book

Agentic GraphRAG

by Anthony Alcaraz, Sam Julien

August 2026

Intermediate to advanced

388 pages

11h 41m

English

O'Reilly Media, Inc.

Read now

Unlock full access

Includes

Quizzes

Preface
Who Should Read This BookWhy We Wrote This BookNavigating This BookConventions Used in This BookUsing Code ExamplesO’Reilly Online LearningHow to Contact UsAcknowledgments
I. Foundations of Graph-Based Agentic Systems
1. The Crisis of Enterprise Agentic AI
Defining Agentic AIWhat Is an Agent, Anyway?Classifying Agentic Systems: The Workflow-Agent SpectrumThe Agent Constraint TriangleAgent ArchitectureThe Limitations of Vector-Based RetrievalA Brief Overview of Vector RetrievalWhere Vector RAG Succeeds and FailsGraphRAG: The Evolution of Contextual Retrieval SystemsA Brief Overview of GraphRAG ArchitectureWhere GraphRAG Succeeds and StrugglesHybrid GraphRAGLearned Graph Reasoning ModelsSummary
2. Agentic Graph Architecture Foundations
From Strings to ThingsThe Dual-Graph ArchitectureThe Vertical Knowledge GraphThe Horizontal Workflow GraphWhere the Two Graphs MeetA Note on Context GraphsThe Eight Pillars of Agentic Graph ArchitectureSetup: Building the Digital Twin of a Tech Stack (Part II)Action: From Firefighting to Foresight (Part III)Production: Creating a Self-Healing System (Part IV)The Autonomous DevOps AgentThe Initial StateThe Transformation ArcA Real-World ParallelWhat the Completed Agent Looks LikeSummary
II. Setup: Building the Digital Twin of a Tech Stack
3. Graph-Based Knowledge Modeling for Agentic Systems
Knowledge Graph FoundationsComponents of GraphRAG SystemsBuilding Versus Buying a GraphRAG SystemGraph Data ModelsTypes of Graph Data ModelsEvaluating Graph ModelsGraph Design Principles for Agentic SystemsThe Three-Graph Architecture for Agent KnowledgeSchema Design PatternsHomoiconic Knowledge RepresentationIntegrating with Existing SystemsKnowledge Organization and Ontology FundamentalsCreating a Unified Semantic FoundationEntity Resolution: The Foundation of Agent KnowledgeBuilding the Knowledge GraphExtraction Approaches for Heterogeneous SourcesAutomating Knowledge Graph Construction with Multi-Agent SystemsEntity Resolution and Linking Across GraphsKnowledge Representation in Practice: DevOps AgentContext Engineering for DevOps AutomationThe DevOps Knowledge Representation ChallengeDesigning the DevOps OntologyPopulating the Knowledge Graph: Building the Digital TwinHomoiconic Knowledge Representation for Agent AdaptabilityQuerying the Knowledge Graph: Making the Digital Twin UsefulEntity Resolution for Infrastructure IdentityMeasuring Context Quality: Ensuring the Digital Twin Stays AccurateSummary
4. Agentic Graph Memory Systems
Architecting Graph Memory SystemsThe Problem: Why Your Current Memory Approach FailsThe Solution: Graph MemoryRepresenting Memory as GraphsUnderstanding the Architecture EvolutionEssential Graph ComponentsTemporal Awareness: Making Memory EvolveImplementation: Building Your First Graph Memory SystemFour Essential Memory OperationsConsolidation: From Experience to KnowledgeIndexing: Organizing for SpeedUpdating: Handling Change GracefullyRetrieval: Finding What MattersProduction Patterns and ArchitecturesThe Letta (MemGPT) Approach: Hierarchical MemoryThe A-Mem Pattern: Evolving Knowledge NetworksThe Graphiti Pattern: Real-Time PerformanceComparing Production SystemsCogneeMem0ZepChoosing Your ApproachProduction-Ready Features and OperationsReasoning and RecommendationsUse Case OptimizationsTraining Your Agents to Use MemoryThe MEM1 Revelation: Let Them Learn, Not Follow RulesFrom Static Rules to Learned StrategiesThe Three Phases of TrainingDynamic Memory Without Dynamic SchemasPractical Training StrategiesEvolving Your Memory SystemMoving Beyond Traditional TrainingAdding Memory to the DevOps AgentDesigning the DevOps Memory SchemaImplementing Memory for Operational ContextTemporal Queries for Root Cause AnalysisConsolidation: Learning Patterns from IncidentsTraining Memory Behavior with Reinforcement LearningProduction ConsiderationsFuture-Proofing Your ImplementationNeurosymbolic IntegrationDistributed ScalingContinual LearningSummary
III. Action: From Firefighting to Foresight
5. Reasoning and Planning
Foundations: Why Graph-Based Reasoning Matters for Agent DevelopmentThe Two Graph Types: Horizontal and VerticalOntological Grounding: Keeping Your Agent in RealityRetrieval MechanismsThe Distracting Effect: Why Your Agents Get MisledPathRAG: Structural Protection Against DistractionsR3-RAG: Teaching Your Agents When and How to RetrieveGraphRAG Pattern TaxonomyReasoning and GenerationInterleaved Thinking: Real-Time Adaptive ReasoningStructured Generation: The Keystone of Reliable CommunicationPipeline ArchitecturesSequential Pipeline: Linear ExecutionLoop Pipeline: Iterative RefinementTree Pipeline: Parallel ExplorationProduction Implementation PatternsHybrid ArchitecturesEvent-Driven Orchestration: The Production Scaling LayerPlanning and CoordinationWhy Planning Deserves Architectural SeparationHierarchical PlanningConstraint-Guided PlanningDynamic DAG ConstructionEmpirical Evidence: Graph Planning at Repository ScaleFrom Architecture to Daily Practice: Task Graphs at Developer ScaleThe Multi-Agent Debate: Reconciling Parallel ExecutionDevOps Diagnostic Agent: Pulling It All TogetherBuilding the Horizontal Workflow GraphIntegrating the Knowledge Graph for Grounded HypothesesConstructing the Investigation DAGA Production Incident WalkthroughWhy the Architecture MattersSummary
6. Agentic Graph-Based Tool Orchestration
MCP: The Protocol That Changes the GameSecuring the Foundation: MCP AuthenticationAgents as Tools: A Powerful PatternThe Prompt Bloat Crisis: When More Tools Mean Less IntelligenceFrom Listing to Finding: The RAG WayThe Gateway PatternSkills: The Judgment LayerChoosing the Right Primitive: CLIs, MCPs, and SkillsThe Knowledge Graph Necessity: Tools That Understand Each OtherTools as Graph NodesTool Dependencies and Type MatchingSecuring Data Flow with Information Flow ControlEvolving Tool EcosystemsOntology-Driven Evolution: When Tools Define ThemselvesThe Trust Problem: When Tools Game the SystemMetatooling: When Tools Create ToolsOrchestration at ScaleThe Intelligent Orchestrator: Modernizing Classical SaaSHierarchical Orchestration: When Scale Demands StructureFunctional Clustering: Resilience Through RedundancyThe Enterprise AI OS Layer: Infrastructure for the Synthetic WorkforceLearning and Advanced PatternsThe Memory Layer: Learning from Every InteractionAdvanced Patterns: The Nested Execution ChallengePerformance at Scale: Making It RealContext Governance: The Missing Layer in Tool OrchestrationTool Orchestration in Practice: The DevOps AgentRegistering Tools in the Knowledge GraphStructured Output for API IntegrationThe Latency Investigation, ContinuedSummary

