What does it mean to exclude text from the Privacy Filter?
While the Omnifact Privacy Filter™ (see the Privacy Filter documentation) automatically masks sensitive information to protect your data, there are cases where you want the AI to access certain details. For example, you might want the AI to process a particular ID, code, or name that would otherwise be filtered. Using the%% syntax allows you to exclude specific words or phrases from the Privacy Filter, ensuring they are sent to the AI model exactly as typed, without being masked.
Note: This feature acts as a proactive version of Click to Reveal. Instead of revealing information after it has been masked, you prevent it from being masked in the first place.
How to Exclude Text from the Privacy Filter
To exclude a word or phrase from the Privacy Filter, simply wrap it in double percentage signs (%%) when typing your message.
Syntax: %%your text%%
Example
If you want to ensure the name “Allie” is not masked by the Privacy Filter, you would type:My name is %%Allie%% and I am a UX designer.When you send this message:
- The Privacy Filter will ignore the text inside the
%%signs. - The name “Allie” will be sent to the AI model in clear text.
- The AI will be able to use the name “Allie” in its response.
- In the chat interface, the text “Allie” will appear with an open eye icon, looking exactly as if you had manually revealed it after sending.
How to Enable Excluding Text from the Privacy Filter
This is an advanced feature and is disabled by default. It is managed via an organization-level feature flag. To enable excluding text from the Privacy Filter for your organization:- Have your Team Administrator contact Omnifact Support.
- Request to enable excluding text from the Privacy Filter for your organization.
%% syntax to bypass the Privacy Filter.