AI Glossary
Your go-to reference for AI terminology. Plain-English definitions, real-world examples, and context for every term.
A private area where staff manage app content, users, and settings.
Measurements and insights about user behavior and system performance.
A way for software systems to communicate and share data with each other.
The task that happens when a trigger fires in an automation.
The process of proving identity to access a system.
Using technology to perform tasks without manual intervention.
Artificial Intelligence—software that can make predictions, generate content, or assist with decisions.
AI that can take multiple steps toward a goal, often using tools or making decisions.
The danger of exposing sensitive information to AI systems.
Systems and policies determining who can access what resources.
A record of who did what and when in a system.
Comparing two versions to see which performs better.
Determining which marketing touchpoints deserve credit for conversions.
A product layer built around another company's AI model or API.
A software layer where AI agents coordinate tasks across tools, systems, or business processes.
A marketplace where users discover, download, and update mobile or platform apps.
The application used to access and view websites, like Chrome, Safari, or Firefox.
The server-side logic, databases, and processes that power a website behind the scenes.
A prioritized list of possible product work, bugs, features, and improvements.
A tool for moving crypto assets or information between blockchains.
Saved copies of website data stored locally to speed up repeat visits.
A network of servers distributed globally to deliver website content faster.
A content management system that lets non-developers edit website content.
Comma-Separated Values—a simple text file format for storing tabular data.
How much information an AI can consider at once when generating a response.
A conversational AI interface that responds to text or voice inputs.
Legal uncertainty about using AI-generated content.
Sensitive information that should not be shared with external AI systems.
Meeting legal, regulatory, or industry standards for data handling and security.
Call to Action—the specific action you want visitors to take.
Customer Relationship Management—software for tracking leads and customers.
A formal request to modify scope after a project has started.
Internet-based computing resources such as servers, databases, storage, and services.
The rate at which customers stop using or paying for a product.
A digital asset issued on a blockchain that can represent value, access, governance, or utility.
The address people type to visit your website, like example.com.
The system that translates domain names into server addresses.
A website that generates pages based on data, user, or context.
A visual interface showing key metrics and information at a glance.
Organized storage for application information that can be searched and updated.
A visual display of key metrics and KPIs from your data.
The structure that defines how data is organized, stored, and related.
Policies for how long data is kept before deletion.
A tangible item or output produced as part of a project.
A central place where business data from many systems is collected for reporting and analysis.
External code, service, or tool that your software relies on.
A claimed advantage where more users create more valuable data, making the product harder to copy.
Decentralized autonomous organization: a group that uses blockchain-based rules or voting to coordinate decisions.
Decentralized finance: financial products built with blockchain-based protocols rather than traditional intermediaries.
Software that adds features to a browser or application.
A specific URL where an API receives requests.
A numerical representation of content that captures its meaning.
Scrambling data so only authorized parties can read it.
Extract, transform, load: moving data from one place, cleaning it, and putting it somewhere useful.
The standards used to judge whether an AI response is good enough for its purpose.
The visual part of a website that users see and interact with.
Training an existing AI model further with specific data for a particular purpose.
The path from awareness to purchase, typically narrowing at each stage.
A switch that lets teams turn a feature on or off without redeploying code.
Providing examples of good input and output before asking AI to complete the real task.
A copy of a code project that can be changed separately from the original.
The process for making decisions about changes, rules, funds, or direction.
The service that stores your website files and makes them accessible online.
When AI generates confident-sounding but false or fabricated information.
Having a person verify AI output before using or publishing it.
A connection between two or more software systems that share data.
Using a trained AI model to generate predictions or outputs.
The cost of generating AI outputs when users ask the model to do work.
A tracked bug, task, question, or feature request in a software project.
Know Your Customer: identity checks used to verify users, especially in finance and crypto.
A focused webpage designed to convert visitors toward a specific action.
Development platforms that minimize hand-coding through visual interfaces and templates.
Large Language Model—AI trained on massive text data to understand and generate language.
The legal permission terms for using software, content, or data.
How easily an asset can be bought or sold without significantly moving its price.
Software designed specifically for smartphones and tablets, downloaded from app stores.
Minimum Viable Product—the simplest version that delivers core value.
A visual automation platform with more flexibility than Zapier for complex workflows.
AI that improves through experience and data rather than explicit programming.
The trained AI system that produces outputs based on inputs.
AI that can work with multiple types of content: text, images, audio, or video.
