Data redaction
Redaction removes sensitive information from stored conversations. When a redaction rule matches — payment card numbers, government identifiers, or any other pattern your organization defines — the matched words are masked in the transcript and the corresponding audio is replaced with silence. Both the transcript and the audio file are modified, so the sensitive content is not retained in either form.
Organizations typically use redaction to keep payment data and personal identifiers out of stored conversations for compliance reasons (for example, PCI DSS).
Where it is configured
Redaction rules live under Administration › Conversation Analytics › Data Redaction (tab Rules), with Add, Delete, Export, and Import actions. Each rule defines the patterns to detect and whether matches are redacted in the audio, the transcript, or both.
Rule configuration — pattern syntax, testing rules before applying them, and the review workflow — is covered in the dedicated Data Redaction User Guide; this guide does not repeat it.
Managed by your service provider
The redaction processing itself runs automatically as part of the conversation-processing pipeline managed by your service provider. If your subscription includes data redaction but the menu item is missing, or you need redaction applied to historical conversations, contact your provider.
What users see
Redacted words appear as masked text (for example *********) in transcripts, and playback is silent
over the redacted intervals. See Transcripts in the
User Guide.
Related pages
- Data Redaction User Guide — rule configuration and workflow.
- Conversation Analytics administration — what is configurable vs provider-managed.