NYC Fabric Community- Kickoff Event - Featured Speakers: Kim Manis / Marco Russo
Kim Manis: FabCon Atlanta Recap: The Latest Microsoft Fabric Announcements, Explained Just a few weeks after FabCon Atlanta, this NYC Fabric User Group session brings the conference highlights straight to the community. We’ll recap the biggest Microsoft Fabric announcements spanning data engineering, analytics, AI experiences, governance, and platform capabilities and discuss how they impact real‑world workloads. Whether you attended FabCon or are catching up after the fact, this session focuses on what you need to know: what’s new, what’s changed, and how teams should think about adopting these capabilities in the months ahead. Join us for a fast‑paced, accessible rundown designed to spark discussion and practical next steps. Marco Russo: Introduction to DAX User Defined Functions (UDFs) Description: A user-defined function is a DAX formula with parameters that can return values or tables. Parameters can be passed by value or by expression. In this session, you learn how to define and consume user-defined functions in your semantic model, how to share them in different models, and how to use function libraries created by the community.
Reintroduce Tenant/Capacity Switch to Control "Users can create Plan items" Post-GA
During the Preview phase of Fabric Plan items, administrators had access to a dedicated tenant/capacity setting: "Users can create Plan items". With General Availability (GA), this granular administration toggle was removed, enabling the capability broadly. Business & Governance Impact Plan items require an underlying SQL Database and consume additional Capacity Units (CUs). In enterprise environments, allowing all users to freely create Plan items creates significant operational and financial risks: Uncontrolled CU Consumption: Unrestricted creation leads to unexpected spikes in Capacity Unit utilization, impacting high-priority workloads across shared capacities. Database Sprawl: Automatic provisioning of underlying SQL DB dependencies without administrative oversight creates governance, compliance, and management overhead. Lack of Cost Allocation: Administrators cannot enforce least-privilege access or limit Plan item creation to authorized departments/workspaces. Proposed Solution Re-introduce the tenant-level and capacity-level governance setting under the Fabric Admin Portal: Tenant-Level Toggle: Allow Fabric Admins to enable/disable Plan item creation globally or restrict it to specific Security Groups. Capacity-Level Override: Allow Capacity Admins to turn off Plan item creation on specific capacities where CU consumption needs to be strictly budgeted or reserved for core analytics pipelines.
Cannot configure writeback for plan with guest account
It seems like it's not possible to configure writeback for a plan with a guest account. Please fix this as we use guest account to work in customer environments to avoid purchase of additional Power BI Pro licenses. Activity ID 8d5c3527-faea-4b7d-95b7-f7e6f10c819a Request ID f5d32f7e-b69d-454d-8531-818f3f165c18 Error Time Fri Jul 31 2026 16:17:28 GMT+0200 (Central European Summer Time) Service version 13.0.28509.443 Client version 2607.1.30035-train Cluster URI https://wabi-west-europe-redirect.analysis.windows.net/ Stack Trace Failed to fetch SQL credentials (status 500). 'tid' claim in the token does not match the tenant ID in the request header.Tooltip Needed for Unbound Tables in Entity Relationship View
This update introduces clearer guidance for situations where a table is not bound within the Entity Relationship view. When an unbound table is detected, a tooltip or directional prompt will appear to help users understand the required binding steps and navigate to the correct configuration area. This enhancement aims to reduce confusion, improve workflow efficiency, and support smoother model‑building experiences.Support image output in Fabric Data Agent, or avoid misleading download links
I would like Fabric Data Agent to support generating, displaying, and downloading analysis results as chart images. My understanding is that the Data Agent Code Interpreter does not currently support generating or retrieving image files such as PNG or JPEG, and can only output results in formats such as JSON. Therefore, I would like to request the following two improvements: 1. Support image output Please enable Data Agent to display generated analysis results as chart images within the response and allow users to download them in formats such as PNG. This would help business users review results more intuitively and reuse them in reports or presentations. 2. Prevent misleading responses when image output is unsupported Currently, Data Agent may respond that a chart image is available for download, but attempting to retrieve the file results in an error. When image generation or download is not supported, please avoid displaying responses or links that imply the image can be generated or downloaded. Instead, clearly indicate that output is currently limited to formats such as JSON. This would reduce user confusion and narrow the gap between Data Agent responses and the features that are actually available.Enable model configuration for Fabric Data Agents
