Docult relies on an in-house AI engine developed for this offer: specialised detection, business rules, reversible pseudonymization and audit, without sending documents to a third-party cloud model.
The Docult engine combines specialised detection, legal rules and human control. It identifies sensitive data, proposes stable codes and prepares exports without attempting to recreate original values.
The Docult engine reads PDF, Word and Excel files, preserves document structure and prepares areas for analysis.
Business rules, regex patterns and specialised detection identify people, entities, addresses, dates, IBANs and confidential information.
Each validated value receives a stable code, linked to an encrypted reversibility vault and an actionable audit report.
The Docult engine is built for one purpose: detect sensitive data, replace it with traceable codes and produce audit evidence useful to legal teams.
Detection and pseudonymization logic is designed by Docult for legal documents, without relying on an external general-purpose model.
Core processing runs on Docult infrastructure hosted in France. Documents are not sent to a third-party cloud model.
Users validate entities before application, adjust suggestions and keep control of the mapping table.
Codes stay stable and can only be restored from the encrypted vault by authorized users.
Each detected category can be validated, corrected, redacted, exported and restored according to firm permissions.