Your AI gets access. Not a master key.
Define AI applications, audiences, policies, and access rules so each assistant uses the right collections and documents.
The right knowledge for the right AI.
Define what each chatbot, copilot, and assistant may read, cite, or never touch — then enforce those boundaries on every single request.
- Per-application policies
- Visual access matrix
- Public-safe controls
- Enforced on every request
Zero-trust access for every AI application.
What permissions covers in Wiki.
AI applications
Model each chatbot, copilot, docs assistant, or internal agent separately.
Audiences
Map customer, employee, partner, or admin contexts to policy behavior.
Policies
Set allowed, restricted, citation-required, and unavailable rules.
Access matrix
See how applications interact with collections at a glance.
Document-level deny
Block sensitive notes even when a collection is otherwise allowed.
Public-safe controls
Keep internal knowledge out of customer-facing AI.
Access simulator
Test what an application can use before you ship changes.
Explore the rest of the product.
Knowledge
Turn scattered information into managed AI knowledge.
Verification
Unverified knowledge never reaches your AI.
Answers
Test the answer before it reaches the customer.
AI Permissions: What Your Chatbot Should Never Be Allowed to Know
The biggest AI knowledge mistake is giving every chatbot access to every document. Permissions should be explicit, teste...
RAG Is Not a Source of Truth
Retrieval can find context. It does not decide whether that context is current, approved, allowed for the audience, or s...
Your AI is already talking.
Make sure it knows what it is talking about.