Context intelligence for your data and AI agents at scale
TL;DR
AWS announced AWS Context, an upcoming service that maps data relationships into a knowledge graph and gives agents governed business rules, domain knowledge and data context at runtime. Glue Data Catalog is getting Business Context and Semantic Search in preview, including descriptions, glossary terms, custom metadata and skill assets for agents. Amazon S3 Annotations are now generally available: each S3 object can carry up to 1 GB of mutable, queryable context data surfaced through Iceberg tables.
Nauti's Take
This sounds like infrastructure, but it solves a very practical agent problem: without clean context, every copilot becomes guesswork with a polished UI. AWS unsurprisingly wraps the answer deeply into its own stack, and the announcement is heavy on product positioning.
Still, the direction is right: serious agent deployments need data knowledge, permissions and operating rules available in a form machines can retrieve at the right moment.
Briefingshow
The real story is not another agent feature, but the move from prompt stuffing to managed enterprise context. If permissions, metadata, business rules and data relationships live in one governed layer, agents can guess less and act with more auditability. The announcement is clearly AWS-centered, but it points to a broader enterprise AI shift: context is becoming infrastructure.
Sources
