Microsoft Fabric is an all-in-one data analytics platform that brings together data engineering, real-time analytics, and business intelligence in a unified environment. Built on a lake-centric architecture, Fabric enables organizations to store and manage data efficiently while leveraging AI-powered capabilities for deeper insights. Designed to empower business users, it simplifies data access, transformation, and analysis, facilitating data-driven decision-making across the organization.

As businesses increasingly adopt data-driven decision-making, Microsoft is expanding Fabric’s capabilities with new AI-driven analytics, enhanced data governance, and deeper integrations across the platform. Recent announcements at the Fabric Conference (FabCon) introduced several powerful features in preview, aimed at streamlining data engineering, improving real-time analytics, and making insights more accessible to business users. In the following sections, we’ll explore these innovations and their potential impact on organizations looking to maximize the value of their data.

First: Our top themes and takeaways

The Fabric Conference provided insight into Microsoft’s focus and investment priorities for Fabric and its workloads. AI remains a central priority, becoming deeply integrated into many facets of Fabric, from Copilot and chat interfaces to AI- and LLM-powered functions embedded into data engineering, embedding AI into your data transformations. This year we saw a huge focus on data governance, between significant investment in advancing Microsoft Purview’s capabilities for governance, AI security, and integrating into the OneLake catalog.

This indicates an important concept: AI use is only safe if it’s secure. Organizations have the ability to use, understand, and act on their data like never before, and part of that is empowering more people to leverage the right tools. However, they need to maintain existing security and access constraints to make sure people don’t see the data they shouldn’t see. By implementing Purview Security Posture Management for AI, embedding quality constraints in materialized views, introducing OneLake security, and more, it’s clear that Microsoft continues to make data security, quality, and governance a priority.

With our expertise in data analytics and governance, we recognize that AI-driven transformation isn’t just about automation – it's about enabling organizations to harness their data with confidence. Microsoft Fabric is equipping enterprises with the tools to integrate AI seamlessly into their data ecosystems while maintaining the highest standards of security and governance. As AI capabilities continue to evolve, businesses that embrace these innovations with a strong data foundation will be positioned to lead in an increasingly AI-powered world.

Copilot & AI

One of the most exciting announcements from FabCon is that Fabric Copilot and AI capabilities will now be accessible to all paid SKUs in Microsoft Fabric, not just F64 capacities and above. This is a huge step toward lowering the barrier to entry for organizations to leverage AI. This change is expected to take effect by the end of April. Other AI updates include:

Data agents

AI Skills has a new name: data agents. Data agents allow you to ask your data questions and get answers in real time. They use data from OneLake, which means they use your organization’s data to answer your questions, while still maintaining roles and permissions set in the OneLake catalog. 

Copilot for Real-Time Intelligence 

Microsoft Fabric’s Real-Time Intelligence enables businesses to process and analyze streaming data instantly with real-time dashboards, AI-powered anomaly detection, and automated alerts. It provides proactive issue resolution – a capability that was previously limited by batch-processing delays. Speed up development of Real-Time dashboards by using Copilot to:

• Create visualizations and aggregate data using narrative prompts with Copilots.
• Generate the KQL queries associated with visuals to help ensure traceability and reliability and enable further customization.

Copilot maintains context from previous interactions, so you can build onto Copilot-generated visuals for more advanced analysis.

AI and operational agents

AI agents in Fabric, now in preview, can reason, plan, and respond across the platform, automating data flows, ingestion, transformation, analysis, and visualization. By integrating with tools like Data Factory, Real-Time Intelligence, and Power BI, these agents enhance efficiency and deliver on the promise of “speed to insights.”

• Analysts can build and deploy agents via a chat interface to resolve issues and take action based on business rules that you define. 
• AI agents analyze how the data across your platform connects, conduct root cause analyses, detect anomalies, and more. 
• Agents use dynamic route optimization to minimize the time it takes to complete actions. You can also monitor agent performance with reported metrics. 

Power BI: What’s new?

Fabric is introducing several enhancements to Power BI that will significantly improve real-time data interaction, application management, and AI-assisted reporting. These updates include more integration with live data sources for instant insights, improved governance, and deployment options for managing enterprise-scale reports. These innovations aim to make data visualization more intuitive, accessible, and powerful for business users.

Power BI write-back

• Integrate Power BI’s user data functions to build and automate custom calculations for more adaptable planning and analysis
•  By facilitating connections to external APIs, translytical task flows allow users to write-back to Fabric databases directly from the Power BI interface – built with automated version tracking and control.

