By default, Dawn stores and processs any and all data you send. This is down according to our Data Sercurity and Privacy guideliness.

If you would like to prevent sensitive data from being stored, processed, or even sent to Dawn in the first place - we provide a few options.

​
Intelligent PII Redaction (Server Side)

You’ll need to request a member of the Dawn team to enable this for you.

When events are received, Dawn can intelligently strip PII from any user and model messages using a combination of pattern matching and AI models to detect and redact sensitive information. This is done at ingestion, meaning we only store the redacted version of the text and have no access to the original text.

Here are some example of the PII being redacted:

  • Email Addresses

  • Credit Card Numbers

  • Passwords

  • Phone Numbers

  • Street Addresses

  • US Social Security Numbers

​
SDK based redaction (Local)

You can also prevent common forms of PII from even being sent over the Internet in the first place. This is achieved by packaging a subset of our PII redaction code directly in ous SDKs (currently only supported in the Typescript SDK with plans to bring it to the other SDKs soon).

Since the this happens client side, only PII that can be matched using RegEx patterns can be redacted. To use intelligent redaction, we’d recommend enabling server side redaction instead.

Check out our SDK docs for instructions on enabling this.

Was this page helpful?