Autonomous Data Product

An autonomous data product is a self-contained, self-managing long-running service or application that encapsulates and orchestrates all necessary components for data generation, transformation, governance, and access.

Each autonomous data product includes data, metadata, code, policies, and semantic models, and operates independently within a larger data ecosystem. ^[1] Designed to be discoverable, addressable, and governed by design, autonomous data products enforce quality, privacy, and access controls programmatically throughout their lifecycle. They self-orchestrate workflows, manage upstream and downstream dependencies, and expose health and usage metrics in real-time.

This concept supports decentralized data architectures, such as data mesh, by enabling domain-oriented teams to independently produce and manage data as a product, while still being programmatically governed and observable to ensure regulatory and policy compliance. Autonomous data products are particularly suited to AI-driven environments, where both human and machine agents require trustworthy, up-to-date, and programmatically accessible data at scale.^{[2] [3]}

The term was originally used by Zhamak Dehghani to describe the behavior of self-contained data products independently interoperating as part of a data mesh architecture, ^[4] a paradigm that she originated while working as a consultant at Thoughtworks. ^{[5] [6]} The term was subsequently popularized by Nextdata, ^[7] the company Dehghani founded in 2022.^[8]

See also

• Data product