Build a semantic layer for agentic AI on AWS with Strategy Mosaic and Amazon Bedrock AgentCore
aws ·
This article details how to build a semantic layer for agentic AI on AWS using Strategy Mosaic and Amazon Bedrock AgentCore. It extends previous work by replacing other platforms with Mosaic’s dimensional modeling approach, which links Amazon Redshift and Amazon Aurora without ETL. The solution enables a Strands agent to query customer data across these sources via the Model Context Protocol (MCP). Key features include governed metrics, row-level security, and identity propagation from the user through AgentCore to Mosaic. The architecture supports live pushdown or in-memory execution to optimize performance for repetitive agent queries. Additionally, the same definitions are exposed to BI tools like Power BI and can be exported via Apache Ossie for interoperability. The walkthrough demonstrates accurate cross-engine joins and secure access control for different user roles.
aws