book

Hands-On Microsoft Fabric AI Architecture

Name: Hands-On Microsoft Fabric AI Architecture
Author: Brian Bønk
ISBN: 9798341673274

by Brian Bønk

April 2027

Intermediate

350 pages

7h 25m

English

O'Reilly Media, Inc.

Read now

Unlock full access

Brief Table of Contents (Not Yet Final)
1. From Aristotle to Agents: A History of Ontology
The Philosophical Roots of OntologyAristotle and the Quest to Classify RealityPorphyry’s Tree: The First Knowledge HierarchyFrom Leibniz to Frege: The Dream of Formal LanguageWhy Philosophy Still Matters for Data ArchitectsThe Computer Science Turn: Knowledge RepresentationThe Birth of Artificial Intelligence and the Knowledge BottleneckExpert Systems: MYCIN, DENDRAL, and the Frame ProblemFormal Ontologies in Computer Science: From Cyc to Biomedical StandardsThe Knowledge Representation Hypothesis and Its ImplicationsThe Semantic Web: A Grand Vision and Its LessonsTim Berners-Lee and the Dream of Linked DataThe Technology Stack: RDF, RDFS, and OWLEnterprise Ontologies: The Complexity TrapLessons Learned: What the Semantic Web Got RightThe Modern Renaissance: Ontologies Meet Generative AIThe Rise of Large Language Models and the Hallucination ProblemRetrieval-Augmented Generation and the Structure GapKnowledge Graphs in Industry: Google, Meta, and MicrosoftThe Convergence: Ontologies as Grounding Truth for Agentic AISummary
2. Introduction to Microsoft Fabric IQ
From Data Platform to Intelligence PlatformThe Three Eras of Enterprise DataThe Missing Layer: Semantic MeaningThe Intelligence Platform ParadigmThe five pillars of Fabric IQPillar 1: The Ontology, The Shared Business ModelPillar 2: Semantic Models, The Analytical DefinitionsPillar 3: Graph in Microsoft Fabric, The Native Graph EnginePillar 4: Data Agents, Conversational Q&APillar 5: Operations Agents, Autonomous ActionThe Role of OneLake: Logical Binding Without Data MovementOneLake as the Unified FoundationThe Binding LayerBinding Targets: Lakehouses, Warehouses, and EventhousesHow the Five Pillars Work TogetherAn End-to-End ScenarioBefore and After: The TransformationSummary
3. The Art of Ontology Design
The Ontology Design Process: From Whiteboard to PlatformThe Four Phases of Ontology DesignPhase 1: Domain Discovery and ScopingPhase 2: Defining Entity TypesPhase 3: Modeling RelationshipsPhase 4: Constraints and Business RulesThe Airport Operations Domain: A Complete Worked ExampleThe SkyLink International Domain NarrativeEntity Types and PropertiesRelationshipsFrom Design to Implementation: Preparing for BindingSummary
4. Binding and Materialization
Understanding the Binding ArchitectureWhat binding means in Fabric IQThe Logical-to-Physical Mapping PrincipleThe Zero-Copy PrincipleSupported Binding TargetsBinding to LakehousesThe Lakehouse as a Binding TargetColumn-to-Property MappingKey Column Binding and IdentityPractical Lakehouse Binding: A Step-by-Step Walk-ThroughBinding to WarehousesThe Warehouse as a Binding TargetDimensional Model ConsiderationsBinding to Warehouse ViewsCross-Schema Binding StrategiesBinding to EventhousesThe Eventhouse as a Binding TargetStreaming Entity PatternsTime-Windowed Entity ResolutionIoT and Operational Use CasesEntity Resolution and Cross-Source BindingThe Entity Resolution ChallengeIdentity Strategies for Cross-Source BindingCross-Source Property AggregationBootstrapping: From Power BI Semantic Models to OntologyThe Bootstrapping ConceptWhat Gets BootstrappedThe Bootstrapping WorkflowBootstrapping in PracticeBootstrapping Limitations and Best PracticesGraph MaterializationWhat Graph Materialization MeansThe Property Graph ModelThe Materialization PipelineRefresh and Update CyclesQuerying the Materialized GraphValidating and Governing BindingsWhy Binding Validation MattersBinding Validation PatternsHandling Schema DriftGovernance Roles and ResponsibilitiesPractical Example: Building an Airport Operations Domain BindingThe ScenarioThe Binding PlanExecuting the BindingVerifying the Materialized GraphThe Complete ArchitectureCommon Binding Challenges and SolutionsThe “Everything Is a String” ProblemThe Missing Foreign Key ProblemThe Granularity Mismatch ProblemThe Slowly Changing Dimension ProblemThe Scale ChallengeSummary
5. Graph Semantics in Fabric
