This Week in Neo4j: AI in Production, Memory, GraphRAG, Architecture and more

Senior Developer Marketing Manager

March 20, 2026

4 min read

Welcome to This Week in Neo4j, your fix for news from the world of graph databases!

This edition brings together some of the most grounded perspectives on building agents, memory and GraphRAG implementations that hold up in the real world.
I also wanted to take the opportunity to highlight the Essential GraphRAG book that you can get for free. Finally, we take a closer look at the AI memory gap and hear how GraphRAG, neuro-symbolic AI and graph-based memory address core limitations of LLMs.

NODES AI, our global graph-and-AI event, is taking place April 15, and the full agenda with themes like Context Graph, GraphRAG, Agents and AI in Production – register now! You can still catch the final Road to NODES workshop events to fulfil your need for something hands-on!

Happy Graphing,

Alexander Erdl

COMING UP!

Livestream: Neo4j Live: NODES AI Agenda with Aura Agents on April 1
Conferences: Find us at Kubecon Europe, Amsterdam on March 22-24, All Things AI, Durham on March 23-24 & MCP Dev Summit, New York on April 2-3
Meetup: Meet us in Mountain View, US on March 23, Menlo Park, US on March 24, Amsterdam, NL on March 25 & San Francisco, US on March 30
All Neo4j Events: Webinars and More
GraphSummit Series: Transform Your Enterprise with Graph and GenAI – Next Stop: Paris on April 14

FEATURED COMMUNITY MEMBER: Vincent Koc

Vincent Koc is an engineering technologist with extensive experience in data-driven disciplines. He holds a fellowship at the Institute of Managers and Leaders Australia, where he serves as a thought leader and mentor to the next generation of data professionals.

Connect with him on LinkedIn.

His session at NODES AI is “Agent Interaction Graphs: Evaluating Multi-Agent Systems with Graph-Based Reasoning”, where he will model agent executions as an interaction graph in Neo4j and use this knowledge graph, attach evaluations and run graph queries to pinpoint critical issues, recurring failure points and bottlenecks based on deep contextual relevancy from a graph.

AI IN PRODUCTION: Useful AI Agent Case Studies: What Actually Works in Production

This article by Jesús Barrasa explores real-world AI agent case studies, focusing on where agents actually deliver value in production – especially in workflows that require memory, multi-step reasoning and tool use. It breaks down how context engineering and structured data (such as knowledge graphs) enable agents to move beyond demos, highlighting lessons learned from deploying agents in messy, real-world enterprise systems.

AI MEMORY: Bridging the Gap: Why We Built Perseus to Solve the AI Memory Problem

Charles Borderie looks at the AI memory gap and argues that better memory architectures are essential for trustworthy AI: LLM-based systems struggle with reliable reasoning due to a lack of structured, persistent memory. Lettria introduces its approach of combining knowledge graphs, improved benchmarks and its Perseus SDK to turn unstructured data into structured, queryable memory—shifting AI from “guessing” to grounded, explainable reasoning

GRAPHRAG: Essential GraphRAG

Essential GraphRAG by Tomaž Bratanic and Oskar Hane is a hands-on guide to building production-ready GraphRAG systems by combining knowledge graphs with vector retrieval, enabling richer context, better accuracy and improved traceability for LLM applications. It walks developers through constructing graphs from unstructured data, generating Cypher queries from natural language and designing hybrid, agentic RAG pipelines that go beyond simple semantic search.

The ebook is available as a free download!

AI Architecture: Graph Chat: Neo4j’s Philip Rathle on Neuro-symbolic AI and Infinigraph

In this GraphGeeks video, Amy Hodler and Neo4j CTO Philip Rathle unpack why the current AI wave is fundamentally a graph moment, not just a vector one. They explore how GraphRAG, neuro-symbolic AI and graph-based memory address core limitations of LLMs – bringing explainability, reasoning and scalable context (via innovations like Infinigraph) to enterprise AI systems.

CONTINUOUS LEARNING

GraphAcademy: Create intelligent AI applications in “Developing with Neo4j MCP Tools“
Learn on Your Schedule: Go deeper into graph technology on Neo4j’s On-Demand webinar library
Workshops: Join our virtual classrooms workshops from Fundamentals to GenAI
New Webinar: Connect the Dots With Context Graphs – Americas, Europe, Middle East & Africa, Asia Pacific