Understand every part of an agent system.
Plain-language guides to the concepts behind AI agents—and the exact way AI Agent turns them into usable product features.
01 · Build
Build the system
Define the reasoning, coordination, and repeatable paths behind the work.
Agents
Learn what an AI agent is and how AI Agent combines instructions, models, memory, knowledge, skills, and connected tools in one visual builder.
Agent teams
Learn how AI agent teams group specialists into bounded squads so agents can delegate to the right peers without exposing every agent in a workspace.
Workflows
Learn what an AI workflow is and how AI Agent combines deterministic steps, agent reasoning, tools, branches, approvals, and durable execution.
Autopilots
Learn how AI autopilots turn proven agent workflows into recurring operations with schedules, runbooks, review gates, and run history.
02 · Context
Ground the system
Give agents trusted sources, useful continuity, reusable methods, and live tool access.
Knowledge
Learn how AI knowledge works: ingest documents and websites, retrieve relevant passages, preserve source context, and inspect retrieval traces in AI Agent.
Memory
Learn how agent memory works across recent messages, semantic recall, working memory, and observations—and how AI Agent scopes memory to useful work.
Skills
Learn how agent skills package repeatable instructions and supporting files into reusable runbooks that can be attached to one or more AI agents.
Connections
Learn how AI agent connections provide scoped access to business tools so agents can read live data and propose or take approved actions.
Company Brain
Query a fresh, read-only analytics copy of your Tables across sources, with freshness and provenance included in successful results.
03 · Operate
Run the system
Direct work, review decisions, and turn agent output into durable operations.
Chat
Learn how agent chat combines conversation with tools, knowledge, files, skills, workflow controls, drafts, and traceable actions in AI Agent.
Tables
Learn how AI Agent Tables combine structured fields, views, relations, imports, and agent-computed cells for operational data and company context.
Overview
Learn how AI Agent Overview brings work, agent activity, priorities, and key signals into one configurable control panel for each company.
Inbox
Learn how AI Agent Inbox collects reviews, decisions, errors, and agent-prepared work into a focused queue for human attention.
Board
Learn how AI Agent Board turns agent proposals and human tasks into structured tickets with status, priority, ownership, and evidence.
Performance
Learn how AI Agent Performance aggregates agent and workflow activity so teams can evaluate outcomes, reliability, usage, and cost over time.
Audit Log
Learn how AI Agent Audit Log records individual agent and workflow events so teams can inspect what happened, when, and in which context.
Human approval
Learn how human-in-the-loop AI pauses sensitive agent actions for review, preserves run state, and records the approve or reject decision.
Projects
Learn how AI Agent projects scope work, context, views, ownership, and agent activity around a shared outcome without changing reusable agent definitions.
Start with one agent. Add the system around it as the work earns it.
You do not need every feature on day one. Build the smallest useful agent, then add repeatability, context, and controls where the real work needs them.
Start building