The AI Coprocessor:
A Groundbreaking
Relational Knowledge
Graph System
Jeff Hollan
Director, Product Management
Seizing on the opportunities presented by AI today requires new innovation that brings together data, business logic, and an environment that supports multiple AI techniques. All in one place.
This is why we built the AI coprocessor for data clouds and language models. Using RelationalAI, organizations can layer multiple AI methodologies to surface knowledge, enhance collaboration, and drive consistency across the enterprise.
Our AI coprocessor uses a groundbreaking relational knowledge graph system. Previously, knowledge graphs modeled business concepts and the relationships between them. Our innovation is reimagining knowledge graphs as systems that execute your business logic and run multiple composite AI workloads from within your data cloud, eliminating data movement and duplication. RelationalAI is designed with a cloud-native architecture and the relational paradigm, making it compatible with your data cloud.
RelationalAI supports multiple composite AI workloads including:
Graph Analytics — Explore the connection within data, discover influential elements, reveal meaningful patterns, and create features to improve predictive models.
Business Rules — Define and maintain dynamic sets of business logic to automate actions, drive consistency, and enhance AI governance.
Optimization — Continually calculate the best course of action, balancing multiple constraints across millions of parameters as situations change.
RelationalAI brings together the best of relational technology and knowledge graphs to create an AI coprocessor that expands your data cloud with new capabilities.
Get zero-copy cloning, workload isolation, and effectively infinite storage and compute.
Represent graphs and relational data models efficiently as atomic, irreducible relations. Benefit from a universal data representation that can be projected as tables, graphs, JSON, tensors and matrices.
Benefit from the latest join algorithm and indexing techniques that dramatically speed up graph traversal, multi-way joins, and complex queries without sacrificing performance.
Gain support for new workloads while adhering to the same architecture, paradigm, and governance. Access knowledge graph capabilities through a single SQL language interface with a shared view of data.
Get native support for a wide range of algorithms for common graph analytics tasks including centrality, community detection, similarity, and path analysis.
Embed and execute rules and business logic in the knowledge graph. Compose applications as modular units of logic to drive consistency and reusability across the organization
Express and experiment with optimization models to solve business objectives with any open source and commercial solvers.
Reason over data with tailored SQL functions and procedures that provide seamless access to knowledge graphs. Utilize the command line interface, web console, and SDKs for popular programming languages.
RelationalAI extends Snowflake’s Data Cloud with composite AI capabilities, enabling customers to harness their existing Snowflake governance and security solutions and innovate on a consistent platform, enhancing performance and simplifying data management. RelationalAI is always in sync with your Snowflake data, enabling teams to develop knowledge graphs by layering concepts, relationships, and logic over Snowflake databases without the complexity of managing copies of data or moving data outside your data cloud.
Learn more >See what organizations are achieving using RelationalAI for business-critical workloads.
"We slashed legacy code by 90% and reduced processing times from over a month to several hours."
— Tax Technology Leader, EY Financial Services Office
"Acting as an AI coprocessor, RelationalAI enables us to enhance our semantic models and perform sophisticated analysis like graph analytics… "
— Mark Austin, VP of Data Science and AI, AT&T
Consumption Based Pricing
Pay-as-you-go with no upfront costs, no annual charges, and no costs for paused workloads.
Only pay for the resources you consume, billed at per hour granularity. Compute and storage are metered separately, with the ability to dynamically add and remove compute resources. Get started for as little as a couple dollars an hour.
Start your journey with RelationalAI today! Sign up to receive our newsletter, invitations to exclusive events, and customer case studies.
The information you provide will be used in accordance with the terms of our Privacy Policy. By submitting this form, you consent to allow RelationalAI to store and process the personal information submitted above to provide you the content requested.