Build a semantic layer for agentic AI with EDB Postgres AI and Amazon Bedrock AgentCore

This article demonstrates how to build a semantic layer for agentic AI using EDB Postgres AI and Amazon Bedrock AgentCore, enabling agents to answer business questions from live data with governed access.

• Create a semantic knowledge base (SemKB) inside Postgres using COMMENT ON statements and foreign keys to define business meaning alongside data
• Expose the AIDB tool catalog to external agents through an MCP endpoint fronted by an Application Load Balancer and registered with AgentCore Gateway
• Enforce access control using Postgres roles, column-level GRANT statements, and row-level security policies that apply to all agents uniformly
• Capture OpenTelemetry traces for every SQL execution as queryable rows in the aidb_otel schema for audit and observability
• Define recurring questions as semantic aliases and business terms as views with comments to improve agent accuracy and consistency
• Route requests from Cognito-authenticated users through AgentCore Gateway to the semantic layer with identity propagation via HTTP Basic auth

This architecture unifies the semantic layer, data, and governance in a single Postgres database, eliminating separate catalog systems and enabling reliable agentic analytics with built-in audit trails.