Skip to content

Commit f1e78a1

Browse files
authored
Merge pull request #53414 from MicrosoftDocs/NEW-fabriciq-fundamentals
New fabriciq fundamentals
2 parents 782112c + ee8dc17 commit f1e78a1

21 files changed

Lines changed: 407 additions & 0 deletions
Lines changed: 16 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,16 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.wwl.understand-fabric-iq-fundamentals.introduction
3+
title: Introduction
4+
metadata:
5+
title: Introduction
6+
description: "Introduction to Microsoft Fabric IQ and how ontologies define business vocabulary for data agents and Graph in Microsoft Fabric."
7+
ms.date: 02/09/2026
8+
author: theresa-i
9+
ms.author: theresai
10+
ms.topic: unit
11+
ai-usage: ai-generated
12+
azureSandbox: false
13+
labModal: false
14+
durationInMinutes: 3
15+
content: |
16+
[!include[](includes/1-introduction.md)]
Lines changed: 14 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,14 @@
1+
### YamlMime: ModuleUnit
2+
uid: learn.wwl.understand-fabric-iq-fundamentals.get-started-with-fabric-iq
3+
title: Get started with Fabric IQ
4+
metadata:
5+
title: Get Started with Fabric IQ
6+
description: Learn what Microsoft Fabric IQ is, where to access it, and how ontologies define business vocabulary and bind to data sources.
7+
ms.date: 02/09/2026
8+
author: theresa-i
9+
ms.author: theresai
10+
ms.topic: unit
11+
ai-usage: ai-generated
12+
durationInMinutes: 8
13+
content: |
14+
[!include[](includes/2-get-started-with-fabric-iq.md)]
Lines changed: 14 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,14 @@
1+
### YamlMime: ModuleUnit
2+
uid: learn.wwl.understand-fabric-iq-fundamentals.explore-fabric-iq-components
3+
title: Explore Microsoft Fabric IQ components
4+
metadata:
5+
title: Explore Microsoft Fabric IQ Components
6+
description: Learn about the components that make up the Fabric IQ ecosystem, including ontology items, data agents, Graph in Microsoft Fabric, and Power BI semantic models.
7+
ms.date: 02/09/2026
8+
author: theresa-i
9+
ms.author: theresai
10+
ms.topic: unit
11+
ai-usage: ai-generated
12+
durationInMinutes: 8
13+
content: |
14+
[!include[](includes/3-explore-fabric-iq-components.md)]
Lines changed: 14 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,14 @@
1+
### YamlMime: ModuleUnit
2+
uid: learn.wwl.understand-fabric-iq-fundamentals.understand-ontology-modeling-paradigm
3+
title: Understand the ontology modeling paradigm
4+
metadata:
5+
title: Understand the Ontology Modeling Paradigm
6+
description: Learn how ontology modeling differs from traditional analytical data modeling by starting with business concepts and relationships rather than designing tables for specific use cases.
7+
ms.date: 02/09/2026
8+
author: theresa-i
9+
ms.author: theresai
10+
ms.topic: unit
11+
ai-usage: ai-generated
12+
durationInMinutes: 10
13+
content: |
14+
[!include[](includes/4-understand-ontology-modeling-paradigm.md)]
Lines changed: 75 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,75 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.wwl.understand-fabric-iq-fundamentals.knowledge-check
3+
title: Module assessment
4+
metadata:
5+
hidden_question_numbers: ["A0348922_22","A0348922_34","A0348922_51","A0348922_55","A0348922_59","A0348922_67","A0348922_83","A0348922_87"]
6+
title: Module assessment
7+
module_assessment: true
8+
ai_generated_module_assessment: true
9+
description: "Check your knowledge of Microsoft Fabric IQ fundamentals"
10+
ms.date: 02/02/2026
11+
author: theresa-i
12+
ms.author: theresai
13+
ms.topic: unit
14+
ai-usage: ai-generated
15+
azureSandbox: false
16+
labModal: false
17+
durationInMinutes: 5
18+
quiz:
19+
title: "Check your knowledge"
20+
questions:
21+
- content: "What are the three core concepts that make up an ontology in Fabric IQ?"
