Build an open data foundation for agentic AI

An open data foundation for agentic AI is a scalable, interoperable data architecture that lets AI agents find, access, trust, and act on data across systems. It uses open-source technologies, open table formats, and managed data services so organizations can avoid vendor lock-in, reduce data reformatting, and support agents that operate across databases, applications, analytics tools, and AI workflows. 

AI agents need an open data architecture because they often work across multiple systems, tools, and services to complete tasks. Open architectures make data usable across environments without forcing proprietary formats or duplicate pipelines. This improves interoperability, supports cross-system agent workflows, and helps organizations connect operational data, analytics data, search indexes, and AI applications more efficiently.

AWS helps make enterprise data AI-ready by combining managed databases, analytics services, open-source technologies, vector search, governance, and security controls. Services such as Amazon Aurora, Amazon RDS, Amazon EMR, Amazon MSK, Amazon OpenSearch Service, Amazon DynamoDB, Amazon ElastiCache, and Amazon SageMaker support scalable data processing, retrieval, access control, and agentic AI application development.

Trusted data is essential for agentic AI because agents use data to make decisions, generate responses, trigger workflows, and take action. Inaccurate, inaccessible, or poorly governed data can reduce reliability. AWS supports trusted data with encryption by default, fine-grained access controls, and secure data movement across humans, applications, and AI agents, including sensitive enterprise workloads. 

Vector search supports agentic AI by enabling context-aware retrieval from large volumes of unstructured and structured data. It helps agents find semantically relevant information, not just exact keyword matches, which improves retrieval-augmented generation, search experiences, recommendations, and task execution. Amazon OpenSearch Service provides vector search capabilities for building performant and reliable AI applications. 

Apache Iceberg supports an open data foundation by providing an open table format that works across different query engines and processing frameworks. This allows organizations to use the same data across analytics and AI workloads without modifying or reformatting it for each tool. On AWS, Iceberg can be used across the data and AI stack to improve interoperability.