IV. Production: Creating a Self-Healing System
7. Self-Evolution and Evaluation
The Foundation of Self-AwarenessThe Execution Graph: The Agent’s AutobiographyThe Multi-Layered Evaluation Framework: A Cognitive AutopsyA Taxonomy for Self-EvolutionTargeted Self-ImprovementChoosing the Right Intervention: Prompt, Fine-Tune, or Constrain?Semantic Backpropagation: Coherent Evolution Across the GraphA Suite of Self-Improvement FrameworksSafe Evolution in ProductionThe Full Evolutionary LoopThe RPO and Graduated Validation ProtocolPersistent Self-Learning Without Entropy CollapseThe DevOps Agent: Predictive Outage Detection in PracticeThe Scenario: A Library Update Triggers a PredictionGenerating the Execution GraphThe Cognitive AutopsySemantic Backpropagation to the Causal Attribution NodeThe Improvement and Its EffectInference-Time Knowledge Augmentation: Learning Without RetrainingSkills as Self-Improving Graph ObjectsCommon PitfallsSummary
8. Optimization
Selective IntelligenceThe Case for Model SpecializationRouting StrategiesMeasuring Cost-Performance Trade-OffsData Governance and Access ControlSubgraph-Level Access ControlPII and Data RetentionThe Execution Graph as a Compliance ArtifactProduction Systems and MaintenanceSchema EvolutionNode and Edge Lifecycle ManagementIncremental UpdatesDeployment PatternsHardware AccelerationGPU-Accelerated Graph OperationsInference Acceleration for Specialist Model FleetsLatency BudgetsThe DevOps Agent: From Prototype to ProductionApplying Selective IntelligenceHardware Acceleration for Dependency AnalysisGovernance for Infrastructure DataThe Completed AgentCommon PitfallsSummaryConclusionWhere the Field Is HeadingChoosing the Right Architecture for Your SystemBuilding What Comes Next
Index
About the Authors

Content preview from Agentic GraphRAG

About the Authors

Anthony Alcaraz is a leading AI/ML strategist at AWS, known for his expertise in decision science, large language models, knowledge graphs, and RAG. With a growing audience of over 38,000 on Medium, Anthony regularly shares his insights into AI and knowledge systems. Recently, he lectured at Oxford on the integration of LLMs with graph theory to solve modern business problems.

Sam Julien is a product leader specializing in AI developer platforms. He’s currently director of product at CopilotKit, focused on transforming agent-user interactions into self-learning agents. Previously, he was the lead product manager for WRITER’s IT and developer platform. Sam has spoken at conferences around the world, including The AI Conference and the AI Engineer World’s Fair.

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.

Read now

Unlock full access

More than 5,000 organizations count on O’Reilly

O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.

Julian F.

Head of Cybersecurity

I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.

Addison B.

Field Engineer

I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.

Amir M.

Data Platform Tech Lead

I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.

Mark W.

Embedded Software Engineer

Publisher Resources

ISBN: 9798341623163Errata Page

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills

Agentic GraphRAG

by Anthony Alcaraz, Sam Julien

About the Authors

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.