LLMs are great at language, but they struggle with logic across disconnected data. In this workshop, we’re using Neo4j and Microsoft Foundry to provide the source of truth that agents need. You’ll build a pipeline that transforms raw Customer 360 data into a structured knowledge graph, giving your AI the context it needs to actually be useful.
In this session, data scientists and engineers will build a functional GraphRAG pipeline within a live Azure environment. Moving beyond high-level theory, you will deploy Neo4j Aura and integrate it with Microsoft Foundry to implement a complete, end-to-end architecture for reasoning over connected data.
The Data & Workflow
We’ll use a Customer 360 dataset—specifically the orders, clickstream events, devices, products, and support tickets of these customers as our primary dataset. This data is ideal for demonstrating how to map complex ownership structures and many-to-many relationships.
During this session, you will:
- Streamline Ingestion: Use Cypher to parse and load data from Azure Blob Storage into Neo4j.
- Model Relationships: Use Cypher and the Neo4j Browser to map connections between thousands of customers orders and their clickstream events.
- Orchestrate Agents: Layer a Microsoft Foundry AI agent over the graph to enable natural language reasoning across the connected customer data.
Practical Application
For those in the customer satisfaction marketspace, this architecture provides a blueprint for algorithmic product recommendation, user behavioral patterns, and support ticket satisfaction. If you work in another industry, this session will still be useful for learning how to build agentic AI pipelines with Neo4j and Microsoft Foundry.
Infrastructure & Integration
Neo4j is integrated into the Azure ecosystem as a managed service. Neo4j Aura is available natively through the Azure Marketplace, enabling unified billing and direct deployment within your existing Microsoft environment. This workshop utilizes the engineering alignment between Neo4j and Microsoft Foundry to demonstrate production-ready AI orchestration.
This event is free, but available seating is limited. Please register today to request a seat; we are reviewing and granting entry to applicants on a rolling basis.