Organizational apps for pro licenses

• Power BI now supports the creation of multiple organizational apps per workspace, offering greater flexibility in content management and distribution.
• Audiences for organizational apps were also introduced, allowing you to further segment content within a Power BI workspace.

Copilot grounding sneak peek 

• Provide Copilot access to corporate jargon, calendars, and reports for enhanced context in its responses and report-creation assistance. 
• Report developers will be able to hide columns, rows, and tables from Copilot in semantic models to refine its responses and understanding. Use grounding to get quality and consistent responses for critical business topics.

Data storage with OneLake

OneLake is a unified, lake-centric storage solution that provides a scalable repository for all organizational data. New enhancements, including OneLake security for advanced access control, expanded mirroring capabilities for seamless data integration, and improved governance with the OneLake catalog, will further strengthen data protection, accessibility, and compliance across your data estate.

OneLake security 

OneLake security, now in preview, provides a centralized lake approach to managing data access, including row-level and column-level controls. You set security once, and it’s automatically enforced across Spark notebooks, SQL endpoints, Excel Online, and Direct Lake semantic models in Power BI. 

This means no more replicating your RLS logic in every individual tool. Instead, you create roles and permissions in OneLake, and Fabric makes sure those rules are applied consistently, no matter how users interact with the data. Whether it’s through a notebook, a report, or a query editor, the same security policies are enforced.

• Centralized role management: You can now create roles directly in OneLake, define which folders or tables each role can access, and assign users accordingly. This aligns with the principle of managing security where the data lives. 
• Row- and column-level security: This one is huge. RLS and column-level security (CLS) are now a part of OneLake. You can define granular permissions, such as limiting users to only see data from their department or hiding sensitive columns like salary or account numbers.
• Automatic enforcement across engines: Whether someone is querying through Spark, exploring with SQL, or viewing a Power BI report powered by Direct Lake, the same security logic applies. This isn’t just a convenience – it’s a governance win. 
• New role management interface: Microsoft is also introducing a new user interface to make defining and managing roles easier. From what’s been shared so far, the UI consolidates access and membership management in one place and introduces T-SQL-style filtering for row-level rules.

Mirroring

New features are coming to Mirroring in Fabric, expanding your ability to bring data into Fabric without complex ETL, extra storage costs, and new CI/CD integrations.

• Mirroring for CI/CD: Use ALM deployment pipelines and Git integration to simplify the deployment process and support smooth updates to mirrored databases.
• Open mirroring: You can now allow any application to directly write change data into a Fabric mirrored database via the public open mirroring APIs.
• New external sources
  • Azure Cosmos DB
  • Azure Databricks
  • Azure PostgreSQL
  • Azure SQL Managed Instance
  • Snowflake

Data governance in OneLake catalog

The OneLake catalog now offers a preview of a centralized data governance experience, implementing several features from Purview’s Unified Catalog in OneLake. Empower data owners to prioritize governance with a more detailed view of their assets.

New features in OneLake catalog

• Add detailed metadata to your data assets (database tables, Power BI reports, dataflows, etc.), including descriptions and use cases.
• View an asset’s lineage to understand its relationship to other assets across your Fabric data estate.
• Control roles and permissions for individual assets or logical domains created in the OneLake catalog.

Purview integration sneak peek

• Sensitivity labels and data classifications from Purview will persist into the OneLake catalog.
• View Purview data quality scores in the OneLake catalog.
• Purview governance domains, data products, and associated roles and permissions will integrate with OneLake security.

Governance dashboard

Built for chief data officers and data managers, the new out-of-the-box governance dashboard provides aggregated insights on their created items in OneLake and suggests recommended actions to improve governance.

Data Factory 

Enhancements to Data Factory announced at FabCon this year focus on bridging some of the gaps between Azure Data Factory and Fabric’s iteration of Data Factory – boosting data integration flexibility and scalability. New features include:

• Variable library: Now in preview, the variable library is a repository that allows you to use the same custom variables in multiple pipelines in a workspace, supporting scalability and consistency across different data lifecycle stages. The variable library is currently only supported by Data Factory pipelines.
• New pipelines activity: With the new pipeline activity in Fabric Data Factory, you can now run and monitor ADF and Synapse pipelines from Fabric, allowing you to integrate these pipelines with Fabric activities like dataflows (Gen2) and notebooks.

