AWS Launches Context Intelligence Stack for AI Agents, Featuring Self-Learning Knowledge Graph

Amazon Web Services (AWS) has entered the competitive market for AI context layers with the launch of a new context intelligence stack for AI agents. The suite includes three new products designed to build and maintain knowledge graphs between enterprise data and artificial intelligence.

The centerpiece of this release is AWS Context, a new knowledge graph service. This service automatically builds a knowledge graph from existing enterprise data, inferring relationships across datasets, business rules, and domain knowledge. It is designed to get smarter over time by learning from how AI agents use it, rather than relying on manual human re-curation. Swami Sivasubramanian, vice president of Agentic AI at AWS, stated that the service allows agents to become more intelligent without requiring rebuilding from scratch, as it infers relationships and makes them available at runtime.

AWS Context maps relationships across data, identifying tables, column meanings, and authoritative sources. It integrates semantic search with graph-level reasoning, improving itself as it learns which sources yield correct results and which components are frequently used. Data stewards can manage the graph via the AWS Management Console, reviewing inferred relationships, promoting them, and attaching business definitions and usage rules. Access to agent data is auditable through existing IAM and Lake Formation permissions, with metadata published in Apache Iceberg format to Amazon S3 Tables.

In addition to AWS Context, the company announced the general availability of Amazon S3 Annotations. This service allows users to attach rich business context directly to individual S3 objects at the storage layer. AWS also introduced a preview of skill assets in AWS Glue Data Catalog, which enables the attachment of domain knowledge at the catalog layer, linking runbooks, query patterns, and usage rules to data assets.

These services are designed to work together, with S3 Annotations and Glue skill assets feeding into AWS Context to synthesize a comprehensive knowledge graph for agents to query. AWS's approach focuses on zero-integration friction for enterprises already utilizing S3, Glue, and Lake Formation, extending their existing identity models without requiring data movement.

The context layer market is becoming increasingly competitive, with offerings from various vendors. Snowflake has its Horizon Context and Cortex Sense services, Microsoft offers context through its Fabric IQ platform, Redis has developed a context platform for data optimization, and Pinecone provides its Nexus context offering. Holger Mueller, VP and Principal Analyst at Constellation Research, noted that performance, especially for transactional data, will be a key consideration for all context offerings.

According to VentureBeat, AWS's entry aims to standardize the previously bespoke work of building context layers for AI agents.