There are three configuration options to anonymize information in the Torch platform and increase data protection: Feedback Invitee, Written Feedback, and User Account.


Feedback Invitee Anonymization

Torch can configure the customer subscription settings to anonymize the email address of invited feedback providers when their survey response is received. For example, the address "name@company.com" would be overwritten with a unique, untraceable value such as "1d34sxd345aoi@company.com",

Limitations

  • Users will not be able to easily re-invite these reviewers using the invitee suggestions feature, since previously used invitee email addresses will no longer be available.

  • The recurring feedback features of the learning goals feature will not be available, as it relies on stored email addresses of invitees. Clients in this case would be limited to the use of “private” Learning Goals, which do not have recurring feedback.


Written Feedback Anonymization

Subscriptions in the Torch application can be configured to limit the probability of PII (personally identifiable information) being used in qualitative client feedback responses. This is done with NLP (Natural Language Processing) technology that is configured to detect proper names, and instruct the feedback invitee to remove the offending names from their statements.

This Feature in Action

Limitations

  • NLP is not perfect, and may result in false positives or negatives depending on the text that is entered.

  • There may be cases in which acceptable usage of proper names in feedback will be blocked, such as the mention of a public figure.

  • For existing Torch subscribers, application of this configuration setting is not applied retroactively.


User Account Anonymization

User accounts can be configured to be anonymized (all PII removed) within the Torch platform after a specified number of days from account deactivation.

Limitations

  • Reactivation of anonymized accounts will require the client to repopulate their user data in the Torch application.

Did this answer your question?