Skip to Content
Docs are being rebuilt — start at Introduction → How it works.
IntroductionCore conceptsCore concepts

Core concepts

These pages exist so the rest of the docs can use a word without re-defining it every time. Each entry is short — 2 to 4 paragraphs — answering one question: what does this word mean in TeamMate, and how does it relate to the others?

If you’re already familiar with most of these, skim. If a word feels overloaded (“is a task the same as a workflow?”), the relevant page is the place to come back to.

The eight concepts

ConceptOne-line definition
WorkspaceYour company’s private space in TeamMate. Holds everything else.
AgentA named AI employee with a job, instructions, and personality.
SkillA reusable behaviour an agent can load on demand — a playbook.
Tool & capabilityWhat an agent can do (a tool); the higher-level switch that turns it on (a capability).
Knowledge baseA searchable collection of documents an agent reads from.
ChannelA place users reach an agent — chat, Slack, Teams, email, voice.
InterfaceA pre-built form on top of chat (e.g. an “intake form” that sends a structured message).
Task vs workflow vs triggerThe three words for “how an agent runs without a user typing.”

Suggested reading order

If you’re brand new, the most useful order is:

  1. Workspace — the container.
  2. Agent — the thing you build.
  3. Knowledge base — what the agent reads.
  4. Tool & capability — what the agent does.
  5. Channel — how users reach the agent.
  6. Task vs workflow vs trigger — how the agent runs by itself.

The other three — Team, Skill, Interface — are best read when you actually need them.

A picture if you prefer

The relationship in one ASCII sketch:

Workspace └─ Agent(s) ├─ Instructions (what to do) ├─ Knowledge (what to read) ├─ Tools (what to do, programmatically) ├─ Skills (loadable playbooks) ├─ Channels (how users reach it) └─ Triggers (when it runs by itself)

For the full picture with model choice and governance, see How it works.