RelationalAI

RelationalAI's graph component is its Knowledge Graph Coprocessor, delivered as a native app running inside Snowflake. Rather than a standalone graph database, it operates as an extension of Snowflake's data cloud, materialising existing Snowflake tables and views into a relational knowledge graph without moving data out of the customer's Snowflake environment. The graph is modelled and queried using Rel, a proprietary declarative language built on relational algebra. It supports graph analytics, rule-based reasoning, prescriptive and predictive analytics, and graph neural networks, all executed directly against data held in Snowflake. The underlying graph model does not use W3C standards — it uses a relational paradigm rather than RDF or property graph conventions. Key capabilities include community detection, entity resolution, fraud graph analysis, supply chain modelling, and knowledge graph construction for AI pipelines. It integrates with Snowflake Cortex AI, allowing generative AI models to be combined with graph reasoning in the same environment. It is aimed at enterprises already on Snowflake that need graph analytics and knowledge graph capabilities without deploying a separate graph database or building data pipelines to move data between systems. Customers include AT&T, Block, and Blue Yonder, primarily in financial services, telecommunications, and retail.