Thirty transcripts, one participant under three names
You have thirty interview and focus-group transcripts in Word, and you want AI to propose a first set of codes. In field notes the same person is "Marko", "Mr. Kovač" and "the former mayor". People name their colleagues, their village and their doctor, and they say things they'd only tell a researcher.
Choose the placeholder style first
Analysis is easier to read with [PERSON_1] and [PERSON_2] than with random codes. Choose Numbers in order in Settings before the first ID, since the style is locked for the set after that. Numbering does show the order in which people appeared, so if that order could say something, keep Random.
[PERSON_1]: After the mine closed, [PERSON_4] at [PLACE_2] stopped speaking to us.
One person, one placeholder, thirty files
Current IDs carries across the folder, so a participant keeps one placeholder in every transcript. Use Group as one for "Marko", "Mr. Kovač" and his nickname. Add "the former mayor" yourself: it isn't a name, but it identifies one person.
Read the transcripts closely for dialect, quotes and local events. In a small community those identify someone as surely as a name does. Text in Word files is what gets read, so photos in field notes stay as they are. Check them yourself.
The ID file is what links people back
The exported ID file is what turns [PERSON_1] back into a person. Store it apart from the protected transcripts, under the same rule as your consent forms and code list, and never upload it. Clearing Current IDs or closing the page leaves you only that file.
When the coding comes back as a workbook, Reveal files can bring the names back with the ID file, and it writes the copy to a folder you choose. Often you won't need to: coded excerpts with placeholders may be what you want to keep.
What the project still has to decide
Pseudonymized transcripts are still personal data. Your consent forms, your ethics approval and your data management plan decide whether an outside AI service may process them. Protect also sends text to Anonymizator for detection, so include that step when you describe processing.
This is an illustration of how the product could be used. It isn't a customer's story.