This feature is in a Limited Availability phase and may be enabled on request by contacting your Druva account representative or Druva Support. When enabled, a Trial icon appears. Governed by the Druva Early Access Addendum, this feature may be modified, discontinued, or offered as a separately priced offering at any time.
Overview
The Chat Violations feature in Sensitive Data Governance (Compliance) allows administrators to monitor and manage instances where sensitive company information is shared with the Claude AI assistant. By tracking these interactions, you can proactively identify compliance risks and take action to protect your organization's data.
Supported Claude data for Chat violations
Chat Violations: Includes session files generated from Claude Code and Cowork.
File Violations: Encompasses configuration files, skill files, and other related assets. For more information, see What data is backed up for Claude?
Accessing Chat Violations for Claude backed up data
To view your organization's AI chat compliance violations:
Log in to the Druva Management Console.
From the Global Navigation Panel on the Cloud Platform console > Data Governance, select Sensitive Data Governance.
On the left-hand menu, click Chat Violations.
Viewing Active Chat Violations for Claude backed up data
The Active tab displays a list of all unresolved sensitive data violations detected in Claude interactions. You can review the following details for each incident:
Conversation: A snippet of the chat message that triggered the violation. Click the information icon (ⓘ) next to the snippet to view the full context.
Last Interaction Time: The exact date and time of latest interaction.
User Name: The user who interacted with the AI agent.
App Name / Agent: Identifies the specific AI application (Claude) and the agent environment (For example: Code).
Sensitive Data Matched: The category of sensitive data identified in the chat (For example: Prescription Details, Bank account details).
Policy Violated: The specific compliance policy that flagged the interaction.
Violation Status: The current state of the violation (For example: Pending on admin).
Investigating Chat Violation Details for Claude backed up data
To get more context on a specific violation:
Click on the conversation snippet in the Active tab.
This opens the Violation Details view, which displays device information (Device Name, OS, and File Path).
Scroll down to the Violations section to see exactly what triggered the policy. The Violation Preview highlights the sensitive text (like confidential document markers or driver's license numbers) directly within the chat transcript so you can accurately assess the risk.
Resolving a Chat Violation for Claude backed up data
Once you have reviewed a violation and taken appropriate administrative action, you can mark it as resolved:
On the Chat Violations > Active tab, check the box next to the conversation(s) you want to resolve.
Click the Resolve button at the top of the list.
The violation will be moved to the Resolved tab.
📝Note: You can download session file that flags the chat violations.
Managing Resolved Violations for Claude backed up data
The Resolved tab keeps a historical record of all processed chat violations.
Reviewing Details: You can see when the violation was resolved (Resolved On) and how it was handled (Resolution Details, such as "Auto Resolved").
Un-Resolving: If a violation was resolved by mistake, select the checkbox next to it and click the Un-Resolve button to move it back to the Active tab for further investigation.