Anonymous Mode

Anonymous mode protects personal data in your conversations with the AI. When enabled, names, addresses, phone numbers, and other sensitive information are automatically hidden before being sent to the language model.

When Should You Use Anonymous Mode?

Enable anonymous mode when your request contains personal data – for example:

• You’re discussing a specific citizen case with names and addresses
• You’re having a response letter created that contains personal data
• You’re analyzing documents with personal information

Practical Example

You want to have a response to a citizen complaint drafted. The text contains the citizen’s name, address, and file number. With anonymous mode enabled, this data is automatically replaced with placeholders before the AI sees the text.

How to Enable Anonymous Mode

1. Start a new chat
2. Enable the Anonymous mode option (toggle in the input area)
3. The mode applies to all messages in this chat

Note

Anonymous mode can only be enabled or disabled when creating a new chat. Once the chat has started, the chosen setting remains in place.

What Is Protected?

Data Type	Example	In Anonymous Mode
Names	John Doe	[NAME_1]
Addresses	Main Street 12, 12345 Berlin	[ADDRESS_1]
Phone Numbers	030 12345678	[PHONE_1]
Email Addresses	john@example.com	[EMAIL_1]
Other Personal Data	Date of birth, file number	Automatically detected and replaced

How Responses Display Protected Data

When anonymous mode is active, the AI only sees placeholders like [NAME_1] instead of real personal data. In the AI’s response, however, the original values are automatically restored and highlighted so you can read the text naturally.

Restored personal data appears with a colored background. If you hover over a highlighted value, a tooltip shows the category (e.g., “Person name” or “Email address”) and confirms that this data was anonymized before being sent to the AI.

This means:

• The AI never sees the real data — it only works with placeholders
• You see the original values in the response, clearly highlighted
• You can always tell which parts of a response contain personal data

Anonymization Exceptions

Your administrator can configure exceptions for specific categories of personal data. When a category is exempt, data of that type is not masked and is sent to the AI as-is.

For example, your administrator might exempt organization names so the AI can reference your department or agency by name, or exempt email addresses matching your organization’s domain so internal email addresses are not hidden.

You can tell which categories are exempt because data of that type will appear in the AI’s response without highlighting. If you’re unsure which exceptions apply, ask your administrator.

Note

Anonymization exceptions are configured by administrators under Admin Settings → Anonymization. See Admin Settings: Anonymization for details.

Important to Know

Anonymous mode does not replace your own diligence

Anonymous mode is an additional protective measure that supports you when working with personal data. It detects many common patterns such as names, addresses, or phone numbers – but not always all sensitive information. Unusual spellings, technical terms, or context-dependent data may not be recognized in every case.

Always review your inputs for sensitive content yourself before sending them. Anonymous mode replaces neither your own duty of care nor your organization’s data protection policies.

Next Steps

• Create and manage chats – Basics of using chat
• Documents, Images, Knowledge Collections, and Integrations – Use files and knowledge collections in chat
• Admin Settings: Anonymization – Configure anonymization exceptions (administrators only)