Skip to main content

Shield Settings

The Shield Settings section allows you to configure the basic operation and behavior of your Generative Shield.

Shield Basic Settings

Configuration Options​

Status​

Enable or disable shield enforcement:

  • Active: The shield is enabled and actively filtering content
  • Inactive: The shield is disabled and content passes through unfiltered

Filter Action​

Define what happens when the shield detects problematic content:

  • Block: Block unsafe content and prevent it from proceeding
  • Detect: Flag risky content while allowing conversations to continue uninterrupted
  • Rewrite (Coming Soon): Rewrite unsafe parts of the content

Filter Direction​

Control which parts of the conversation are filtered:

  • Input Only: Filter user messages and prompts before they reach the AI
  • Output Only: Filter AI responses before they're delivered to users
  • Both: Filter both user inputs and AI outputs (recommended for maximum protection)

Shield ID​

Each shield configuration has a unique identifier (shown in the interface) that can be used for API integration, monitoring, and troubleshooting purposes.

Shield Priority and Inheritance​

The organization-level shield applies baseline protection to all agents and conversations within your organization. When a shield configuration is set for a specific Organization Agent, the agent-level shield takes complete precedence and the organization-level shield is not applied to that agent. The organization-level shield remains active for all other agents and conversations.

Common Issues​

Troubleshooting

Unexpected Blocking: Check Filter Action settings and review Filter Direction scope. Switch to "Detect" temporarily to monitor flagged content without interrupting conversations.

Shield Not Working: Verify Status is "Active", confirm changes were saved, and check that filters are enabled.

Next Steps​

After configuring basic shield settings:

  1. Set up Standard Filters for common content categories
  2. Create Custom Filters for organization-specific needs
  3. Enable Cybersecurity Protection for security threats
  4. Test your configuration using the Management and Testing tools