AI teammates
AI teammates that do real work, inside your real projects
Give an agent a job — plan this launch, write Friday's update, rebalance next week — and it works on your actual tasks, with the permissions you set.
What you get
Agents with a role
Planner, reporter, analyst, writer — each with its own skills.
Your permissions apply
An agent can only see and do what you allow.
You stay in control
Review before it sends, or let trusted agents run.
Shared memory
Agents learn your projects and your preferences.
Hand off the work nobody enjoys
Status updates, weekly digests, follow-ups and first drafts — done in the background and ready for review.
- Runs on your real data
- Every action is logged
- Approve, edit or reject
Agents that know the project
They read tasks, comments and plans, so answers are specific to your work, not generic.
- Grounded in your workspace
- Source links on every answer
- Private to your organisation
FAQ
About ai teammates
No. Your content is not used to train shared models.
Yes. Choose its projects, actions and whether it needs approval.