Data, logic, and learning
in one single architecture
Data-centricSemanticAgentic
Encode business logic, semantics, and relationships
Relational knowledge graphs bring context by encoding business logic, semantics, and relationships. Data alone doesn’t explain itself, RKGs make data understandable.
The RKG serves as the semantic brain of your enterprise— empowering deeper insights, more responsive decisions, and automation inside the Snowflake AI Data Cloud.
Relational knowledge graph
Models concepts and relationships in graph normal form (GNF) — bringing meaning and structure beyond tables and joins and transposing data into an interconnected graph of entities, properties, and relationships.
Semantic layer
Defines shared meaning, enabling consistent interpretation across models, reasoning workloads, and applications.
AI reasoning workloads
- Graph reasoning: Understand complex relationships and hierarchies through interconnected data models
Rule-based reasoning: Enforce business logic dynamically
Predictive reasoning: Surface patterns and trends from structured and interconnected data
Prescriptive reasoning: Recommend optimized decisions based on operational goals
GraphRAG: Combine semantic graph structures with retrieval-augmented generation (RAG) techniques for richer generative AI applications
Enablement for applications and agents
Applications query the knowledge graph through familiar APIs, receiving dynamic, contextual, and responsive insights grounded in governed relational data.
Our latest benchmarks
<1s
Graph query performance across billions of rows
1s/1000
Latency for rule-based reasoning across thousands of rules
10x
Faster semantic joins compared ot manual SQL logic
~100ms
Response time for GraphRAG use cases
KNOWLEDGE GRAPH MODELING
Transform relational tables into rich graphs, natively inside your data cloud.
SEMANTIC LAYER INTEGRATION
Unify business concepts, data models, and application logic into a coherent layer.
REASONING WORKLOADS
Execute graph, rule-based, predictive, and prescriptive reasoning workloads directly over your graph.
COST AND ACCESS MANAGEMENT
Fine-grained access control and cost transparency, integrated with Snowflake governance.
DYNAMIC GRAPH REASONING
Run responsive queries and computations over evolving graph structures.
DEVELOPER TOOLKIT
Python SDK, graph visualization tools, stream APIs, and Snowflake-native SQL functions.
How it works
| Faster development | accelerate intelligent app development with graph modeling and semantic unification. |
|---|---|
| Responsive insights | provide dynamic, contextual answers instead of static dashboards. |
| Operationalized reasoning | embed rule-driven, predictive, and prescriptive decision support directly into applications. |
| Snowflake-native trust | stay compliant with your existing Snowflake security, audit, and governance processes. |
| Scalable performance | optimize reasoning workloads for performance across large, complex datasets. |
Why RelationalAI?
Key
differentiators
| RelationalAI | Traditional graphs | Point solution apps |
|---|---|---|
| Built for relational and graph workloads | Graph only | App-specific |
| Native semantic layer | Trivial/basic semantics | Semantics are embedded in imperative code |
| Snowflake-native deployment | Separate infrastructure needed | Separate infrastructure needed |
| Dynamic reasoning at scale | Static graph queries | Siloed models |
| Unified platform for apps + reasoning | Graph-only reasoning | No shared reasoning |
This demo illustrates how to leverage RelationalAI’s native integration with Snowflake to implement GraphRAG (Graph Retrieval-Augmented Generation). It covers setting up the environment and executing reasoning workloads within Snowflake.
Explore how to detect fraudulent activities by applying graph analytics and reasoning over relational data using RelationalAI within Snowpark Container Services.
Learn how to construct a customer social graph to identify relationships and communities among customers, enhancing personalization and engagement strategies.
Querying
Access responsive insights via SQL, Python SDK, or Snowflake functions.
Visualization
Explore graph structures visually with built-in graph viewers.
Integration
Use Snowflake apps, external APIs, or direct app integration.
Reasoning
Write declarative rules, predictive models, and prescriptive flows.
Modeling
Build knowledge graphs using intuitive APIs.
| Data residency | Your data remains inside Snowflake at all times. |
|---|---|
| Access control | Fine-grained role-based access and resource isolation. |
| Auditing and compliance | Fully integrated with Snowflake security, ensuring traceability and compliance. |
| Cost management | Monitor and optimize costs with transparent compute and storage metrics. |