Neo4j Live: Graph Engineering - Making Agent Workflows Repeatable

Your agent takes a different path every run, and you can’t explain why. Graph engineering fixes that by designing agent work as an explicit graph: nodes do the work, edges decide what runs next.
We cover:
- The three graphs behind the term: execution, context, and procedure graphs
- How graphs give you predictable paths, an audit trail, and portability across agent frameworks
- How to turn a skill into a procedure graph (research result: pass rate up from 53% to 67% on 27 tasks)
- A live demo: a loan-approval agent that runs faster on a smaller model

Guest: Zach Blumenfeld ( https://www.linkedin.com/in/zachblumenfeld/ ) Blog:

#neo4j #graphdatabase #agenticai #knowledgelayer #knowledgegraph #graphrag #graphengineering