Classify sensitive entities in any document. Five protection categories.

Five protection categories with legal basis

Shield's Document Classification Mode maps every detected entity to one of five categories: PII Standard (DSG/FADP), Sensitive PII (DSG Art. 5 lit. c), Client ID (Art. 47 BankG / Art. 69 FINIG), Own Secrets (Art. 162 StGB), and Third-Party Secrets (Contractual / NDA; Art. 162/273 StGB; UWG). Each category carries its legal basis for defensible processing.

• PII Standard: names, addresses, dates of birth, phone numbers
• Sensitive PII: health records, religion, biometrics, AHV numbers
• Client ID: banking identifiers, IBAN, account numbers
• Own Secrets: financials, pricing, IP, strategy
• Third-Party Secrets: NDA-covered data, counterparty information

Structured placeholders or full redaction

Two output modes for different use cases. Structured placeholders ([PERSON_1], [COMPANY_2]) preserve readability and cross-references for internal review. Full redaction replaces entities with solid blocks for external publication. Both modes maintain document formatting.

Document flags for workflow integration

Tag documents with processing directives: Legal Hold (must not be altered), Art. 271 StGB (must not leave Switzerland), or Not Relevant (excluded from proceedings). Flags carry through your entire document pipeline as metadata.

Zero persistence architecture

Classification runs entirely in memory. No database. No disk writes. No logs. When processing completes, the document data ceases to exist. Deploy air-gapped, on-premises, or on managed Swiss cloud infrastructure.