Microsoft Ignite: What are the key insights around Microsoft Fabric?

Introducing Microsoft's latest news around Microsoft Fabric

At this year’s Microsoft Ignite, Microsoft Fabric took center stage with some exciting updates. A major highlight was the introduction of IQ—a brand-new suite of products within the Fabric ecosystem. In addition to this, new capabilities such as IDENTITY columns, enhanced Mirroring options, and Data Agents have been added, expanding functionality and flexibility for users.

Fabric IQ

Microsoft Fabric IQ is a semantic layer that elevates Fabric from a unified data platform to a unified intelligence platform. Instead of focusing only on storage and computer. IQ delivers a unified business context by defining entities, relationships, and rules - and seamlessly connects this meaning to live data in OneLake. The result is a governed, self-service knowledge graph of your organisation that people and AI can use to reason and act. 

At Ignite 2025, Microsoft positioned IQ as one pillar of a broader "Unified Context Layer" alongside Work IQ (signals and memory from Microsoft 365) and Foundry IQ (agent runtime and governance). Together, these layers are designed to make agentic AI useful and auditable at enterprise scale. 

Why Fabric IQ matters:

There is consistent definitions across tools, your team can define entities like "customer", "contract" or "gross revenue" once and reuse this everywhere. This works across Power BI, notebooks, agents and apps, cutting reworks and eliminating semantic drift. 

• Faster solution delivery: Teams don't have to reconstruct business meaning for every dashboard or AI experience; they build against a shared ontology and get immediate alignment. 
• Trust & governance: Ontology items are versioned and validated; constraints and rules improve data quality, lineage clarifies impact, and policies propagate consistently.
• Cross domain reasoning: Native Graph in Microsoft Fabric supports multi-hop queries to explain outcomes across connected data, (e.g., Order → Shipment → Temperature Sensor → Cold chain breach)
• AI that speaks your language:  IQ provides structured grounding for copilots and agents so answers reflect the enterprise vocabulary rather than raw columns, reducing hallucinations and increasing adoption. 

The core pieces of IQ and how they work

• Ontology (preview):A no-code solution that enables domain experts to model entity types, relationships, properties, rules and constraints - and then link these concepts to OneLake sources such as tables, streams, and existing Power BI semantic models. Once published, the ontology automatically generates a navigable graph across your data

• Graph in Microsoft Fabric (preview): Provides native graph storage and compute for nodes, edges, and traversals - including pathfinding, dependency analysis, and advanced graph algorithms. This enables connected insights directly on OneLake without the need for additional ETL.
• Fabric data agents (preview): “Virtual analysts” that answer business questions grounded in your ontology and policies, rather than scraping adhoc tables.  
  

Together, these components turn Fabric into a semantic system where questions, reasoning, and actions share the same governed context. 

Where IQ fits in the Fabric ecosystem

IQ builds on what customers are already using in Fabric:

• Ingest & store: IQ unifies data from lakehouse tables, eventhouse streams, and existing Power BI semantic models in OneLake.  

• Analyse & visualise: Ontology integrates with Graph for visual exploration; it also grounds Power BI models and domain aware agents.  

• Operate & govern: Ontologies are versioned and validated; monitoring is available via Fabric’s admin and capacity tools.  
  

What new features will Fabric users actually feel?

• Less rework, more reuse: BI and data science teams reuse trusted semantic definitions from Power BI across notebooks, pipelines, and agents.  

• Connected insights out of the box: With Graph, analysts can traverse relationships (customers → devices → transactions) to detect fraud or supply chain risks without stitching together bespoke joins.  

• Smarter co-pilots & agents: Fabric co-pilots and data agents now answer with business concepts (e.g., “late shipments affecting VIP customers”) instead of table names.  

• Governed autonomy: When paired with Foundry IQ, operational agents can monitor live signals, weigh trade offs (cost/speed/risk), and take actions under policy.  

Typical use cases (with quick wins)

• Customer 360 with causality: Link CRM orders, logistics telemetry, and service tickets; use Graph traversals to explain churn drivers and trigger proactive retention actions.  

• Regulatory reporting with one truth: Encode statutory definitions in the ontology so every report and AI workflow uses the same audited meaning.  

• Intelligent operations: Define entities (Asset, Job, Incident) and policies; let agents escalate or schedule interventions based on semantic rules.  

Getting started - Practical steps 

• Identify 10–15 core business entities (Customer, Order, Asset, Contract) and the relationships that drive outcomes. Start small and high value.  
• Create an Ontology (preview) in Fabric and bind it to your existing OneLake sources and Power BI semantic models. Validate with real scenarios (queries, reports, agent prompts).  
• Pilot Graph queries for two use cases (e.g., cold chain breach root cause; multi-account fraud patterns). Measure time to insight and error reduction vs. table only approaches.  
• Introduce a Fabric data agent to answer stakeholder Q&A using your ontology and policies. Keep humans in the loop before automating actions.  

Note: As of today, IQ and its ontology/graph components are in preview. Plan pilots accordingly, and track roadmap updates via the Fabric blog. 

How IQ relates to Work IQ and Foundry IQ

Think of Work IQ as your collaboration context, Fabric IQ as your business data semantics, and Foundry IQ as the agent runtime and governance layer.  

For years, the industry has centralised where data lives but left what it means fragmented. Fabric IQ is Microsoft’s answer: shared business meaning bound to live data, delivered as a first class workload. If you’re already standardising on Fabric and Power BI, IQ is the next step.  

IDENTIY Columns

Microsoft Fabric now supports IDENTITY columns, allowing users to automatically generate unique numeric values for each new row in their tables. This feature streamlines data ingestion and simplifies scenarios where surrogate keys are required, making it easier to manage large datasets and maintain data integrity without manual intervention. 

New options for Mirroring

Fabric has expanded its Mirroring feature, adding support for a wider range of source systems and offering more granular control over replication settings. Organisations can now mirror data from both on-premises and cloud sources with improved performance, enabling near real-time analytics and reducing complexity in hybrid data environments. 

Data Agents 

The introduction of Data Agents allows organisations to securely connect and synchronise data from various on-premises sources to Microsoft Fabric without exposing sensitive network endpoints. Data Agents offer enhanced security, simplified configuration, and support for a broader array of data connectors, making integration across diverse environments more seamless than ever.