What is Zahen

Zahen is a governed agentic AI platform. It helps you get answers and complete tasks using your organisation’s own approved knowledge — without the risks people usually worry about with AI.

It is not a general chatbot. Every answer is drawn from documents your organisation has approved and that your role is allowed to read, and anything sensitive stops for a human before it happens.

The four promises

Zahen is built around four guarantees. You’ll see them throughout this guide.

Grounded. Answers come from your approved documents, with citations you can check. If the documents don’t contain the answer, the assistant says so instead of guessing.
Role-aware. You only ever see information your role and department are permitted to access. The same question can return different results for different people — by design.
Human-approved. When the agent proposes something high-risk, it pauses and waits for a person to approve or reject it.
Audited. Every meaningful action — sign-ins, questions, document access, approvals — is recorded in an append-only log.

What you can do with it

Ask questions and get answers grounded in approved policy, with sources. See Asking questions.
Run agent tasks — multi-step work the agent plans and carries out, pausing for approval when needed. See Agent tasks.
Approve or reject sensitive actions, if you’re an approver. See Approvals.
Manage knowledge and your workspace, if you’re an admin. See Managing knowledge and Administration.

What you can do depends on your role — see What your role lets you do.