Guard
Intelligent governance that protects sensitive data, ensures compliance, and builds trust at scale.
At a glance
Automated discovery
Automated discovery of PII, PHI, and financial data.
Real-time protection
Real-time masking, tokenization, and access control.
Prebuilt policies
Prebuilt policy packs for GDPR, CCPA, PIPEDA, PIPA, SOC 2, HIPAA.
Continuous monitoring
Continuous monitoring and audit-ready reporting.
Native integration
Natively integrated with Fabric, Studio, and Signals.
Why Guard
Trust is the foundation of enterprise AI. Data without governance is noise, and AI without compliance is risk. Guard establishes control, transparency, and resilience so organizations can innovate without fear.
What Guard does
Discover
AI-powered scanners classify structured and unstructured data with intelligent context detection.
- •AI-powered scanners classify structured and unstructured data.
- •Sensitive field detection across text, logs, contracts, and telemetry.
- •Entity recognition for names, addresses, IDs, credit cards, health codes.
Protect
Real-time data protection with advanced masking, tokenization, and access controls.
- •Real-time masking, redaction, and tokenization.
- •Role-based access down to the column and cell level.
- •Key management with rotation and customer-managed keys.
Govern
Policy enforcement aligned to global frameworks with approval workflows and lineage tracking.
- •Policy packs aligned to global frameworks.
- •Approval workflows for data usage and model training.
- •Lineage and change history for every dataset.
Report
Comprehensive monitoring and audit-ready reporting with compliance dashboards.
- •Continuous monitoring of data access and movement.
- •Audit-ready reports exportable for regulators and partners.
- •Dashboards with compliance scores and trend analysis.
AI inside Guard
Contextual classification
LLMs detect sensitive context even when fields are mislabeled.
Adaptive policies
Rules learn from user decisions and extend to new domains.
Explainable actions
Each mask or block includes rationale, confidence, and lineage.
Natural language governance
Describe a requirement in plain English, Guard generates and enforces the policy.
Architecture overview
Comprehensive governance architecture from ingestion to reporting with tight integration across the suite.
Three-layer inputs into a four-stage governance pipeline. Policies compiled and enforced by the Policy Engine, with identity, keys, lineage, and audit spanning end to end.
Common use cases
Financial services
Enforce PCI DSS, detect account numbers, and redact before export.
Healthcare
Identify PHI, mask in logs, and generate HIPAA audit reports.
Global enterprises
Apply GDPR and CCPA policies across multi-region warehouses.
AI training pipelines
Screen datasets before model fine-tuning to prevent leakage.
Government contracts
Ensure classified data is handled only by cleared accounts.
Example policy packs
GDPR Europe
Detect and restrict personal identifiers, auto-generate DSAR reports.
CCPA California
Handle "Do Not Sell" flags, consent tracking.
HIPAA US
Classify PHI, mask fields, redact notes.
SOC 2
Control access to financials, monitor use.
Custom
Define bespoke rules in natural language.
End-to-end workflow
Guard scans datasets from Fabric or external sources.
Sensitive data is flagged, masked, or routed for approval.
Policies are applied in real time as data flows.
Access events and enforcement are logged.
Audit reports generate automatically.
Studio and Signals consume governed datasets with lineage attached.