Data Engineering

Data Engineering improvements were a central focus this year, many of them geared toward the developer experience and the integration of AI. A highlight is the availability of T-SQL scripting in notebooks. Other new features include:

• Engineering seamless migration: The new “migrate” feature allows you to convert Synapse Data Warehouse assets into Fabric Data Warehouse assets in a few clicks. Organizations using Synapse Data Warehouses will no longer have to use Mirroring and will be able to retire their Synapse warehouses if they want to.
• AI functions: Use AI functions to apply LLM-powered transformations to your OneLake data in data science and engineering workflows and notebooks. Some of the functions available in preview now can perform text translations, variable classifications, response generations, similarity analyses, and more.
• Materialized views: You can now build materialized views in your Lakehouse using T-SQL syntax and built-in data quality constraints, logs, diagnostics, and Power BI quality report. Like any other Delta table, you can query these views across all engines.

Databases

The home in Fabric for OLTP workloads, the SQL database reflects the functionality of your standard SQL databases but is optimized for analytics by replicating the data into OneLake in near real time.

• SQL database for Fabric: Natively utilize other Fabric workloads on your database – Engineering, Power BI, Spark analytics, data agents (AI skill) – without the need for mirroring from external sources. Databases also enable automatic index creation with automatic tuning.

Governance with Microsoft Purview

Microsoft Purview is a unified data management platform for security, compliance, governance, and auditing. It helps classify, protect, and monitor data while supporting regulatory compliance. Here we focus on Purview’s Data Governance capabilities, crucial for maintaining data quality, lineage, and accessibility. Microsoft continues to enhance features and further integrate Purview with Fabric, making it an increasingly mature and unified enterprise data management solution.

Data Observability 

If you’re familiar with Purview, you might know of the out-of-the-box data health report available in the Unified Catalog. The report provides a bird’s-eye view of the quality across your data estate. The new Data Observability feature now in public preview, offers similar insights but focuses on the conceptual relationships within your data. If you have your catalog designed around reporting, data observability allows you to easily identify if there are quality issues that could affect reporting, what those issues are, and where they are coming from.

• View your data estate and its quality under a logical lens, versus just physical. 
• Drill down to see lineage at the column level or by glossary term.
• Identify data quality by dimension (accuracy, conformity, completeness, etc.) at the estate, domain, product, asset, and column levels.

Copilot for Microsoft Purview 

Initially focused on security, Copilot in Purview is expanding to Data Governance. It will offer guided steps for Unified Catalog configuration, auto-generate summaries for data products and assets, and streamline documentation. Its effectiveness depends on high-quality user-added metadata, making catalog accuracy and thoroughness essential for meaningful insights.

Data Quality enhancements 

Data Quality in Purview allows you to easily validate, test, and score the quality of your data. Enhancements coming to Data Quality include:

• Enabling row-level monitoring for data quality tests, allowing data stewards to pinpoint specific records that need attention. 
• Build custom reports with your Purview metadata in Power BI with the new Purview semantic model. Use it to visualize data quality across your organization how you want to see it. 
• Critical Data Elements (CDEs) and Glossary terms will be able to be mapped to the asset and column level to enhance quality testing and traceability.

Wrapping up

 This year’s FabCon showcased advancements that are set to redefine the landscape of data analytics and governance. The introduction of new features such as Fabric Copilot, AI agents, and enhanced data governance tools underscores Microsoft's commitment to empowering organizations with cutting-edge technology to streamline data processes, enhance real-time analytics, and build robust data security and governance.

As AI continues to integrate deeply into business operations, the importance of high-quality, well-managed data cannot be overstated. Fabric's lake-centric architecture and AI-powered capabilities provide a solid foundation for organizations to harness their data effectively. The new features announced at FabCon, including expanded access to AI tools, improved data observability, and advanced security measures, are pivotal in enabling businesses to make data-driven decisions with confidence.

The emphasis on data governance, particularly through Microsoft Purview, highlights the critical role of security and compliance in the AI-driven era. With Fabric’s comprehensive tools for data classification, protection, and monitoring, organizations can leverage AI while maintaining the highest standards of data integrity and privacy.

In conclusion, the advancements presented at FabCon are set to drive significant business transformation. Organizations that embrace these innovations will be well-positioned to lead in an increasingly AI-powered world, unlocking new opportunities for growth and efficiency. Microsoft Fabric continues to be at the forefront of enabling smarter, data-driven decision-making, paving the way for a future where data and AI work seamlessly together to drive success.

^{This has been prepared for information purposes and general guidance only and does not constitute legal or professional advice. You should not act upon the information contained in this publication without obtaining specific professional advice. No representation or warranty (express or implied) is made as to the accuracy or completeness of the information contained in this publication, and CohnReznick, its partners, employees and agents accept no liability, and disclaim all responsibility, for the consequences of you or anyone else acting, or refraining to act, in reliance on the information contained in this publication or for any decision based on it.}