22+
choices:
23+
- content: "Tables, columns, and foreign keys"
24+
isCorrect: false
25+
explanation: "Incorrect. These are relational database concepts. Ontology uses things (entity types), facts (properties), and connections (relationships)."
26+
- content: "Things (entity types), facts (properties), and connections (relationships)"
27+
isCorrect: true
28+
explanation: "Correct. An ontology is made up of the things in your environment (entity types), their facts (properties), and the ways they connect (relationships)."
29+
- content: "Lakehouses, warehouses, and semantic models"
30+
isCorrect: false
31+
explanation: "Incorrect. These are data storage and modeling options in Fabric. Ontology uses entity types, properties, and relationships."
32+
- content: "What is the primary role of a Fabric data agent?"
33+
choices:
34+
- content: "To move data from lakehouses to warehouses"
35+
isCorrect: false
36+
explanation: "Incorrect. Data agents don't move data. They enable natural language queries across data sources."
37+
- content: "To visualize relationships between entities in a graph"
38+
isCorrect: false
39+
explanation: "Incorrect. This is the role of Graph in Microsoft Fabric, not the data agent."
40+
- content: "To process natural language questions and generate queries grounded in ontology definitions"
41+
isCorrect: true
42+
explanation: "Correct. Data agents use Azure OpenAI to process natural language questions, identify relevant data sources, and generate appropriate queries (SQL, DAX, or KQL) based on ontology vocabulary."
43+
- content: "How does data binding in ontology modeling differ from traditional ETL processes?"
44+
choices:
45+
- content: "Data binding copies data into a centralized warehouse, while ETL leaves data in place"
46+
isCorrect: false
47+
explanation: "Incorrect. This is reversed. ETL copies data into warehouses, while data binding creates a semantic layer without copying data."
48+
- content: "Data binding creates a semantic layer that references data in place, while ETL copies and transforms data into a warehouse"
49+
isCorrect: true
50+
explanation: "Correct. Data binding connects ontology definitions to existing data sources in OneLake without copying data, unlike ETL which extracts, transforms, and loads data into a centralized warehouse."
51+
- content: "Data binding and ETL are the same process with different names"
52+
isCorrect: false
53+
explanation: "Incorrect. Data binding creates a semantic layer without moving data, while ETL physically copies and transforms data into a warehouse."
54+
- content: "Which Fabric IQ component uses GQL (Graph Query Language) for querying?"
55+
choices:
56+
- content: "Ontology items"
57+
isCorrect: false
58+
explanation: "Incorrect. While ontology items integrate with Graph, they use NL2Ontology queries and federated query execution."
59+
- content: "Data agents"
60+
isCorrect: false
61+
explanation: "Incorrect. Data agents generate SQL, DAX, or KQL queries depending on the data source."
62+
- content: "Graph in Microsoft Fabric"
63+
isCorrect: true
64+
explanation: "Correct. Graph in Microsoft Fabric uses GQL (Graph Query Language), an international standard for graph queries, to traverse relationships and analyze connected data."
65+
- content: "What makes entity types different from traditional database tables?"
66+
choices:
67+
- content: "Entity types are tied to specific databases and schemas"
68+
isCorrect: false
69+
explanation: "Incorrect. This describes database tables. Entity types exist independently of any particular data source."
70+
- content: "Entity types are reusable logical models that exist independently of any storage system"
71+
isCorrect: true
72+
explanation: "Correct. Entity types are reusable logical models elevated above any single table, allowing the same concept to be bound to multiple data sources while maintaining consistent business meaning."
73+
- content: "Entity types can only contain static data, not time-series data"
74+
isCorrect: false
75+
explanation: "Incorrect. Entity types support both static data bindings and time-series data bindings from eventhouse streams."
