AI glossary.

Plain-English explanations for the words people throw around when building AI websites, automations, and business systems.

Security

API key

A private code that lets software call an AI service. Keep it on the backend, never inside public browser code.

Connection

API

A way for two apps to talk to each other, like a website sending a lead to a CRM.

Automation

Webhook

An automatic message sent from one app to another when something happens, like a form submission.

AI input

Prompt

The instruction you give AI. Better prompts include goal, context, rules, examples, and output format.

AI system

Agent

An AI workflow that can take steps, use tools, and move toward a goal with rules and approvals.

Knowledge

RAG

Retrieval augmented generation: AI searches your docs first, then answers using the most relevant information.

Data

Embedding

A numerical representation of text or content that helps AI search by meaning, not just exact words.

Data

Vector database

A database built for similarity search, often used when AI needs to find relevant docs or FAQs.

Cost

Token

A small chunk of text used by AI models. Usage and pricing often depend on how many tokens are processed.

Build

Backend

The server-side part of a website. This is where API keys, scheduled jobs, and private logic should live.

Build

Frontend

The part visitors see in the browser: pages, buttons, forms, animations, and visible content.

Schedule

Cron job

A scheduled task that runs automatically, like publishing a weekly AI learning post.

Business

CRM

A system for tracking leads, contacts, pipeline stages, follow-up, and booked work.

Workflow

Automation

A rule-based process that moves work forward after a trigger, like sending a reminder after no reply.

Safety

Guardrail

A rule that keeps AI in bounds, like requiring human approval before sending a customer message.