Neo4j NODES 2025 – rewatch our experts’ sessions

Organised crime networks, from local fraud rings to complex international syndicates, operate through intricate webs of relationships that traditional analytical approaches often fail to uncover. In this session, Alessandro demonstrates how a combination of Neo4j’s GDS library with agentic AI architectures creates a powerful platform for law enforcement intelligence analysis.

The session explores a complete workflow that transforms raw criminal data into actionable intelligence reports through two complementary approaches.

First, you’ll learn how to leverage Neo4j GDS algorithms, including community detection (Louvain, Label Propagation), centrality analysis (PageRank, Betweenness), and other graph-specific techniques to identify criminal organisations and key players.

Second, you’ll discover how to implement LangGraph-based agentic systems that automatically recognise temporal and geographical patterns as well as activity evolution, and generate comprehensive, human-readable reports from these analytical findings.

Alessandro showcases two distinct agentic architectures:

• A controlled parallel workflow where specialised agents handle specific analytical tasks (group demographic analysis, temporal and geographical evolution, threat assessment, report generation) through coordinated but well-defined responsibilities
• A fully autonomous approach where agents dynamically determine their own analysis paths.

By the end of this technical session, you will understand practical implementation details, including Cypher queries for data modelling, GDS algorithm selection and tuning, LangGraph workflow orchestration, and prompt engineering strategies for generating accurate, contextually-aware reports.

The session also covers critical considerations for deploying such systems in sensitive law enforcement contexts, including data privacy, algorithmic transparency, and human-in-the-loop validation.

Graph visualisation matters. It’s the vital link between connected data and human insight.

Relying on third-party libraries is initially convenient but may limit your flexibility, both in terms of critical features for a great user experience and your business model.

This session covers how we engineered a fully custom graph visualisation from scratch. From key architectural patterns to rendering with WebGL, spatial indexing for interactivity, animations, and GPU-accelerated layouts, you’ll gain insight into the key components and how they fit together.

By the end, you’ll understand what challenges to expect, how to approach key design decisions, and whether building in-house is right for your product.

In a world overflowing with unstructured data, from emails and reports to PDFs and legal documents, transforming this textual overload into actionable knowledge is more urgent than ever.

This session explores how the fusion of GenAI and graph technology can reveal the stories hidden deep within complex text.

We’ll dive into a powerful process that mimics human reasoning, identifying facts, events, and relationships in natural language, and transforming them into rich, queryable graph structures.

You’ll discover how to specify custom entities and relationships using one-shot prompting, and how to normalise them into a coherent graph. We’ll also showcase a human-in-the-loop process that empowers analysts to verify, refine, or discard extracted content with just a few clicks, ensuring quality, relevance, and trust in the resulting knowledge graph.

By combining the contextual power of GenAI with the structural clarity of graphs, this presentation will demonstrate how unstructured documents can evolve into dynamic knowledge networks, ready to power smarter systems and decisions.

In a world increasingly driven by Large Language Models, organisations face a difficult question: how can we leverage these powerful tools without compromising sensitive data?

In some industries – finance, healthcare, government, for example – security and privacy are non-negotiable.

This talk will walk through real-world challenges we’ve encountered when dealing with anonymisation in the LLM era. We’ll look at the limits of traditional techniques, explore emerging strategies, and discuss the trade-offs between privacy, utility, and cost.

This talk is less graph-heavy than usual, but highly relevant for the Neo4j community. After all, we’re the people building knowledge graphs, handling sensitive relationships, and connecting the dots in complex datasets.

Privacy isn’t someone else’s problem, it’s ours too!