You are here:
Employee Confidential Reporting Subagent
Handle the secure detection, classification, and reporting of confidential concerns such as harassment, discrimination, and misconduct.
Required Editions
| View supported editions. |
| User Permissions Needed | |
|---|---|
| To enable agents in Salesforce: | Manage AI Agents AND Employee Service: Agentforce OR Customize Application |
| To build and manage HR service agents: | Manage Agentforce Service Agents AND Manage AI Agents OR Customize Application |
Topic Details
This topic contains potentially sensitive actions. To use this topic, customize this action according to your business’s security requirements. See Maintain Trust with Agentforce Actions.
| API Name | inappropriate_content |
| Included Agent Tools | Getting Items in Picklist Creating of Confidential Case |
Examples of Utterances Classified to This Subagent
- "I need to report someone for harassment."
- "I am being discriminated against and want to file a complaint."
- "I want to report a workplace safety concern privately."
- "Someone is being toxic and offensive in my department."
- "I need to speak to HR about a confidential misconduct issue."
Considerations
- To handle organization-specific keywords or unique reporting scenarios, override the default
topic with custom instructions so the agent understands workplace nuances and routes
confidential reports that standard guardrails might otherwise block.
- If you override or recreate this topic, set the API Name to inappropriate_content. This identifier is required to bypass standard AI safety blocks.
- In the Scope field, paste the categories to capture (for example, violence, harassment, bias, or offensive language) so the agent can identify the correct triggers.
- Configure the instructions to initiate the confidential case flow when the topic is triggered.
- Manually add the Getting Items in Picklist action to categorize the issue and the Create Confidential Case action to route the report to authorized specialists.
- To route confidential conversations from the agent to a human specialist, override your standard escalation flow or merge the logic from the Escalate Confidential Complaint flow into your existing configuration.
Did this article solve your issue?
Let us know so we can improve!

