Martinfowler publishes guide to data contracts
Data contracts apply API design practices to datasets by defining clear contents, schemas, quality rules, and service levels, thereby increasing the reliability of large-scale data and providing explicit metadata for AI agents. The author explains that contracts are machine-readable, verifiable, versioned, and provider-driven, acting as the single source of truth for datasets and functioning as product pages in data marketplaces. Furthermore, they can be integrated into data products and linked to shared ontologies to ensure semantic consistency across an organization. Open article →