Why Graph? The Limits of Relational ThinkingThe JOIN Wall: When SQL Becomes IntractableThe DAX Ceiling: When Measures Cannot Navigate RelationshipsGraph Theory Fundamentals for the Data ProfessionalNodes: The Entities in Your GraphEdges: The Relationships That ConnectDirected vs. Undirected RelationshipsThe Property Graph ModelGraph in Microsoft Fabric: Architecture and CapabilitiesHow the Materialized Graph Is StoredThe Graph Explorer in the Fabric PortalThe Relationship Between Ontology, Bindings, and the Queryable GraphIntroduction to GQL (Graph Query Language)Pattern Matching: The Core of GQLPattern Matching and Path NavigationConditional Filtering on Path ElementsDiscovering Hidden ConnectionsFlight Disruption Cascade AnalysisCrew Scheduling Conflict DetectionRoute Connectivity and Passenger Flow NetworksCross-Domain ReasoningBreaking Down Organizational Silos with GraphLinking Flight Performance to Ground OperationsConnecting Weather Events with Passenger Impact and Ground AccountabilityAdvanced Graph Patterns and AggregationsSubgraph ExtractionTemporal Patterns in Graph DataIntegrating Graph Results with the Data EcosystemFrom GQL Results to Spark DataFramesCombining Graph Insights with Power BIFeeding Graph Patterns to Data AgentsPerformance Considerations and Best PracticesQuery Optimization PrinciplesGraph Semantics in the Broader Fabric IQ VisionSummary
6. RAG vs. GraphRAG
Understanding RAG: The FoundationWhat RAG Is and Why It EmergedThe Standard RAG PipelineWhen RAG Excels: Unstructured Document IntelligenceThe Limits of Vector SearchThe Semantic Similarity AssumptionThe Counting ProblemThe Relationship ProblemThe Multi-Hop ProblemThe Aggregation ProblemThe Consistency ProblemEnter GraphRAG: Retrieval Through StructureFrom Text Chunks to Knowledge StructuresThe GraphRAG PipelineHow the Ontology Guides RetrievalWalking the Graph: Structural Context AssemblyStandard RAG vs. GraphRAG: A Detailed ComparisonSide-by-Side: The Same Question, Two ApproachesStructural Questions AnsweredMulti-Hop Questions AnsweredAggregation Questions AnsweredThe Verdict: Complementary, Not CompetingGraphRAG Patterns in Fabric IQPattern 1: Entity-Centric RetrievalPattern 2: Relationship-Aware Context AssemblyPattern 3: Subgraph ExtractionPattern 4: Hybrid RAG (Vector + Graph)Context Assembly and Token EfficiencyThe Context Window ChallengeStructured Context vs. Unstructured ChunksHow GraphRAG Reduces HallucinationMeasuring Retrieval QualityImplementing GraphRAG in Fabric IQPrerequisite: A Well-Designed OntologyConfiguring Graph-Aware RetrievalNL2GQL: Natural Language to Graph TraversalCombining GQL Results with LLM GenerationArchitecture and Performance ConsiderationsWhen to Use Each ApproachScaling GraphRAG for Enterprise WorkloadsCaching and Pre-Computation StrategiesGraphRAG in the Broader Fabric IQ VisionSummaryWhat You Have Learned
About the Author

Overview

Microsoft Fabric IQ represents a paradigm shift in how enterprises manage data, moving from a unified data platform to a unified intelligence platform. At the center of this shift is the ontology, a semantic intelligence layer that connects raw data to business meaning. By linking entities, relationships, and logic within OneLake, Fabric IQ addresses semantic drift and enables a governed, shared understanding of enterprise data.

In this practical playbook, author Brian Bønk shows how this new architecture changes analytics and AI in Microsoft Fabric. You'll move from the foundations of ontology to hands-on implementation, learning how to design an enterprise ontology, bind it to OneLake artifacts, and operationalize graph semantics. The book explains how Fabric data agents use this semantic layer as grounding, allowing them to answer complex business questions with accuracy, governance, and explainability instead of probability.

Understand the Fabric IQ architecture and its core components
Design an enterprise ontology using entity types, properties, and relationships
Bind ontological definitions to OneLake data assets
Apply graph semantics and GraphRAG for multihop reasoning
Build Fabric data agents grounded in governed semantic meaning

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: 0642572308322Errata Page

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills