AWS details semantic-layer architecture for agentic AI with Stardog

AWS published a reference architecture for building a semantic layer for agentic AI with Stardog and Amazon Bedrock AgentCore. The example connects Stardog's semantic AI application over Aurora and Redshift, then uses a Strands Agents agent to query across both sources.

The goal is to give agents reliable business context without forcing teams to move all data through a separate ETL process. Semantic layers can help agents understand relationships among customers, records, and systems instead of relying on brittle prompt context alone.

For enterprise AI teams, the post highlights how knowledge modeling, data access, and agent orchestration are becoming part of the same application stack.