When AI behavior changes over time, often unexpectedly.
Multi-Factor Authentication—requiring multiple proofs of identity to log in.
Malicious software designed to damage, disrupt, or gain unauthorized access.
A significant checkpoint in a project timeline.
Ongoing care and updates required after launch.
A company that supplies AI models through an API or platform.
A person or team responsible for reviewing, updating, and guiding a software project.
A durable advantage that makes a business hard to copy or compete with.
Model Context Protocol: a standard way for AI apps to connect to external tools and data sources.
An app built specifically for one platform using that platform's programming language.
Tools that let you build apps and automations without writing code.
A standard for granting applications limited access without sharing passwords.
Work explicitly not included in the current project agreement.
The structure you request for an AI response, such as a table, checklist, email, or summary.
Coordinating multiple tools, steps, models, or systems in an automated workflow.
Add-on software that extends the functionality of an existing application.
The actual data sent in an API request or response.
The instruction or question you give to an AI to get a response.
Malicious input designed to manipulate AI behavior.
Software that securely stores and generates strong, unique passwords.
Controls defining who can access, view, or modify specific data or features.
Fraudulent attempts to steal information by impersonating trusted entities.
Personally Identifiable Information—data that can identify a specific individual.
A small tracking script that monitors visitor behavior.
Designing instructions and context so an AI model produces better, more consistent results.
The perspective or expertise you ask an AI model to use when responding.
The background information an AI model needs to complete a task well.
A boundary that tells AI what to include, avoid, limit, or prioritize.
Improving a prompt over multiple attempts by diagnosing what was wrong with the output.
A proposed code change submitted for review before it is merged.
Product-market fit: evidence that a product strongly satisfies a real market need.
A meaningful change in strategy based on evidence that the current approach is not working well enough.
A shared set of rules that systems or participants use to interact.
A service that securely processes online payments between customers, merchants, and banks.
Quality Assurance—testing to find and fix problems before launch.
Website design that automatically adjusts layout based on screen size.
A timeline of planned features, improvements, and milestones.
A cap on how many requests can be made to an API in a given time period.
Processing and displaying data immediately as events occur.
Retrieval-Augmented Generation—giving AI access to external information to improve accuracy.
A storage place for a software project's code, history, and collaboration notes.
Remote procedure call: a way for one system to ask another system to run a specific function.
How long a company can keep operating before it runs out of money at its current burn rate.
Using already staked crypto assets to help secure additional services or protocols.
A computer that stores files and responds to requests from other computers.
The security protocol that encrypts data between browsers and servers, shown by the padlock icon.
A website with pre-built pages that are the same for every visitor.
Software as a Service—software you rent through a subscription instead of buying outright.
A group of users or data points that share common characteristics.
Verifying AI claims against reliable sources.
Search Engine Optimization—improving visibility in search results.
The defined boundaries of what is included in a project.
A software development kit: prebuilt tools developers use to work with a platform faster.
Single sign-on: one login that gives users access to multiple approved tools.
The defined structure of data, including fields, types, and relationships.
A short planned work period, often one or two weeks, used by software teams.
Increasing a system's ability to handle more users, data, or activity.
Code on a blockchain that automatically executes rules or transactions.
Locking or committing crypto assets to help secure a network or earn rewards.
Searching by meaning rather than exact keyword matches.
A collection of related data organized into rows and columns within a database.
A chunk of text that AI reads or writes, roughly 3-4 characters or about 0.75 words.
Shortcuts taken now that create problems or extra work later.
Evidence that customers or users are adopting and valuing a product.
The economic design of a token, including supply, incentives, distribution, and utility.
A saved identity that lets someone log in and access personalized features.
Tracking parameters added to URLs to identify traffic sources.
A short description of what a user needs and why.
The percentage of time a website, app, or service is working and available.
User interface: the screens, controls, and visual elements people use to interact with software.
User experience: how easy, useful, and satisfying a product feels across the whole journey.
A database designed to store and search AI-friendly representations of content.
Being stuck with a vendor because switching is too difficult or expensive.
Software that runs in a web browser and does not require installation.
A defined sequence of steps to complete a task or process.
A notification sent automatically from one system to another when something happens.
A tool that lets someone hold and use crypto assets by controlling private keys.
A popular no-code tool for connecting apps and automating workflows.
A blockchain scaling method that bundles transactions and uses zero-knowledge proofs to verify them.