Fabric Data Agents currently do not allow developers to configure or swap the underlying LLM, which limits flexibility compared to Azure AI Foundry and Copilot Studio. To better support advanced scenarios, consider: Allowing custom model selection Supporting hybrid patterns (e.g. bring your own model for NL→SQL generation) This would: Reduce the need to wrap Fabric Data Agents in external orchestration layers Enable advanced optimisation for domain-specific use cases
Undo functionality in Fabric IQ Plan
Hi everyone, I’ve recently started using Fabric IQ Plan, which requires working entirely within the platform. One limitation I’ve encountered is the lack of an undo functionality - there doesn’t seem to be a button for it, and standard shortcuts like Ctrl+Z don’t appear to work either. In other Microsoft tools, this is a basic capability. For example, even in Power BI Service it’s possible to use Ctrl+Z to go back one step, which makes it much easier to safely experiment and correct mistakes. Having a similar behaviour in Fabric IQ Plan would significantly improve usability and user confidence when making changes. Is there an existing feature or workaround to revert changes (e.g. going back n steps)? Thanks in advance! MiriamSetting AI Instructions for semantic models programatically
I've been looking into how I can add AI metadata to a semantic model programatically. This allows the push back of AI instruction metadata back to the users in a data ingestable process, notebook and pipeline, bypassing the Fabric 'Prep data for AI' user interface, which you may not want to give to users for iterating through instruction sets. This works well for some metadata (like table and column descriptions), but does not allow for AI instructions to be written to the model progamatically. I can edit the culture tmdl file via GIT, but when updating back into Fabric from GIT, the AI instructions come back blank. It would be great to have a REST API endpoint for writing Prep Data for AI configuration. Cheers, Kev
Support Invoking Fabric Data Agents Directly from Notebooks via API/SDK
It would be extremely valuable if Microsoft Fabric allowed developers to invoke Fabric Data Agents directly from Fabric Notebooks (PySpark/Python) through an official API or SDK. Current limitation: Data Agents are primarily accessible through the Fabric UI/Copilot experience, but there is no clear notebook-native integration for programmatic usage. Suggested capability: Allow notebooks to: - Send prompts or questions to a Data Agent - Receive grounded responses programmatically - Maintain conversational context/session state - Integrate agent workflows into data engineering and AI pipelines Example use cases: - Automated business insight generation - AI-driven anomaly detection workflows - Agentic analytics pipelines - Intelligent report generation - Integration with Spark/PySpark transformations - Enterprise AI orchestration inside Fabric Example ideal workflow: response = fabric_data_agent.ask( agent="RetailInsightsAgent", prompt="Summarize customer churn trends for Q4" ) Why this matters: Currently, many teams must rebuild semantic grounding, orchestration, and retrieval pipelines manually inside notebooks using external LLM frameworks. Native notebook integration would significantly accelerate enterprise AI adoption within Microsoft Fabric. This feature would make Fabric Data Agents much more production-ready for advanced AI engineering and analytics workflows.The Technical Roadmap: Building the Fabric Community Mobile App
I was thinking to start a project to build a Cross-Platform Mobile App (iOS & Android) using Flutter, dedicated specifically to the Microsoft Fabric community. The goal is to provide a central hub for tracking updates, community discussions, and learning resources on the go. I am currently in the infrastructure planning phase and designing the data schema. Below is the technical roadmap I’ll be following to bring this from an idea to the app stores. Phase 1: Architecture & Data Schema Schema Design: Defining entities for Users, Community Posts, Knowledge Base articles, and Certification tracks. Backend Selection: Using Firebase for real-time synchronization and authentication, integrated with Azure Functions to fetch official Fabric updates. Phase 2: UI/UX & Brand Identity 3D Integration: Leveraging my background in 3D motion graphics to create a high-end, cinematic Splash Screen and interactive UI elements using Rive for Flutter. Design System: Adopting Microsoft’s Fluent Design System principles to ensure the app feels like a native part of the Fabric ecosystem. Phase 3: Flutter Development (The Core) State Management: Implementing BLoC (Business Logic Component) to ensure the app is scalable and testable. Clean Architecture: Separating the app into three layers: Data: Repositories and Data Sources (API/Firebase). Domain: Use Cases and Entities. Presentation: UI Widgets and ViewModels. Networking: Using the dio package for robust API handling and interceptors. Phase 4: Testing & Quality Assurance Unit & Widget Testing: Ensuring every button and logic flow works across different screen sizes. Integration Testing: Testing the end-to-end flow from login to data retrieval. Phase 5: CI/CD & Production Deployment Pipeline: Setting up GitHub Actions for Continuous Integration and Continuous Deployment. App Store Optimization (ASO): Preparing the Google Play Store and Apple App Store listings with high-quality visuals. Monitoring: Using Firebase Crashlytics and Google Analytics to monitor performance and user engagement post-launch. I’d love to hear your thoughts! What features would make your life as a Fabric professional easier? Looking forward to building this together!
