AWS walkthrough: build a semantic layer for agentic AI with Strategy Mosaic and Bedrock AgentCore
AWS ML Blog · rss · 2026-10-20
AWS ML Blog's third semantic-layer walkthrough swaps Strategy Mosaic into the same Customer 360 pattern, running a Strands agent on Amazon Bedrock AgentCore against Mosaic's dimensional semantic layer.
- The series previously used Stardog (ontology/RDF) and Denodo (data virtualization); Mosaic uses linked dimensional models with a metric expression syntax.
- One model definition serves four protocols: MCP for agents, JDBC/ODBC via Trino, XMLA for Power BI/Excel, and REST, backed by certified sources including Snowflake, Databricks, RDS, Athena, and Salesforce.
- No-ETL federation: a Purchase model over Redshift and a Customer model over Aurora are linked on shared cid attributes; queries push down natively and joins resolve in the Mosaic engine.
- Governance lives in the semantic layer, not the prompt: joins, metric definitions, and per-caller RLS/OLS propagate end-to-end through the MCP endpoint.
- Each model can run live pushdown or in-memory execution. Requires Strategy One (March 2026), when MCP access became default.
More from coding & agent
- Claude Code Pushes Urgent Security Update Across All Installations — umeshonai · 2026-10-23
- Claude cut a 2.6-hour video export to 20 minutes — but only after being told to rebuild the server — byhewar · 2026-10-23
- 3 flight-booking rules my AI agent checks that I'd never do by hand — Efistoffeles · 2026-10-23
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Building an autonomous SEO agent in Hermes with three skills: programmatic pages, Keywords Everywhere, DataForSEO — EXM7777 · 2026-10-23
- Building a sports betting AI agent with onchain execution and human approval by default: architecture and lessons — Dry-Comfortable-2678 · 2026-10-23