Bedrock is a data management and security platform built on patented data discovery and classification technology. The platform creates a Metadata Lake, a graph knowledge base that serves as a central repository for enterprise data context, mapping relationships between metadata elements such as data classification, entitlements, lineage, usage patterns, and ownership without storing customer data. The platform performs Full-Context Entitlement Analysis, mapping the complete access chain for every identity by resolving nested groups and integrating federated identity data to calculate effective permissions for users and service accounts. An Impact Score ranks identities by the sensitivity and volume of data they can access. Bedrock uses Adaptive Scanning to group similar data objects by analyzing file paths, structure, and metadata to select representative samples, achieving comprehensive coverage across petabyte-scale environments. The platform deploys via a Serverless Outpost Architecture using infrastructure-as-code, operating as ephemeral functions that perform discovery and classification directly within the customer's environment. The platform adheres to a Zero Data Access principle, where all scanning, classification, and analysis occur in the customer's environment through Serverless Outposts. Only anonymized metadata is transmitted to Bedrock's SaaS for visualization and policy management. A Graph API provides programmatic access to contextual information stored within the Metadata Lake, enabling automation and integration with existing systems. The Natural Language Policy Engine ingests policies from GRC documents, interprets intent, identifies relevant data types, and maps them to the Metadata Lake's query language for automated enforcement.