AI Glossary

Agentic AI is software that plans, acts, and uses tools to complete multi-step goals with limited human input. Definition, examples, and SMB use cases.

An AI agent is an LLM-driven program that uses tools to complete tasks autonomously. Definition, architecture, and real SMB examples.

An AI receptionist answers calls 24/7, books appointments, and writes to your CRM. Definition, pricing, and how it compares to a human receptionist.

An embedding is a numeric vector that represents the meaning of text, an image, or audio. Definition, top embedding models, and how they power search.

Fine-tuning adapts a pre-trained LLM to a narrow task by training it further on labeled examples. Definition, cost, and when it beats prompting.

A Large Language Model is a transformer-based neural network trained on trillions of tokens to predict the next token. Definition, key models, and business use.

Prompt engineering is the practice of writing instructions to LLMs to get reliable, structured output. Definition, techniques, and when to stop optimizing.

RAG is the technique of fetching documents from a database and feeding them to an LLM before it answers. Definition, architecture, and SMB use cases.

Tool use is when an LLM calls external APIs, databases, or code on its own. Definition, function calling, and how it powers AI agents.

A vector database stores embeddings and finds similar items by approximate nearest-neighbor search. Definition, top vendors, and when you actually need one.

The Claude Agent SDK is Anthropic's open-source framework for building production agents on top of Claude. Definition, capabilities, and when to use it.

Claude Code is Anthropic's terminal-based AI coding agent — reads your repo, runs commands, edits files, and ships PRs. Definition, pricing, and use cases.

Computer Use lets Claude operate a virtual desktop — moving the cursor, clicking, and typing — to complete tasks in any software. Definition and limits.

Claude Haiku is Anthropic's fastest, cheapest model — built for high-volume, low-latency workloads. Definition, pricing, and when to use it.

Claude Opus is Anthropic's most capable model, tuned for deep reasoning, long context, and agentic coding. Definition, pricing, and when to use it.

Claude Skills are reusable, modular instructions and capabilities you can drop into a Claude project. Definition, ecosystem, and use cases.

Claude Sonnet is Anthropic's balanced model — strong reasoning, lower cost than Opus, faster latency. Definition, pricing, and use cases.

Constitutional AI is Anthropic's method for training models to be helpful, harmless, and honest using a written constitution and AI feedback. Definition explained.

Dynamic workflows are LLM-orchestrated processes where the model decides the next step at runtime. Definition, examples, and how they differ from static automation.

MCP is an open standard for connecting AI models to tools, data, and APIs. Definition, ecosystem, and why it matters for AI interoperability.

AI alignment is the problem of making AI systems pursue goals that match human values. Definition, methods, and why it matters for production systems.

AI ethics is the field examining what AI systems should and should not do, and who decides. Definition, principles, and practical SMB implications.

AI evaluation is how you measure whether an AI system actually works. Definition, methods, and why evals are the bottleneck in production AI.

AI governance is the policy and process layer for managing AI risk in an organization. Definition, frameworks, and what SMBs actually need.

Grounding is the practice of tying AI outputs to verified source material. Definition, techniques, and why it is the primary defense against hallucination.

AI guardrails are runtime rules and filters that constrain LLM behavior. Definition, types, and how SMBs should use them in production.

An AI hallucination is when an LLM generates plausible but false information. Definition, why it happens, and how to mitigate it in production.

AI orchestration is the layer that coordinates LLM calls, tools, and data into a working application. Definition, top frameworks, and how to choose.

AI safety is the field focused on making AI systems behave as intended without harmful side effects. Definition, practical risks, and what SMBs should know.

A multi-agent system is a coordinated set of AI agents that divide work and communicate. Definition, patterns, and when it beats a single agent.

An agentic workflow is a multi-step process driven by an AI agent that decides what to do next at each step. Definition, examples, and how to design one.

AI automation uses LLMs and agents to handle work that traditional automation cannot. Definition, examples, and the build-vs-buy math.

An AI co-pilot is an assistive AI embedded in a workflow where a human stays in control. Definition, examples, and how it differs from autonomous agents.

An AI knowledge base is a structured corpus of documents an AI agent retrieves from to answer questions. Definition, architecture, and SMB setup tips.

AI readiness is whether an organization can actually deploy AI safely and usefully. Definition, dimensions, and a practical SMB checklist.

AI ROI is the measurable financial return from an AI deployment. Definition, calculation, and the common traps that fake the numbers.

Conversational AI is software that holds natural dialogue with users in text or voice. Definition, evolution, and what separates working systems from demos.

Customer service AI is the stack of LLM-powered agents handling support tickets, chat, voice, and email. Definition, top vendors, and ROI math.

Voice AI is the stack that lets computers understand and speak natural conversation. Definition, components, top platforms, and SMB use cases.

Workflow automation connects apps and triggers actions across them without human clicks. Definition, top platforms, and where AI changes the game.

AI data privacy covers how personal data is collected, processed, retained, and shared by AI systems. Definition, key laws, and a vendor checklist.

AI disclosure is the legal and ethical obligation to tell users they are interacting with AI. Definition, applicable laws, and SMB practical guidance.

AI implementation is the end-to-end process of deploying an AI workflow from scoping through production. Phases, timeline, and SMB common pitfalls.

An AI pilot is a bounded test of an AI workflow before broader rollout. Definition, structure, and the common reasons pilots fail to graduate to production.

An AI system of record is the authoritative log of AI decisions and interactions. Definition, regulatory drivers, and what SMBs should keep.

AI vendor selection is how SMBs evaluate AI vendors on capability, cost, and risk. A practical 12-question checklist and decision framework.

The build-vs-buy decision for AI depends on scope, talent, time horizon, and total cost. A practical decision framework for SMB owners.

Enterprise and SMB AI projects share technology but differ in budget, scope, timeline, and vendor type. Comparison framework for SMB buyers.

HIPAA-compliant AI handles protected health information under a Business Associate Agreement and meets the HIPAA Security Rule. Definition and vendor checklist.

SOC 2 is an audit framework for vendors handling customer data, including AI services. Definition, Type 1 vs Type 2, and what SMBs should demand.