The Nonymize field guide
Practical guidance for safer AI workflows
Clear, practical guidance for preparing sensitive conversations and documents before AI—without losing the context that makes them useful.
Privacy · AI workflows · Client trustLatest resources
View all →- 01
General · Aug 19
What are Structured Outputs? And why they are just a gate on reliability and not privacy?
- 02
General · Aug 13
When it comes to AI tool usage, the “Not Used for Training” data promise does not mean you're out of the woods
- 03
General · Aug 11
The Biggest Companies in Tech Just Validated the Architecture We Built Nonymize On
General
Anonymization Is More Than Removing a Person's Name
Removing names from a client transcript isn't anonymization — indirect identifiers like employer, job title, city, and deal details can re-identify someone with no name in the document at all. Research from Carnegie Mellon found 87% of Americans are uniquely identifiable from just ZIP code, birth date, and gender, and a peer-reviewed 2019 study put re-identification at 99.98% with 15 attributes, even in incomplete datasets. Real anonymization means catching the full combination of identifiers, keeping replacements consistent across documents, and putting a human review step in front of anything that reaches an AI tool.
Read the guide →Put the guidance to work
Prepare your next document before it reaches AI.
Nonymize suggests replacements for identifying details. You review and approve the result. We do not use your text to train AI models.