Lines changed: 14 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,14 @@
1+
### YamlMime: ModuleUnit
2+
uid: learn.wwl.understand-fabric-iq-fundamentals.summary
3+
title: Summary
4+
metadata:
5+
title: Summary
6+
description: "Summary of Microsoft Fabric IQ fundamentals and key concepts"
7+
ms.date: 02/09/2026
8+
author: theresa-i
9+
ms.author: theresai
10+
ms.topic: unit
11+
ai-usage: ai-generated
12+
durationInMinutes: 3
13+
content: |
14+
[!include[](includes/6-summary.md)]
Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,7 @@
1+
Imagine you're a data analyst at Lamna Healthcare, responsible for helping clinical operations teams understand patient care patterns across your facility. Patient records sit in lakehouse tables while vital signs stream continuously from ICU monitoring equipment into an eventhouse. When hospital administrators ask questions like "Which patients in the ICU have elevated vital signs?" or "How many beds are occupied on the surgical floor?", you need to manually join lakehouse tables with eventhouse streams, translate business terms into technical column names, and write complex queries.
2+
3+
Business users can't explore the data themselves—they depend on you to write queries each time they have a question. By the time you deliver answers, clinical conditions may have already changed.
4+
5+
Fabric IQ solves this challenge by letting you define business vocabulary in an ontology, then bind those concepts to your data sources in OneLake. You define concepts such as Patient, Department, and Room with their properties and relationships, creating a semantic layer. Business users can then ask questions in natural language through data agents or visually explore relationships through Graph in Microsoft Fabric—exploring the data themselves without needing you to write queries.
6+
7+
In this module, you'll discover what Fabric IQ is and how it works. You'll explore the components that work together—ontology items, data agents, Graph in Microsoft Fabric, and Power BI semantic models—and learn when to use each one. You'll also see how ontology modeling shifts your approach from use-case-driven thinking to concept-driven thinking, fundamentally changing how teams collaborate around data.
Lines changed: 84 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,84 @@
1+
Fabric IQ is a workload in Microsoft Fabric for creating ontologies that define your business vocabulary. It sits alongside other Fabric workloads like Data Engineering, Data Factory, Data Science, Data Warehouse, Real-Time Intelligence, and Power BI. Within the IQ workload, you create **ontology items**—Fabric artifacts that contain your ontology definitions and data bindings.".
2+
3+
An ontology is a shared vocabulary of your business. It's made up of the things in your environment (represented as entity types), their facts (represented as properties of entity types), and the ways they connect (represented as relationships).
4+
5+
You can also think of an ontology like a business context layer, containing:
6+
7+
- A catalog of concepts (like Hospital, Patient, Department) with their properties and relationships
8+
- Data bindings to your lakehouse tables and eventhouse streams
9+
- A graphical representation that links related concepts for navigation and analysis
10+
- A query surface that lets you ask questions about concepts (not just tables), supporting federated queries across sources
11+
12+
Instead of requiring data experts to translate business questions into SQL queries, you model data using business concepts that everyone understands.
13+
14+
The ontology provides a single definition of each concept that can be used by data agents and Graph in Microsoft Fabric for querying and visualization.
15+
16+
## Where Fabric IQ fits in the data platform
17+
18+
Fabric IQ builds on Microsoft Fabric's unified data platform. Here's how it connects to other capabilities:
19+
20+
**Ingest and store:** Fabric IQ works with data you already have in lakehouse tables and eventhouse streams. It doesn't move or duplicate your data—it creates a semantic layer that references your existing data sources.
21+
22+
**Model and represent semantics:** The ontology item offers modeling capabilities by defining entity types, properties, and relationship types. You can generate an ontology structure from existing Power BI semantic models, or create your own from scratch. Then bind these definitions to your data and explore them in a navigable graph that builds automatically.
23+
24+
**Analyze and visualize:** The ontology item integrates with Graph in Microsoft Fabric to provide a visual graph and query experience based on your business concepts. You can use the ontology to ground data agents with business context.
25+
26+
## Access Fabric IQ in your workspace
27+
28+
You create ontology items the same way you create other Fabric items:
29+
30+
1. Navigate to your Fabric workspace
31+
2. Select **+ New item**
32+
3. Search for and select **Ontology (preview)**
33+
4. Enter a name for your ontology (use numbers, letters, and underscores—no spaces or dashes)
34+
5. Select **Create**
35+
36+
:::image type="content" source="../media/new-ontology-item.png" alt-text="Screenshot showing the New item dialog with Ontology option selected." lightbox="../media/new-ontology-item.png":::
37+
38+
The ontology opens when it's ready. You'll see two main areas: the configuration canvas where you define entity types and relationships, and the preview experience where you explore your data.
