The Entities & Redaction feature is designed to give you control over what types of sensitive and personal data you store and view inside EvaluAgent.
An Entity is a specific type of sensitive, private or confidential information. Examples include customer name, address, postcode, account reference. Whenever you configure an integration to full-fetch mode you’ll be able to use this feature to firstly automatically identify specific entities and, secondly redact entities that you do not wish to store or view within EvaluAgent.
Each time an imported contact reaches our platform, your pre-configured entity & redaction rules are applied to identify and remove specific sensitive information.
Rules can be configured, tested, activated and deactivated by any user with the Manage Integration permission.
- (1) The platform comes seeded with a number of common entities
- (2) The platform also includes a wide range of pre-defined geographically appropriate entities that you may also wish to identify and redact.
- (3) To test how the platform finds and processes an entity, simply tick the test box alongside the entity or entities you wish to test, add some relevant text that includes examples of entities you are wishing to test into the testing panel, and click Run test.
Once you are comfortable that the entity rule works, click to activate the entities you wish to find and, if you would like the platform to redact an entity, simply switch redaction on.
Your rules are applied to all contacts imported after that point.
- Once text has been redacted, it cannot be re-instated - as the rules imply, redaction is used to scrub / delete selected data from our records.
- Redaction can only be applied to future imported contacts, not to contacts already imported.
- As soon as redaction is switched off for a specific entity, that entity on all future imported contacts will no longer be redacted, but for the reason highlighted above, previously redacted examples of said entity cannot be reinstated.
When entity and redaction rules have been applied to a conversation, the text looks something like this
Build your own customised Entities
You can build your own customised entity by clicking the + Create a customised entity button to reveal the slide in panel
Entity types that can be added include the following
A single string prefixed with
Matching product or customer numbers that all begin with a specific code eg GB432345
A single string suffixed with
Finding any reference to a particular word stem. For example, ‘complain’ would also return ‘complaint’, ‘complaints’ or ‘complaining’
A single string containing
Finding a string pattern at any point in a word. For example, searching for ‘author’ in an insurance claims call would also return ‘unauthorised’, ‘preauthorised’, ‘authorisation’ and so on.
A single string that exactly matches
Matching a word or phrase reference. For example, looking for specific organisations such as ‘Financial Ombudsman Service’ or purchase terms and conditions such as ’money back guarantee’ being mentioned.
Multiple strings matching
Matching one of many words such a customer mentioning competitor names or specific products.
A RegEx pattern that matches
Complex searches for custom strings that cannot be matched using the simple options above. We recommend using specialist resources online such as https://regexlib.com/ or https://regex101.com/ to learn more about building successful RegEx patterns