Policies and procedures are the lifeblood of a consistent organization, but let's be honest: finding the exact rule you need in a massive library of documents can feel like searching for a needle in a haystack.
That’s why we’ve introduced a smarter way to search. Instead of hunting for keywords, you can now use our AI-powered search to simply "ask" your documentation a question.
How It Works
We have trained an AI to "read" and understand your policies just like a person would. This means you can ask questions in plain English—like you're asking a coworker—and get a meaningful answer back. This assistant runs on models from OpenAI and Anthropic (the makers of Claude) working together behind the scenes, so you don't need to think about which one is answering.
To get started:
Look for the special icon (white circle with stars) at the top of the editor.
Click it to open the conversational window.
Type your question (e.g., "What is the process for requesting time off?").
Ask and Plan Modes
At the bottom of the AI window, the mode button lets you switch between two modes:
Ask: Ask questions about your existing policies and procedures. Answers include numbered sources you can click to open the document the answer came from.
Plan: Describe a change you want made, such as "Add a sentence to article 1.5 saying it is reviewed annually." The AI reads the document and shows you a proposed edit, with new text highlighted in green, before anything is changed.
When the AI shows a proposed edit, you choose what happens next:
Accept: Apply the proposed change to the document.
Accept and Request Review: Apply the change and send it for review.
Reject: Discard the proposal. The AI will ask what you would like done instead.
To start fresh, click New conversation.
Smart & Secure
We know security matters, so we built this with strict boundaries:
Permissions are King: The AI respects user roles. A user will never get an answer derived from a document they don't have permission to see.
Only the "Real" Stuff: To prevent confusion, the AI only learns from finalized content. It searches the most current Policy Releases and Approved Procedures. It completely ignores drafts, candidates, or anything currently under review.
For Owners & Admins
This integration is enabled by default to help your team immediately. However, if you need to turn it off for any reason, you have full control.
Go to Settings > LLM.
Click Deactivate.