# Workspace Administration

Each workspace in OneByZero Neo is a self-contained collaborative environment for a department or team. Members can build internal agents, organisation-wide agents, or customer-facing agents — depending on their role. You can switch between workspaces using the workspace switcher in the top-left corner of the platform.

Workspace Administrators manage everything within that boundary: who can access the workspace, what they can do, which resources they can reach, and the variables and secrets that agents use at runtime.

Workspace settings are opened from **Settings** in the sidebar. Each tab is shown only to members whose role includes the permission to view it — Workspace Admins see all of them.

## Workspace Settings Tabs

The Workspace Settings page has the following tabs:

- **Members** — a menu with four sections:
  - **Users** — invite people to the workspace and manage their role and policies. See [Member Management](member_management.html).
  - **Identities** and **User directories** — the people your agents have talked to on external channels, and how they are linked across channels. See [Identities and User Directories](identities.html).
  - **Service accounts** — credentials that let external systems call the OneByZero Neo API. See [Service Accounts](credentials.html).
- **Policies** — define which resources members and service accounts can access, and at what level. See [Policies](policies.html).
- **Roles** — define roles as named collections of permissions. See [Roles and Permissions](roles_permissions.html).
- **Secrets** — a menu with **AWS Secrets Manager**, **HashiCorp Vault**, and **Variables**. See [Variables and Secrets](variables_secrets.html).
- **Workspace** — the workspace’s icon, name, and ID. See [Workspace Setup](workspace_setup.html).

Roles and policies work together. A **role** controls which features a member can use — for example, whether they can create agents or manage knowledge bases. A **policy** controls which specific resources they can use — for example, which agents, datasets, or knowledge bases, and whether they can view, run, or edit them.