RelationalAI Now Available In Snowflake

Today at the Snowflake Summit we announced the public availability of RelationalAI as a Native App for Snowflake. Built as a Snowflake Native App on Snowpark Container Services, RelationalAI now runs fully embedded within Snowflake as a relational knowledge graph coprocessor. It extends the Snowflake Data Cloud by supporting graph analytics and rules-based reasoning workloads, leading to faster and higher quality decisions. This all happens within your Snowflake account, maintaining the same ease of use, scalability, security, and governance you are accustomed to.

Decision-making within organizations involves many domain experts, manual processes and siloed systems. This complexity makes the process expensive, time-consuming, and prone to errors. Despite advances in cloud data storage and computing power, creating intelligent applications that leverage AI to make sense of the relationships in data and utilize business knowledge remains a challenge. Until now.

Knowledge graphs turn an organization's collective understanding into a comprehensive model of the business. This digital representation captures the essential details of your operations. When combined with Generative AI, knowledge graphs use symbolic logic to boost AI performance.1,2 This pairing enables us to reason through and solve complex business problems that were previously too difficult to tackle.

How does RelationalAI work?

With RelationalAI, you can capture the distributed knowledge in your organization and model your business as a relational knowledge graph. Think of this as bringing siloed knowledge together from across your organization to form a cohesive model for decision-making. The model represents important concepts, relationships, and rules grounded in business data, to power fast, precise and repeatable decisions.

The relationships in the knowledge graph surface connections that are difficult to see otherwise. RelationalAI provides built-in graph algorithms to utilize these relationships for detecting centrality, similarity, and communities, as well as predicting links and paths. Additionally, rule-based reasoning offers a compact, reusable approach to representing both knowledge and business logic. Customers have experienced a significant reduction in business logic complexity, through a 10 to 100x reduction in code, by leveraging RelationalAI to condense existing application logic into a concise model of the business.

Knowledge graphs often demand that users learn specialized modeling and query languages, which can be a barrier to adoption within organizations. RelationalAI, however, is designed to be accessible to everyone in your organization working with data. Through RelationalAI’s Python package, users can easily create models, rules, and relationships derived from your Snowflake data. The Python query builder simplifies the process of utilizing the knowledge graph to uncover new insights. By exporting Python functions to Snowflake procedures, models and logic become accessible to anyone using SQL through existing tools.

What does fully embedded in Snowflake mean?

RelationalAI is natively integrated with the Snowflake Data Cloud and runs entirely within a customer’s Snowflake account. This integration drives a number of key benefits for customers.

  1. Data governed by Snowflake. Customer data never leaves the boundary of a Snowflake account. RelationalAI does not open outbound network access or communicate with services outside of an account boundary. All data stored by the RelationalAI service is persisted in Snowflake storage containers including tables and stages and is protected via Snowflake’s security and governance capabilities.
  2. Operational ease. RelationalAI is operated as a cloud service using Snowflake’s built-in support for eventing and diagnostics. RelationalAI resources are secured by Snowflake’s authentication and authorization systems including support for role based access control. Management and monitoring commands are provided through SQL commands for administrators. Data can be materialized into a knowledge graph without extracting data or configuring pipelines and results returned to Snowflake tables for use by anyone through existing tools and processes within the organization.
  3. Simplified procurement and purchasing. The RelationalAI service can be procured and provisioned through the Snowflake marketplace and customers can use their existing Snowflake capacity commitments towards RelationalAI through Snowflake’s Marketplace Capacity Drawdown program. Updates to the service are continuously delivered through the Native App Framework. Customers only pay for what they use with a consumption-based pricing model and RelationalAI’s separation of compute and storage allows for flexible consumption and cost optimization.


Fully embedded in Snowflake


Customers using RelationalAI today

See what some of our customers are able to do today with RelationalAI and Snowflake.

“Thanks to RelationalAI, we are developing a deeper understanding of our customers. RelationalAI’s knowledge graph coprocessor enables my team to see patterns, identify customer needs, and better serve our customers. With RelationalAI, we can now do these sophisticated graph analyses within our Snowflake environment, saving us significant time and money. And RelationalAI’s performance is superb. What used to be done in days, can now be done in minutes.”

Cristian Figueroa
Head of Network Science and Behavioral Modeling, Cash App

“RelationalAI is foundational to our efforts to transform and modernize our company. With RelationalAI's knowledge graph, we are easily able to encode a deep understanding of our business, resulting in a highly personalized online marketplace. RelationalAI's state-of-the-art technology applies business knowledge to our data, enabling key capabilities that we consider key to our competitive advantage in our industry, powered by a knowledge graph, all within our Snowflake environment. Thanks to RelationalAI, we are looking to lean on modern AI to remove manual processes and more importantly drive accuracy and commerce flow. The resulting customer experience will be game-changing."

Ranbir Chawla
Senior Vice President of Engineering, Ritchie Brothers

“RelationalAI has been instrumental in our quest to modernize our solutions to be cloud native working off of our Blue Yonder Platform, Microsoft Azure and Snowflake Data Cloud. Running as a Knowledge Graph Coprocessor inside of Snowflake, RelationalAI provides the semantic understanding and reasoning capabilities that enable our customers to predict any disruptions and proactively drive mitigation actions. RelationalAI provides us with the foundation to offer a scalable and extensible solution while reducing our legacy code by over 80%.”

Amanpreet Singh
Corporate Vice President, Blue Yonder

Get started today

Find us on the Snowflake App Marketplace today and request access to RelationalAI. Once you have provisioned RelationalAI, you can get started in your favorite Python environment by using one of the published solutions on Snowflake’s Solution Center, such as community detection.

¹ Compound AI

² Increasing the LLM Accuracy for Question Answering: Ontologies to the Rescue!, GraphRAG: Unlocking LLM discovery on narrative private data