Adds safety/unapproved-topics to settings

Introduces "safety/unapproved-topics" to enhance content management and filtering settings by addressing potentially unapproved topics.

docs/mint.json

@@ -159,6 +159,7 @@
         "safety/inspect-mode",
         "safety/debug-mode",
         "safety/pii-redaction",
+        "safety/unapproved-topics",
         "resources/assist-mode"
       ]
     },

docs/safety/unapproved-topics.mdx

@@ -0,0 +1,93 @@
+---
+title: "Unapproved Topics"
+description: "Define topics you don't want AI to handle"
+icon: "shield-exclamation"
+---
+
+## What's This?
+
+By default, our AI will attempt to handle all customer inquiries. However, you can define specific topics that should always be handled by human agents instead.
+
+<Note>
+The AI will handle everything unless you specifically tell it not to. Configure restricted topics in your dashboard under: **Settings → Autopilot and Allowed Topics**
+</Note>
+
+## How It Works
+
+<Steps>
+  <Step title="Default Behavior">
+    AI handles all topics unless specifically restricted
+  </Step>
+
+  <Step title="Smart Topic Detection">
+    Our AI intelligently understands context and intent - no need to set specific keywords. It recognizes when a conversation relates to your restricted topics.
+  </Step>
+
+  <Step title="Customer Notice">
+    Customer is informed they'll be speaking with a human agent
+  </Step>
+
+  <Step title="Context Preservation">
+    Full conversation history is provided to the human agent
+  </Step>
+</Steps>
+
+## Customization Options
+
+<AccordionGroup>
+  <Accordion title="Define Restricted Topics">
+    Add topics specific to your business:
+    - Industry-specific issues
+    - Company policies
+    - Special procedures
+    - VIP customers
+  </Accordion>
+
+  <Accordion title="Routing Rules">
+    Configure where conversations go:
+    - Specific departments
+    - Agent skill levels
+    - Priority queues
+    - Time zones
+  </Accordion>
+</AccordionGroup>
+
+## Best Practices
+
+<CardGroup cols={2}>
+  <Card title="Start Open" icon="door-open">
+    Let AI handle everything initially, then restrict topics as needed
+  </Card>
+
+  <Card title="Monitor First" icon="eye">
+    Watch AI performance before restricting topics
+  </Card>
+
+  <Card title="Regular Reviews" icon="clipboard-check">
+    Periodically review if restrictions are still needed
+  </Card>
+
+  <Card title="Team Input" icon="users">
+    Get feedback from agents on what to restrict
+  </Card>
+</CardGroup>
+
+## Configuration
+
+<Steps>
+  <Step title="Access Settings">
+    Go to Settings → Autopilot and Allowed Topics
+  </Step>
+
+  <Step title="Add Topics">
+    Simply list the topics you want to restrict - our AI understands natural language and context
+  </Step>
+
+  <Step title="Test Settings">
+    Verify handoffs work as expected
+  </Step>
+</Steps>
+
+<Card title="Need Help?" icon="question" color="#0091FF">
+  Contact our support team to help configure your topic restrictions and handoff rules.
+</Card>