Make Your AI Agents Predictable With Graph Engineering - Zachary Blumenfeld, Neo4j

By ai.engineer

Categories: AI, Tools

Summary

Graph engineering—designing agent workflows as explicit graphs with nodes and edges—enables predictable AI agents by constraining execution paths and adding a portable knowledge layer between data sources and any runtime, solving the unpredictability problem plaguing current agent frameworks.

Key Takeaways

  1. Three graph types power agent engineering: execution graphs (workflow routers), context graphs (decision traces and domain data connections), and procedural graphs (skill steps, validation, escalation logic)—focus on the latter two for portability and efficiency gains.
  2. Graph engineering delivers three core benefits: predictability (constraining agent paths for accuracy), governance (controlling, auditing, and iterating agent behavior), and portability (storing graphs in a knowledge layer that works with any runtime like LangGraph or Claude).
  3. Neo4j's approach creates a knowledge layer between systems of record (Databricks, BigQuery, Snowflake, documentation) and agent runtimes using standard protocols—enabling one graph engineering process to power any agent framework without rebuilding for each.
  4. Context graphs store multiple memory types (short-term conversation history, long-term patterns, domain context) that agents reference—making it possible to audit decisions, understand agent reasoning, and improve performance iteratively with explicit data connections.
  5. Graph engineering evolved from prompt engineering and context engineering—the term is new but mechanisms (state machines, workflow graphs, knowledge graphs) are established; what's changing is applying these rigorously to make agents handle bigger, more complex work predictably.

Related topics

Transcript Excerpt

[music] Uh we're going to talk about graph engineering today. How many people have heard about that term graph engineering? Okay. So so a fair amount of the room and then just so I understand how many people here are familiar with Neo forj and have used Neo forj. Okay. So we got a few people. So that's good for uh just letting me know who we're talking to today. Um so just some history on some of the verbiage that's being used. uh there was a tweet that went out uh by Peter Steinberger I think uh earlier this year in the summer uh that talked about graph engineering and it was almost like is this sort of a new discipline that's happening as this evolution from prompt engineering and context engineering uh and there's a lot of hype around this um and the name for this is new but the mechani…

More from ai.engineer