Policy assistant
A private employee chat that answers from your organisation's uploaded policies — with citations, not guesses.
This is different from MOOSH Assist (the HR drafting workspace). The policy assistant is a private employee conversation grounded on company policies your organisation uploaded during onboarding.
Every answer names the policy and version it used. HR, managers, and org admins cannot read the chat. Memory is treated as highly sensitive personal data and can be cleared. Canonical actions need explicit preview and confirmation; the bot can link to human help but does not file reports for you.
If AI is unavailable, a deterministic search fallback still points at the right policy without inventing wording. Live use remains gated pending agent-safety approval.
For example
Jess asks how to raise a confidential grievance; the assistant cites Grievance Policy v3 and offers a “Talk to a human” path to Support — HR never sees the conversation.
Worth knowing
- Citation
- The exact policy document and version the answer was grounded on.
- Policy document
- A versioned company policy file kept in a separate, access-controlled corpus — not mixed with anonymous check-in answers.
What you can do here
- 1Ask a sample policy question and read the cited answer.
- 2Clear conversation history when you want it gone.
- 3Use the degraded-AI fallback when the assistant is unavailable.
- 4Open Support to talk to a human.