Removing personal data from a document

You need this when a document has to go out but the people’s details must not: a contract shown as an example, a case sent to a consultant, a document for a tender.

How to do it

To remove the personal data from a document:

1. Right-click on the document in the list — or use the three dots button at the end of the row, which appears when you hover over it.
2. Hover over «AI»: it has an arrow on the right and opens a submenu.
3. Choose «Anonymise document».
4. In the window that appears in the middle of the screen, choose how to mask: «Labels (default)» replaces each detail with a code, and the document still reads as a structure; «Redaction (black bar)» simply covers them. Then press «Start».
5. Wait: the work goes on by itself (text extraction, then finding the data). When the copy is ready the «Anonymisation review» window opens on its own — that is where you look at it and correct it.
6. In that window, press «Download anonymised» to take the copy away.

The mask badge on the file’s row reopens the review whenever you need to go back to it.

What comes out, exactly

⚠️ The original is not touched. What comes out is a separate copy, named like the original plus (anonymised) — so you must send that one, not the file you started from. The copy does not appear as a row in the list: you reach it from the review.

⚠️ Only Word files stay Word. An anonymised .docx is still a .docx; everything else — PDFs included — comes out as a text file (.md). All the content is there, the layout is not. If the recipient expects a PDF identical to the original, this is not the right tool.

Review it before trusting it

The review is not a formality: automatic recognition can miss a name written in an odd way, or mask something you needed.

In the «Anonymisation review» window you go through the occurrences one by one and correct them: you can take an entry out of the masking with «Remove (reveal the real value)», or select text that recognition skipped and mask it by hand. Your changes only reach the copy when you press «Save».

If the document is a scan, the window tells you: those are the pages read with optical recognition, where mistakes are easier. Always look, especially if the document is leaving the company.

Who sees the real data

The mapping between the codes and the real values holds the sensitive data: only someone with the «Anonymizer» permission on that folder can see it. It is not a technical detail — it is why that permission is given to few people.

If it doesn’t work

If «Anonymisation failed» appears, the reason is written there when it is known. The two typical cases: a document made of scanned images with no readable text, or a format the system cannot read.

If the entry isn’t there

«Anonymise document» is missing from the «AI» menu in three cases, in this order:

• the «Anonymizer» extension is not active for the company → whoever administers Kamzan turns it on;
• you don’t have the «Anonymizer» permission on the folder → whoever manages access to that folder grants it;
• the document is of a type no text can be extracted from.

Questions this guide answers

• how do I black out personal data
• I have to send a document without the names
• how do I anonymise a file
• what is the anonymiser
• anonymise document
• black out the names in a pdf
• anonymisation review
• download anonymised
• mask sensitive data