Who can use this feature?
Plan: Call Center AI
Managed from: Admin Portal
User type: Admin, Analyst
Compliance monitoring automatically detects potential rule violations across all calls where both Krisp Microphone and Krisp Speaker were active. When a possible breach is identified, it is flagged in the Admin Portal with a reference to the exact moment in the conversation where the issue occurred, so supervisors can review and act without listening to the full recording.
Default compliance rules
The following six rules are active by default when Krisp AI Speech Analytics is enabled. Admins can disable any rule that does not apply to their operation.
Flags cases where the agent processed or accessed personal information before completing proper customer verification. Verification requires the customer to confirm two pieces of identifying information, such as full name, date of birth, account number, address, or phone number.
Flags cases where the agent or customer verbally disclosed sensitive payment card data, including a full card number (all 16 digits), CVV/security code, or PIN. Sharing only the last four digits of a card number for verification purposes is not flagged.
Flags cases where the agent requested credentials that should never be collected, such as a full social security number, complete account password, full card number, CVV, or PIN. Asking for partial information (for example, the last four digits of an SSN) is not flagged.
Flags cases where the agent processed a financial transaction without receiving explicit customer consent. Treating silence or ambiguous responses as agreement, or using implied consent language without waiting for confirmation, are considered violations.
Flags cases where the agent shared protected customer information with a third-party caller (such as a family member or unverified legal representative) without proper authorization from the account holder. Sharing information with internal departments or partners as part of standard procedures is not flagged.
Flags cases where the agent contacted a third party about a customer's protected health information (PHI) without first obtaining the customer's explicit consent, including permission to contact the third party, disclosure of what information will be shared, identification of who will be contacted, and clear verbal approval.
How violations appear in the Admin Portal
When a potential violation is detected, it is displayed in the call details page in the Admin Portal. Each alert includes:
- The name of the rule that was triggered
- A justification explaining why the violation was detected
- The specific part of the call transcript where the issue occurred
If a recording is available for the call, Admins and Analysts can play it to confirm the context of the alert.
Reviewing and dismissing alerts
Admins can mark a detected violation as Not an Issue if, after review, they determine the alert does not represent an actual breach. A brief explanation can be added when dismissing.
Hint
The two identity verification rules (Processed Personal Information Without Speaker Verification and Disclosed Information to Unverified Third Parties) can occasionally produce false positives in ambiguous scenarios. Use the Not an Issue option to dismiss alerts that do not represent actual violations after review.
Managing compliance rules
Admins can enable or disable individual rules from the Admin Portal. Analysts can view compliance alerts but cannot change rule settings. Rules are organized by category: Identity Verification, Sensitive Data Handling, Payment Security, and Customer Authorization. You can use the category filters at the top of the page to narrow the list. To manage compliance rules:
- Go to Products >>> Speech Analytics >>> Manage rules
- Toggle individual rules on or off as needed.
Hint
If your team requires a compliance rule that is not in the default set:
- If Speech Analytics is not yet enabled for your team: Contact your Account Executive or reach out to sales@krisp.ai.
- If your team is already using Speech Analytics: Contact the Customer Engineering team at support@krisp.ai. All custom rule requests are reviewed and evaluated before implementation.