Intelligent Graphs in Action: From Agentic RAG to Neo4j in Manufacturing

Step into the next frontier of AI and graph technology at our upcoming Tokyo meetup! 🚀

Agenda:
* 5:40pm~6:00pm: Doors open, Checkin and networking
* 6:00pm~6:30pm: First Talk
* 6:30pm~7:00pm: Second Talk
* 7:00pm~7:30pm: Third talk
* 7:30pm~8:00pm: Open discussion & Mixer, Doors Close

We’ll dive into two exciting themes shaping the future of intelligent applications:

• Agentic GraphRAG: Discover how GraphRAG goes beyond traditional RAG by enabling agents to reason over knowledge graphs, delivering more factual, explainable, and context-rich responses.
• MCP for Neo4j in Manufacturing: Learn how Model Context Protocol (MCP) is driving Neo4j adoption in Japanese manufacturing, transforming data workflows and powering real-world applications at scale.

Topics/Speakers:
🎤 Tomaz Bratanic, GenAI Research @ Neo4j
Talk Topic: Agentic GraphRAG
Talk Description: RAG works—until complex context kicks in. GraphRAG upgrades it by weaving in knowledge graphs to structure retrieval, boost relevance, and enable explainable, precise generation. Learn how it fuses symbolic reasoning with neural search to power next-gen, context-aware agentic AI.
Speaker Bio: TOMAŽ BRATANIC has extensive experience with graphs, machine learning, and generative AI. He has written an in-depth book about using graph algorithms and GraphRAG.

🎤Takeshi Kuwada, Data Engineer, National Institute for Materials Science (NIMS), former Director of iASYS Technology Solutions japan (neo4j distributor)
Talk Topic: ​MCP: A Game-Changer for Neo4j Adoption in Japanese Manufacturing
Talk Description: ​Through my many years of experience introducing Neo4j to the R&D departments of Japanese manufacturing companies, I've seen both its incredible potential and the challenges of its on-site use. A major hurdle, especially for non-IT engineers, has been mastering Cypher queries and understanding complex data structures, which has often prevented them from fully leveraging the value of a graph database.
​This talk will focus on MCP (Model Context Protocol), an innovative technology that makes connecting Neo4j with LLMs incredibly simple. I will share my insights from hands-on experience with it. MCP allows for interaction with Neo4j using natural language, making graph database use dramatically more accessible.
​I will also share stories from my experience working with Japanese manufacturing clients, discussing both the strengths and weaknesses of Neo4j. Through demonstrations, I will explain the breakthrough that MCP represents. I look forward to discussing the potential and practicality of this technology with fellow Neo4j enthusiasts.
Speaker Bio: ​Since 2011, I have led the adoption of Neo4j in Japanese manufacturing as the Director of iASYS Technology Solutions. My work has primarily focused on data utilization projects in R&D departments, including automotive and materials development. Currently, I am involved in a materials and data open-data project at the National Institute for Materials Science (NIMS), where I also promote data science in materials experimentation.

🎤Changhwan Lee, Senior Architect, AI & Data Platform Team, Creationline Inc.
Talk Topic: ​Automating Knowledge Graphs with Large Language Models (LLMs) — Transforming Tech Blogs into GraphRAGs for Natural Language Exploration
Talk Description: Imagine if users could simply push their source data, and a “pretty good graph” would be automatically generated and ready to use.
If we could achieve this, companies could visualize and reuse the vast amount of knowledge and data hidden within their organizations — bringing us closer to a world where anyone can easily find and use information.
With that vision, we at Creationline conducted a proof of concept (PoC) on automating knowledge graph generation using large language models (LLMs).
Using our own tech blog as the data source, we explored:
- How LLMs can assist in data modeling
- How well they can structure semantic models and extract relationships
- How to separate roles between LLMs and traditional code processing
- The pathway to GraphRAG and building a Neo4j Vector Index
- Hands-on experimentation with GraphRAG search using natural language
While full automation remains a challenge, this project revealed significant potential and deep insights into how LLMs can revolutionize the way we build and explore knowledge graphs.

Whether you’re building GenAI apps, exploring graph-powered reasoning, or working in industries like manufacturing, this session will give you fresh insights into the practical and transformative power of graphs.

Join us for an evening of knowledge-sharing, networking, and inspiration with the Tokyo graph and GenAI community!