Power every decision with intelligence

To make fast, repeatable, future-ready decisions across the company, you have to put your organization’s knowledge to work. RelationalAI uses a relational knowledge graph and 4 different AI workloads to operationalize all the data, rules, relationships, and decision systems that drive your business—all while staying entirely within your Snowflake data cloud.

get started

The Knowledge Graph…

Think of the knowledge graph as a “collective common sense.” It doesn’t just capture that you have a sales table, it captures that your team keeps an eye on transactions over a certain amount that come from one of your more unpredictable customers.

dive in

… Coprocessor

Instead of creating another silo, we integrate powerful decision-making tools like graph reasoning, rules-based reasoning, and mathematical optimization directly into your data cloud. Our native application, built on Snowpark Container Services keeps everything securely within the Snowflake perimeter.

explore

Decisions, simplified

The knowledge graph is a map for the decisions a business makes. By traversing and annotating that map, you carve a path around all the tangled systems that slow decisions down and start pushing the company forward. Let’s look at a few examples.

Which financial transactions need human reviewUnderstanding how money, items, and data flow through your business is a key for detecting issues like fraud before they become significant liabilities. Use AI to label suspicious actors and then traverse the transaction graph for situations that need review.

with model.rule():
    t = Transaction(from_=Customer(), to=Suspicious(Person))
    t.set(Suspicious)
    with t.amount > 40000:
        t.set(NeedsReview, 
              reason="Large transaction to suspicious person")

Which customers in a social network are the most influentialIdentifying key influencers within a customer base enhances marketing strategies and engagement. Using graph algorithms like PageRank, businesses can determine the most influential customers based on their connections and interactions, enabling targeted marketing and efficient resource allocation.

with model.rule():
    msg = Message()
    weight = sum(msg, msg.length, per=[msg.sender, msg.to])
    Edge.add(msg.sender, msg.to, weight=weight)

with model.rule():
    user = User()
    rank = graph.compute.pagerank(user)
    user.set(rank=rank)

Which nodes in a power grid need repairBy leveraging the graph structure of the power grid, repairs can by easily prioritized by impact, ensuring critical loads and distribution centers are automatically prioritized first.

with model.rule():
    upstream, downstream = NeedsRepair(), NeedsRepair()
    graph.compute.is_reachable(upstream, downstream)
    upstream.set(
        load_priority = aggregates.max(
            downstream.load_priority.or_(0), per = [upstream]
        ),
        connection_priority = aggregates.count(
            downstream, per = [upstream]
        ),
    )

Beyond Count(*)

For business queries over massive amounts of data, there’s nothing better than Snowflake, but SQL struggles to go beyond aggregations over tables.

Decisions require you to utilize every form of automated intelligence at your disposal, from community detection and centrality analysis for finding your most influential users to mixed integer solvers that determine the optimal allocation of goods in your warehouses.

RelationalAI extends the best that Snowflake has to offer with an array of powerful decision making tools to help deliver on the promise of “all data, all workloads.”

Uncover hidden patterns with advanced graph queries for path finding and graph algorithms for community detection, centrality, similarity, and more

LEARN MORE

Graph Reasoning

Simplify intelligent application development with expressive and scalable rule-based based reasoning

LEARN MORE

Rules-based Reasoning

Predict the impact of your decisions by leveraging GenAI and Graph Neural Networks

LEARN MORE

Predictive Reasoning

Optimize business operations with sophisticated solvers for mixed integer programming (MIP) and satisfiability (SAT)

LEARN MORE

Prescriptive Reasoning