39+
40+
> [!IMPORTANT]
41+
> Your Fabric administrator needs to enable certain tenant settings before you can create ontology items. For more information, see [Ontology (preview) required tenant settings](/fabric/iq/ontology/overview-tenant-settings?azure-portal=true).
42+
43+
## Explore the ontology interface
44+
45+
The ontology item has two primary views:
46+
47+
**Configuration canvas:** This is where you build your ontology. You create entity types (like Hospital, Department, Room, Patient), define properties on those entities (like FirstName, DateOfBirth, AdmissionDate), and establish relationship types between entities (like "contains" or "assigned to").
48+
49+
:::image type="content" source="../media/configuration-canvas.png" alt-text="Screenshot showing the ontology configuration canvas with entity types, properties, and relationships." lightbox="../media/configuration-canvas.png":::
50+
51+
**Preview experience:** This view shows your instantiated ontology. You can see entity instances (specific rooms, departments, and patients), explore relationships in a graph visualization, and query your data using business language instead of SQL. The preview experience integrates with Graph in Microsoft Fabric to provide rich visual exploration.
52+
53+
:::image type="content" source="../media/preview-ontology.png" alt-text="Screenshot showing the ontology preview experience with graph visualization and entity instances." lightbox="../media/preview-ontology.png":::
54+
55+
## Understand the build-bind-query workflow
56+
57+
Creating an ontology in Fabric IQ follows three main steps:
58+
59+
**Build:** Define your business vocabulary by creating entity types, properties, and relationship types. You're modeling the concepts that matter to your business. For example, in a healthcare scenario, this means defining what a Patient is, what properties describe a patient (like name, date of birth, admission date), and how patients relate to other entities (like rooms and departments).
60+
61+
**Bind:** Connect your ontology definitions to data sources. This includes lakehouse tables for static data (like patient records and room assignments) and eventhouse streams for time-series data (like vital signs monitoring). Data binding maps your source data columns to ontology properties.
62+
63+
**Query:** Once your ontology is bound to data, you can query it using business concepts instead of database tables. Use Graph in Microsoft Fabric to visualize relationships and traverse connections. Use Query Builder to filter and explore entity instances without writing SQL. Or connect AI agents that can answer natural language questions using your business vocabulary.
64+
65+
This workflow separates business meaning from physical data structures.
66+
67+
## How Fabric IQ connects to OneLake
68+
69+
Fabric IQ doesn't move or duplicate your data. Instead, it creates a semantic layer that references existing data sources:
70+
71+
- **Lakehouse tables** contain static data (patient records, hospital information, room assignments)
72+
- **Eventhouse streams** contain time-series data (continuous vital signs from medical monitoring equipment)
73+
74+
When you query your ontology, Fabric IQ automatically sends your queries to the most efficient system to get results quickly. For graph traversals, it uses GQL for Graph in Microsoft Fabric. For time-series queries, it uses KQL for Eventhouse. This federated query capability means you can ask business-level questions that span multiple data sources without knowing the technical details of where data lives.
75+
76+
## Two paths to create an ontology
77+
78+
Fabric IQ offers two approaches for creating ontologies:
79+
80+
**Generate from Power BI semantic model:** If you already have a well-structured Power BI semantic model, you can automatically generate an initial ontology structure from it. Fabric IQ creates entity types matching your tables, properties matching your columns, and relationship types following your model relationships. You then refine this generated ontology by renaming entity types, verifying keys and bindings, and enhancing it with additional data sources like time-series eventhouse streams.
81+
82+
**Build from OneLake data:** If you don't have a semantic model, or want full control over ontology design, you can build directly from lakehouse and eventhouse data. You manually create entity types, define properties, and establish relationships. This approach gives you complete control over how you model your business vocabulary.
83+
84+
Both paths lead to the same result: a complete ontology that defines your business concepts.

0 commit comments

Comments
 (0)