> ## Documentation Index
> Fetch the complete documentation index at: https://docs.hiperai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Data Protection (DLP & PII)

# Data Protection

Data Protection prevents accidental or malicious leakage of confidential business information and personally identifiable information (PII) during AI interactions.

***

## Core Protection Capabilities

### 1. DLP Management (Business Data Loss Prevention)

Protects proprietary business assets in prompts, uploaded files, and model responses:

* **API Keys & Credentials**: Prevents AWS keys, OpenAI tokens, passwords, and private certificates from leaking.
* **Source Code & Architecture**: Detects internal repositories and code snippets.
* **Contracts & Financials**: Identifies confidential agreement drafts and financial data.

### 2. PII Redaction (Privacy & Compliance)

Automatically detects and masks personal identifiers to satisfy GDPR, HIPAA, and CCPA standards:

* **Identifiers**: Names, email addresses, phone numbers, national IDs (SSN, DNI, Passport numbers), credit cards.
* **Redaction Modes**:
  * **Mask**: Obfuscates sensitive strings (e.g., user@\*\*\*\*\*.com, ****-****-\*\*\*\*-1234).
  * **Remove**: Strips the sensitive text entirely.
  * **Tokenize / Hash**: Replaces sensitive data with deterministic pseudonyms for internal processing.

### 3. Semantic Verifier

Assists in security and legal investigations by determining if a suspicious piece of text was generated by your SecureAI tenant:

* **Confidence Levels**: *Verified*, *Probable Match*, *Ambiguous*, *No Match*.
* **Usage**: Go to **Admin ? Data Protection ? Semantic Verifier** and paste the text snippet to analyze origin probability.

***

## Configuration & Setup

1. **Global Settings**: Go to **Admin ? Data Protection** to enable real-time scanning for inputs, outputs, and file attachments.
2. **Templates & Custom Rules**: Load regional/industry starter templates or add customized regex patterns.
3. **Simulation Mode**: Test policies against sample datasets before enforcing them in production.
4. **Incident Monitoring**: Review redaction and blocking metrics under the Data Protection incident logs.