RelationalAI introduces new graph processing features for its Snowflake app

Startup RelationalAI Inc. today introduced new features for its software that will enable companies to analyze their data more efficiently.

The capabilities debuted at the annual Snowflake Summit in San Francisco.

Snowflake provides a popular cloud data platform that companies use to store, analyze and visualize their data. Customers can optionally extend the platform’s feature set by installing apps from third-party providers. Those apps run directly inside Snowflake as software containers.

The company provides a Snowflake app that can be used to find useful patterns in business data. The service also lends itself to other tasks, notably enhancing the output of Snowflake-powered artificial intelligence applications.

RelationalAI can hold the data that it processes in the form of graphs. A graph is a data structure that represents records as points and the connections between those records as lines. The points could, for example, represent a company’s logistics facilities while the lines may represent the roads that connect them. This abstraction can be analyzed to find shorter delivery routes.

Graphs are also useful for a range of other data analysis tasks. RelationalAI lists demand prediction and customer churn detection among the use cases supported by its newly upgraded Snowflake app.

As part of the update, the company is rolling out new algorithms for analyzing datasets stored as graphs. The additions include path finding algorithms, which are useful for tasks such as identifying the fastest delivery route between two stores. There’s now also support for egonet analysis. This is a popular method of studying relationships between data points.

According to RelationalAI, today’s update will also enable customers to analyze their graph-based datasets with GNNs, or graph neural networks. Those are AI models optimized specifically to work with graphs. GNNs are useful for tasks such as forecasting future demand for a product.

The new AI models are joined by a text-to-reasoner capability and mathematical optimization solvers. According to RelationalAI, the former feature can use AI to help companies predict business developments. Mathematical optimization solvers, in turn, are algorithms that calculate the optical way of going about meeting a business goal such as lowering product delivery costs.

Rounding out the list of new features is support for semantic views. Those are data structures in Snowflake that can hold information about relationships between business entities, such a company’s subsidiaries or suppliers. RelationalAI uses semantic views to enhance Cortex Analyst, another Snowflake feature that lets workers ask questions about data in natural language.

“RelationalAI’s knowledge graph has the potential to be a game changer for customers looking to harness AI within their existing Snowflake environments, making the process simple and streamlined,” said Unmesh Jagtap, Snowflake’s director of product management for applications.

Image: RelationalAI