From the source
Amazon's walkthrough shows how to build a semantic layer for agentic AI on AWS using Strategy Mosaic over Amazon Redshift and Amazon Aurora, with a Strands agent hosted on Amazon Bedrock AgentCore reaching Mosaic through its MCP endpoint.
- It is the third post in a series that keeps the same customer 360 pattern, after earlier installments used Stardog's knowledge graph and Denodo's data virtualization platform.
- Mosaic's semantic layer is dimensional, composing attributes, hierarchies, facts and metrics through Strategy's metric expression syntax, with linked models as the unit of federation.
- One model definition serves four native protocols: Model Context Protocol for AI agents, JDBC/ODBC over Trino for SQL clients, XMLA for Power BI and Excel, and REST for custom applications.
- The walkthrough models the C360 domain across the two AWS sources using two linked models with no ETL, and keeps joins, metric definitions and per-caller row and column filters inside Mosaic rather than in the agent prompt.
- It also covers choosing between live pushdown and in-memory execution for agent traffic, and publishing the same governed definitions to Power BI, Tableau, notebooks and custom apps.
- The source of truth is Strategy One (March 2026), the release when Mosaic access through the MCP server became available by default, and only certified models are reachable through that server.