---
sourceDocument: Yokohama Enable AI
sourceDocumentLink: https://servicenow-prod.fluidtopics.net/r/yokohama/intelligent-experiences

 Release :

    - yokohama

ft:locale :

    - en-US

ft:publication_title :

    - Yokohama Enable AI

ft:clusterId :

    - platai

bundleId :

    - platai

workflow :

    - Platform


---

# Enable AI Guardian in AI agents

# Enable AI Guardian in AI agents {#ariaid-title1}

Release version: Yokohama  
Updated September 10, 2026  
![](https://www.servicenow.com/docs/portal-asset/ico-clock) 1 minute to read  
Identify and block offensive messages that are sent by human agents automatically by enabling AI Guardian in AI agents. With this capability, you can help reduce your agentic workflow or test from being exposed to harmful content.

## Before you begin

Role required: admin

## About this task

The AI Guardian, which is a ServiceNow AI Platform capability in the Now Assist panel, is a set of guardrails that are designed to intercept and mitigate offensive, sensitive, or security-related issues that may arise during interactions with the Now Assist application.

For example, let's say that AI Guardian detects an offensive message in the execution plan of an agentic workflow. When you try to trigger the plan or test it, AI Guardian can step in to terminate the plan or test because it detected harmful content at the first step of the execution plan.

For more information about the different guardrails, see [Now Assist Guardian](https://servicenow-prod.fluidtopics.net/H0bVbM_NdghPjzLAF~OnJg#now-assist-guardian "AI Guardian is built on the ServiceNow Small Language Model (SLM) and monitors generative AI interactions to detect offensive content, prompt injection attacks, and sensitive topics.").

## Procedure

1. Configure Offensiveness for AI agents.
   1. In AI Agent Studio, navigate to the Settings tab.  
      You're directed to the Offensiveness page.

   2. Turn on the Offensiveness setting for AI agents by using the toggle button.  
   3. Configure the detection impact by selecting the options icon (![More options icon.]()) to enable the detection impact to use the following options:  
      * Edit: Choose the detection impact between logging or both blocking and logging.  
        Enable the detection impact:
        1. Select the Edit option.
        2. On the Detection impact page, select either the Log or Block and log button based on your requirements.  
           Note:  
           You can switch the detection impact from Log to Block and log or from Block and Log to Log at any time.
        3. Select Save.
        {#enable-aia-na-guardian__ol_g2w_rhb_w2c}
      * Export: Exports the offensiveness-based logs as a .CSV file. Logs can be analyzed for the types of content that are being identified so you can take other preventive measures, such as changing conversational questions.
      {#enable-aia-na-guardian__ul_cnk_4vd_zdc}
   {#enable-aia-na-guardian__substeps_ohm_11l_xdc} {#enable-aia-na-guardian__offensiveness-detection}
{#enable-aia-na-guardian__offensiveness-detection}
2. Configure Prompt Injection for AI agents.
   1. In the AI Agent Studio, navigate to SettingsPrompt Injection and select Configure.  
      You're directed to the AI Admin Hub page to configure the Prompt Injection.

      Note:  
      For more information about configuring the Prompt Injection, see [Configure prompt injection attack protection](https://servicenow-prod.fluidtopics.net/zWeDIFV9sl_NEik2nxGEKw "Activate or deactivate prompt injection attack detection to protect all generative AI applications and AI-generated text and conversations.").

      When you configure the Prompt Injection for an agentic workflow by using the required instructions, the system is designed to detect
      harmful content and block the conversation.
   {#enable-aia-na-guardian__substeps_dmp_21l_xdc} {#enable-aia-na-guardian__prompt-injection-detection}
{#enable-aia-na-guardian__prompt-injection-detection}{#enable-aia-na-guardian__steps_yfg_k2p_vdc}

*[\>]: and then