping a human do the job better.
Practical Example
"Automation: a factory robot painting a car. Augmentation: an AI suggesting marketing copy to a writer."
AI Readiness
How prepared a person or organization is to effectively use AI.
Plain English
Checking if you have the right mindset, tools, and data to actually benefit from AI without failing.
Practical Example
"Realizing your messy Google Drive means you aren't ready to let an AI read your files yet."
ROI (Return on Investment)
Value gained from AI relative to cost. This is what leadership actually cares about.
Plain English
Figuring out if the money saved by using AI is greater than the money spent building it.
Practical Example
"A $5,000 AI tool saving $50,000 in customer service hours."
Ready to test your knowledge?
Now that you know the lingo, find out exactly how ready you are for the AI era.
Take the Assessment