Skip to main content

Labels

Data Labels is a new capability in the console. It helps you classify data across four dimensions: Data Type, Data Provenance, Data Pattern, and Data Sensitivity. You can also create custom label sets that are specific to your organization. All labeling features are in the classification sidebar in the console.

Labels give you a clear, consistent way to classify and organize your data. They work by combining automatic AI classification with your own custom labels. This makes it much easier to find the data you need, apply the right security and compliance policies, and create reports on your most important information.

Features and benefits​

  • Data Type: Automatically identifies file content (e.g., contracts, source code, financial records) using LLM-driven content classification to maximize accuracy and coverage with minimal tuning.
  • Data Provenance: AI-driven determination of data ownership (Internal, Personal, or Public). This helps teams separate proprietary data from noise, allowing them to target policies and investigations more effectively.
  • Data Pattern: Detects regulated and sensitive patterns (e.g., PII, PCI, PHI) using content inspection rules for precise and comprehensive compliance controls.
  • Data Sensitivity: Aggregates and rolls up other labels into a simple, ordered sensitivity scale, enabling policies and reports to focus on the highest-impact data first.
  • Custom Label Sets: Allows you to define organization-specific concepts once (e.g., proprietary project names) and reuse them consistently across both discovery and policy enforcement.

Use cases​

  • Speed up data discovery and improve reporting by automatically classifying data based on its type and ownership. Examples include Source Code, Contracts, Invoices, Internal, Personal, and Public
  • Strengthen controls by using pattern matches like PII and PCI and by using sensitivity levels to set policy severity.
  • Capture project or team concepts with custom label sets. These labels can be combined with other classification methods to create highly precise policy scope and segmentation.
  • Validate the behavior and effectiveness of new classification labels and policies through a test and tune phase before performing a wider, organization-wide rollout. This validation capability is initially available for Data Type classifications and will be expanded to support other label types in the future.