June 2025 Summaries

Christoffer Bergman introduces an innovative approach to using Agentic AI with Java and Neo4j, focusing on enhancing the GraphRAG framework by integrating AI agents. While traditional GraphRAG combines vector search and graph traversal to provide context to language models like LLMs, the agentic model empowers AI with tools to navigate and retrieve data independently. The blog includes a practical example where data from Stack Overflow for Teams is imported into Neo4j, enabling AI to answer user queries based on this structured information. Bergman explains how he developed a Java library to facilitate the creation and registration of AI agents with OpenAI, using reflection and annotations to define tool methods, thus enabling the agent to perform tasks like retrieving posts and comments from the graph. Although the library is presented as a learning tool rather than a production-ready solution, it demonstrates the potential of integrating AI agents with graph databases to improve data interaction and retrieval.

Generative AI (GenAI) is rapidly expanding across industries, with significant investments projected to reach $202 billion by 2028, highlighting its growing importance within the AI market. However, deploying GenAI on platforms like Amazon Web Services (AWS) can present challenges, such as factual inaccuracies, traceability issues, and intellectual property concerns, which can be detrimental to businesses. Graph databases offer a solution by providing complex data hierarchies and maintaining contextual relationships, which enhance AI accuracy and understanding. The Graph Retrieval Augmented Generation (GraphRAG) approach, combining Large Language Models with graph databases, addresses these challenges by enabling real-time data access and improved traceability, benefiting industries like financial services and manufacturing. Partnerships between AWS and Neo4j, including integrations with various AWS tools, facilitate the seamless incorporation of GraphRAG into existing environments, optimizing GenAI applications for better reliability, transparency, and outcomes.

The blog explores the utilization of the Model Context Protocol (MCP) and Neo4j Data Modeling MCP server to enhance graph data modeling workflows. It highlights how MCP facilitates the integration of standardized tooling for applications, enabling efficient graph data model generation, validation, and visualization for Neo4j-backed applications. The process involves defining entities and relationships, understanding use cases, and refining data models to boost query performance. The blog demonstrates using tools like Claude Desktop and the Neo4j Cypher MCP server for creating and querying graph models, emphasizing the importance of iterative data modeling and visualization through Mermaid diagrams. Additionally, it discusses the benefits of combining the Data Modeling MCP server with other MCP servers for comprehensive development workflows and underscores the ease of refactoring data models in Neo4j's schemaless database environment.

Neo4j certification offers developers a valuable opportunity to enhance their skills in graph database technology and prove their capability in building complex systems. Unlike passive learning methods, the Neo4j certification emphasizes hands-on experience through writing Cypher queries and modeling real-world data, ensuring participants demonstrate a deep understanding rather than mere recall. This certification is particularly relevant as knowledge graphs become integral to AI and emerging technologies, providing an edge in fields such as retrieval-augmented generation and context-aware applications. Additionally, the Neo4j community, accessible via platforms like Discord, supports developers in collaborative learning and problem-solving. Earning this certification is not just about gaining a credential but also about signaling to peers and employers a commitment to advancing in the field, enhancing one's developer toolkit, and preparing for an AI-driven future.

This Week in Neo4j highlights a range of developments and activities within the Neo4j community, emphasizing the ongoing Certification Month with community activities designed to enhance Neo4j expertise. New tools like Neo4j Desktop 2.0 and AI-powered dashboards in Aura are introduced, offering developers advanced capabilities for data visualization and management. The blog post also features the GraphRAG sessions from the AI Engineer World’s Fair, which discuss building knowledge graphs, multi-hop information retrieval, and optimizing AI agent memory. Additionally, the post features community member Christophe Willemsen's insights on anonymization in the LLM era, a narrative by Saswati Rao on creating personal knowledge graphs, and offers resources for continuous learning through GraphAcademy and webinars.

