Skip to main content
PII Redaction

Learn about the ability to redact (mask) PII and other forms of sensitive data.

Mikhail Dubov avatar
Written by Mikhail Dubov
Updated over a week ago

Chattermill allows customers to redact Personally Identifiable Information (PII) and other forms of sensitive data. In this article we answer common questions related to this functionality.

Can Chattermill redact PII data?

Yes, Chattermill supports redaction of PII data in which case it is scrabbed from each piece of feedback or conversation (or transcript in the case of voice data). For security, this happens at the very beginning of the data ingestion pipeline and the operation is therefore not reversible. Once redacted the piece of sensitive information is replaced with a specific sequence of characters.

What formats of PII does Chattermill redact?

Chattermill has several prebuilt formats it supports:

  • Email

  • Phone Number

  • Post Code (UK Format)

  • Zipcode (US Format)

  • Credit Card Number

  • Social Security Number (US Format)

The rules can be enabled separately or together. They can be enabled for specific projects or for all projects, including any future projects.

Additionally, Chattermill supports custom rules supplied by customers. This is particularly useful for custom User IDs, Account IDs, or regional PII formats. For this, we typically utilize Regular Expressions. You can provide us with a regular expression pattern to look for or consult with our team to build a pattern if you are not sure. We utilize regular expressions for this purpose for full reliability.

How do I enable PII Redaction for my Chattermill account?

Please contact our Support team via email or the in-app chat to enable PII redaction for your account or a specific project. It would help us if you could provide the following with your request:

  • Projects you want to enable PII redaction for (eg. All projects)

  • Formats you want to redact (eg. All built-in formats)

Do you redact PII in comments only or metadata as well?

PII redaction described here is only relevant for redacting PII inadvertently included in the unstructured piece of feedback (or conversation transcript) itself rather than metadata. For metadata, the much more secure method is to avoid fetching data that contains PII data in the first place. This can be customized as part of the implementation/integration. Please consult our Customer Success team if you are unsure or would like to make a change.

How does PII redaction work?

At ingestion, as soon as there is a match the relevant sequence in the comment will be replaced with a specific pattern. For example:

My name is John and my email is [email protected]

could be replaced with

My name is John and my email is [redacted]

The email in this case will not be available to any of Chattermill's models, will not be stored in our database and will not be visible on the Chattermill application.

Is PII redaction part of my package?

Yes, PII Redaction is provided to all Chattermill customers as part of their package.

Did this answer your question?