Add response scans to a guardrail

Overview

Use this task to add response scans when you create a guardrail. By default, F5 AI Guardrails scans prompts. AI Guardrails can also scan responses.

AI Guardrails includes two guardrail groups:

• F5 guardrails: Ready-to-use packages from F5, such as PII, Restricted topics, Prompt injection, and EU AI Act packages.
• Custom guardrails: Guardrails you create for your own policies.

You can set AI Guardrails to scan responses for both categories.

INFO

Efficacy metrics apply to prompts, not responses. If you set a guardrail to Responses or Prompts and responses, this can increase latency and token costs.

Before you begin

Before you begin, make sure you have:

• Admin privileges: You need permissions to create or update guardrails.
• Access to the AI Security UI: If you use SaaS, the URL is typically https://us1.calypsoai.app.
• One or more F5 guardrails turned on: Turn on the F5 guardrails you want to scan responses.

Set scan direction for an F5 guardrail

1. Log in to the AI Security UI.
2. Select Guardrails.
3. Select the F5 guardrail you want to change.
4. For a guardrail you've turned on, select the Scan direction list, then select one option:
   • Prompts
   • Responses
   • Prompts and responses

AI Guardrails saves your changes automatically.

References

For more information, see:

• Prompts and scans in AI Security