Retrieval-augmented generation (RAG) applications, designed to enhance large language model (LLM) responses by incorporating external data, often struggle with multi-hop reasoning tasks that require connecting disparate pieces of information. Knowledge graphs, which organize data as interconnected nodes and relationships, offer a solution to this challenge through a technique known as GraphRAG. This method improves the accuracy, context, and explainability of LLM-generated responses by integrating RAG with knowledge graphs, allowing for more effective navigation of complex queries involving multiple topics. GraphRAG enhances retrieval by broadening context, prioritizing relevant data, and providing a structured framework that supports reasoning across tools and data sources. The Neo4j LLM Knowledge Graph Builder facilitates the automatic creation of knowledge graphs from unstructured data, streamlining the process of turning raw content into structured insights that power retrieval-augmented applications. This approach not only improves the explainability and accuracy of LLM outputs but also reduces the workload at query time, making GraphRAG a robust solution for enterprise use cases requiring comprehensive and traceable AI insights.

Cypher is a powerful and expressive query language for Neo4j, designed to simplify querying graph databases by using a visual syntax that resembles sketching relationships and patterns. Unlike traditional SQL, Cypher emphasizes the importance of relationships within data, making it particularly useful for scenarios where data interconnections are crucial. Key elements of Cypher include nodes, which represent entities like people or products, and relationships that define connections between these nodes. Cypher employs keywords such as MATCH, WHERE, and RETURN to structure queries, allowing users to specify patterns, filter results, and define output from the database. Advanced operations like CREATE, MERGE, and SET enable data manipulation and ensure flexible data handling in Neo4j. As an intuitive tool, Cypher mirrors real-world relationships, making it easy to read and understand complex queries by breaking them down into sequential steps. Engaging with Cypher involves adopting a new way of thinking about data, encouraging users to practice through experimentation and leverage community resources for deeper learning.

The article introduces AI-powered dashboard creation in the Aura Console, a tool for users of Neo4j AuraDB. The new Dashboards feature allows users to create interactive graph-powered dashboards in minutes, using an AI-powered dashboard generator and the GenAI copilot in the chart editor. Users can visualize and explore their Neo4j data, work across tools and collaborate with ease, and share their work within the Aura console. The article provides a step-by-step guide on how to create dashboards from Spotify data, including importing data into AuraDB, generating a graph model, creating a dashboard, and using Graph Analytics to enrich the graph. The feature offers more functionality for customizing visualizations, enhancing interactivity, and sharing dashboards.

Neo4j is a high-speed graph database that offers two professional certifications, both of which are free, self-paced, and globally recognized. The Neo4j Certified Professional certification is for developers and architects working with Cypher, graph modeling, and transactional applications, while the Graph Data Science Certification is for data scientists, machine learning engineers, and anyone using graphs for analytics, predictions, and insights. To prepare for these exams, it's recommended to take courses such as Importing CSV Data Into Neo4j and Intermediate Cypher Queries, and to have a solid understanding of Cypher query coding and graph algorithms. The exams are non-proctored open-book exams that can be retaken every 24 hours, with the Certified Professional exam consisting of 80 questions within 60 minutes and the Graph Data Science Certification exam consisting of 40 questions within 60 minutes. To pass these exams, it's essential to have thorough notes, including literal copying of Cypher query code from both theoretical and sandbox modules, as well as having access to documentation during the exam. The certifications are not only a great goal to achieve but also demonstrate that you know something about Neo4j and are ready to apply it in real life.

Neo4j Desktop 2.0 offers a unified local developer experience with a new design, improved navigation, and integrated tools for managing databases, querying, visualization, and data modeling, as well as enterprise features like enhanced security and remote connection capabilities. This version is suitable for developers who prefer to keep data on their local machine or need restricted internet access, and it provides consistent behavior with the Aura console's fully managed cloud experience. Desktop 2.0 also includes updated tools like Query and Explore, with modernized Cypher editor, richer table visualization, and improved saved Cypher experience. The new version is available for Mac, Windows, and Linux, and users can download it now to start using its features.

