Use sensitive data redaction - Amazon Connect

Use sensitive data redaction

To help you protect your customer's privacy, Contact Lens lets you automatically redact sensitive data from conversation transcripts and audio files. It redacts sensitive data, such as name, address, and credit card information using Natural Language Understanding.

To enable redaction, choose the option on the Set recording and analytics behavior block. For more information, see Enable redaction of sensitive data.

Sensitive data redaction is applied after a call disconnects.


The redaction feature is designed to identify and remove sensitive data. However, due to the predictive nature of machine learning, it may not identify and remove all instances of sensitive data in a transcript generated by Contact Lens. We recommend you review any redacted output to ensure it meets your needs.

The redaction feature does not meet the requirements for de-identification under medical privacy laws like the U.S. Health Insurance Portability and Accountability Act of 1996 (HIPAA), so we recommend you continue to treat it as protected health information after redaction.

For a list of the languages supported by Contact Lens redaction, see Languages supported by Amazon Connect.

About redacted files

Redacted voice files are stored in your Voice Amazon S3 bucket, for example: connect-instanceARN/Analysis.

Redacted chat files are stored in your chat Amazon S3 bucket, for example: connect-instanceARN/Chat

You can access all files (redacted, unredacted, raw, etc.) through the AWS console, by using the Amazon S3 console.

Following is a list of what you can access by using the Amazon Connect admin website (such as on the Contact details page), assuming you have the appropriate security profile permissions:

  • Access redacted voice and chat files.

  • Download redacted voice recordings.


Currently, you cannot download redacted chat files and voice transcripts.

When redaction is enabled, Contact Lens generates the following files:

  • A redacted file. This file is generated by default when Redaction is enabled. It's the output schema, with sensitive data redacted. For an example file, see Example redacted file for a call.

  • An original (raw), analyzed file. This file is generated only when you choose Get redacted and original transcripts with redacted audio in the Set recording and analytics behavior block. For an example file, see Example original, analyzed file for a call.


    For voice contacts, the original analyzed file is the only place where the complete conversation is stored. If you delete it, there will be no record of the sensitive data that was redacted.

  • A redacted audio file (wav) for voice contacts. Sensitive data in audio files is redacted as silence. These silent times are not flagged in the Amazon Connect admin website or elsewhere as non-talk time.

Use your file retention policies to determine how long to keep these files.