Creating a context layer between enterprise data stores and AI agents is a specialized task, with no standard service to automate or maintain graphs over time. Amazon is making a direct effort to change that.
Amazon entered this field on Wednesday, announcing a series of three products it is offering as a reference intelligence stack for AI agents. The centerpiece is AWS Context, a new knowledge graph service that gets smarter over time through agent use. AWS also announced the general availability of Amazon S3 annotations and a preview of skill assets in the AWS Glue data catalog.
The context layer is now a contested architectural category with no shortage of options from different vendors. AWS is entering that market with a different architectural foundation: The graph must learn how agents use it automatically, without human reengineering.
"Your agents are now smarter without having to rebuild anything," said Swami Sivasubramaniam, vice president of agentic AI at AWS, during his AWS Summit NYC keynote.
"This service automatically creates a knowledge graph from all your existing data," He said. "The service infers relationships across your data set, business rules, and domain knowledge, and makes it all available to your agents and your organization at runtime."
AWS Context builds a self-learning knowledge graph from existing data
AWS says this is a problem it has seen repeatedly in customer deployments.
AWS Context automatically maps relationships in existing data: which tables exist, what columns mean, how sources are related, and which sources are authoritative. It combines semantic search with graph-level reasoning and infers relationships across datasets, business rules, and domain knowledge, making it all available to agents at runtime.
"The knowledge graph improves itself over time as it learns which sources give the correct results and which parts are used," Sivasubramaniam said.
Data managers manage graphs through the AWS Management Console, review inferred relationships, promote them to production, and attach business definitions and usage rules. Each query inherits the calling user’s IAM and Lake Formation permissions, allowing agent data access to be audited based on identity through the controls that enterprises already trust.
All metadata is published to Amazon S3 tables in Apache Iceberg format, which can be queried via Athena, Redshift, Spark, or any Iceberg-compatible engine, with no proprietary APIs. Third-party catalog connections are supported, so context from systems outside AWS can be pulled into the same graph. Agents query through the Agentic Search API and MCP tools in Bedrock AgentCore, EKS, or any MCP-compliant framework.
Reference is more than just a service
Context is a complex space and AWS is building multiple layers of services to help enterprises create context across the data stack.
Amazon S3 Annotations. This service enables users to link rich business context directly to individual S3 objects at the storage level.
AWS Glue Data Catalog Skills Assets. Glue Skills attach domain knowledge at the asset catalog layer, connecting runbooks, query patterns, and usage rules to data assets across the estate.
AWS Context then synthesizes both into knowledge graphs that agents query at runtime, combining semantic search with graph-level logic across structured and unstructured sources. Each layer feeds the next layer.
AWS is entering a highly competitive reference space
Snowflake announced its context approach earlier this month with its Horizon Context and Cortex Sense services. Microsoft is providing context through its Fabric IQ platform which provides a semantic ontology for data. Redis has developed a context platform that optimizes data for retrieval. Vector database vendor Pinecone has a Nexus Context offering that compiles enterprise data into task-specific artifacts before agents can query them.
AWS’s structural logic is straightforward: For enterprises already running S3, Glue, and Lake Formation, AWS Context extends an existing identity model with no data movement required. The pitch is zero-integration friction – not just cost consolidation.
"Context makes agents more powerful and since the whole world is building agents, every agentic platform vendor needs a context capability," Holger Mueller, vice president and principal analyst at Constellation Research, told VentureBeat.
Mueller said AWS is no exception. "The concern – as with all reference offerings – is around performance, especially for transactional data, we will see," He said.
<a href