Neo4j is making June a Certification Month with various community activities to help developers get certified and showcase their skills. The company is also highlighting the integration of Neo4j Graph Analytics directly in Snowflake, agentic AI workflows using LangGraph, and the importance of knowledge graphs for reliable enterprise-grade AI. Meanwhile, NODES 2025 is approaching on November 6 with a call for papers closing soon. Various events, including livestreams, conferences, meetups, and webinars, are scheduled throughout June to engage with the community and promote Neo4j's offerings.

Join Neo4j's User Research Panel and contribute to shaping the future of their products. As a panel member, you'll participate in various research studies such as interviews, usability tests, surveys, and more to help better understand user needs. By joining, you'll have the opportunity to share your voice directly with product development teams, influence how Neo4j evolves and improves, get early insights into upcoming features, receive paid compensation for your input, and contribute to real product decisions. The panel is open to individuals of all backgrounds and experience levels, and you can opt out or request data removal at any time.

RAG has improved AI and LLMs by using context-aware generation, but it's not enough on its own. A case study for a large real estate company showed that RAG was unable to answer complex questions that required reasoning across documents, entities, and relationships. To bridge the gap, a reasoning layer was added using Neo4j, a graph database built for representing relationships. This structure allowed for relational thinking, enabling answers such as who did what, where, and when. The outcome was an 80-percent reduction in time-to-answer for complex questions, 70 percent of internal queries handled without human escalation, 90-percent accuracy in multi-hop responses, more trust in AI-generated answers, and the system being adopted by three other departments. Embeddings aren't enough; graph databases add structure to what RAG can only guess. LangChain + Neo4j + Gemini = a production-grade reasoning system, and grounded prompts win. Smarter AI isn't just bigger models — it's smarter retrieval and reasoning. Adding a reasoning layer makes your AI smarter, more reliable, and more trustable, especially for businesses with domain-specific knowledge that spans documents and systems.

GraphRAG is a game-changing retrieval technology that builds upon vector search by incorporating reasoning and relationships. It starts with vector search to narrow down relevant documents or nodes, but then uses large language models (LLMs) to generate Cypher queries based on the ontology of a knowledge graph. This enables GraphRAG to traverse multi-hop relationships, enrich responses with structured context, execute precise queries over a graph, and generate answers that are grounded, explainable, and connected. The author has reworked a movie chatbot using Neo4j and Google Vertex AI to incorporate GraphRAG, resulting in a system that understands the plot of movies, pulls in vector-similar movies, identifies directors, genres, and timeframes, and responds with contextual, multi-hop, personalized suggestions. Deploying the chatbot to Google Cloud Run provides serverless and scalable deployment capabilities.

Neo4j Graph Analytics for Snowflake is a new offering that brings highly optimized graph algorithms to the Snowflake AI Data Cloud, allowing users to generate deeper insights from their data using familiar SQL without requiring specialized graph expertise. This zero-ETL solution uses elastic compute and provides enterprise-ready graph analytics, improving decision-making by uncovering hidden patterns and relationships in complex data. With this offering, business decision-makers and data teams can focus on outcomes rather than overhead, and it prioritizes speed, delivering insights 2x faster than open-source alternatives with parallelized in-memory processing of graph algorithms. The solution also provides cost-efficient pricing, using serverless architecture and a pay-as-you-go model to simplify cost management, allowing users to align their infrastructure spending with their business needs.

Symfony is being integrated with Neo4j, a graph database, for efficient debugging and performance optimization. To achieve this, Symfony's Profiler and Logger are utilized to monitor and optimize Neo4j database interactions. The setup involves installing the required packages using Composer, configuring connections between Symfony and Neo4j, and enabling profiling and logging in the application. This integration enhances development efficiency by removing performance guesswork, providing valuable insights into query execution times, debug issues, and database performance issues.