Unlike a personal assistant tied mainly to one user's history, a team agent operates around a shared function or project. Members can provide task-specific knowledge and review how the agent uses common systems.
The shared design needs role-based permissions, source provenance and visible action history. The agent should not combine private personal context with team work unless users explicitly authorize that boundary.
ELI5
A shared team agent is an AI helper that works inside a group's collaborative space. It uses team-approved context and tools instead of belonging only to one person's private assistant history.
For example, a support team can share an agent that knows current procedures and helps with recurring cases. Each member's role should control what the agent can access, and its actions should be visible to the team.
