Why Context Graphs Replace Siloed Data

▶ Watch (0:17)

Enterprise data sits in Slack threads, customer tickets, and separate systems. Agents making decisions without this context produce generic answers. Gartner added context graphs to the AI hype cycle. Foundation Capital called it a $3 trillion startup opportunity. Knowledge graphs aggregate information and create relationships. They combine with LLMs for reasoning and context.

Three Memory Layers for Agent Workflows

▶ Watch (6:33)

Short-term memory holds current pipeline state. Long-term memory organizes business processes and entities across interactions. Reasoning traces capture why decisions were made. All three persist in a knowledge graph. Neo4j’s open-source agent memory package combines them into a context graph structure.

Demo: Lenny’s Podcast Memory Graph

▶ Watch (9:58)

The Lenny memory podcast demo loads all episodes into a knowledge graph. It provides tools for accessing memory via Neo4j agent memory APIs. The graph extracts locations mentioned across episodes and builds a map. Because in graph format, queries return a holistic view of the entire dataset, not just similar chunks.

Demo: Financial Services Context Graph

▶ Watch (13:00)

The financial services demo connects a support ticket, CRM, and internal business system via 10 MCP tools. It uses OpenAI embeddings and Neo4j context graph. The UI asks about Jessica Norris. The graph retrieves her bank account, margin trades, and a previous rejection. The model recommends denying the loan with risk factors and fraud detection patterns.

Notable Quotes

We are using AI coding tools or or maybe they’re using us. Stephen Chin · ▶ Watch (0:55)

Gartner has now officially made context graphs as part of the AI hype cycle. Stephen Chin · ▶ Watch (2:36)

we can do kung fu, we can shoot bullets, we can do all this amazing stuff with the right memory structure in place. Stephen Chin · ▶ Watch (6:05)

the AI model recommends not giving her the loan, but it gives us the reasons, the risk factors, it gives us previous decisions which should influence this. Stephen Chin · ▶ Watch (15:17)

Key Takeaways

  • Context graphs combine short-term, long-term, and reasoning memory in one graph.
  • Neo4j’s open-source agent memory package enables explainable AI decisions.
  • Gartner recognized context graphs as a key AI technology.