Power BI Updates & Release Notes

AI powered Power BI reporting skills is a collection of multiple skills that enables AI agents to automate report creation, from designing pages to publishing to Fabric. Now available through the Power BI authoring plugin in Skills for Fabric —a first-party catalog of agent skills for Microsoft Fabric optimized for GitHub Copilot CLI—this capability allows agents to author reports through natural language, write schema-correct PBIR files, reload Power BI Desktop, capture screenshots, and iteratively refine reports for a seamless, end-to-end authoring experience.

Original source

Microsoft Build 2026 marks a major shift in how developers build data experiences with AI agents. Today we're announcing two capabilities that bring agentic analytics to the forefront: Agent Skills for Power BI, which let developers prompt an AI agent to build and refine semantic models and reports, and Fabric Apps for Semantic Models, which enable AI agents to build and deploy Fabric-native web apps on semantic models. Both capabilities will accelerate the time it takes to go from raw data to a polished analytics solution with just natural language prompts.

Original source

The May 2026 Power BI update brings visual calculations and custom totals to general availability, adds Copilot summary shortcuts on the report ribbon and visual header, introduces a new Get Data experience in Power BI Desktop preview, and ships multiple reporting features including set as landing page, subscriptions support for reports in org apps, and column fixed width for table and matrix.

For a quick summary of the May features, read on. For a detailed look at these updates and more enhancements, see the Power BI May 2026 Feature Summary.

Download the May 2026 version of Power BI Desktop.

To stay up to date on bug fixes and improvements as they're announced, visit the change log for Power BI Desktop.

Copilot and AI

Explore now supports perspectives for large data models—create a perspective and set it in File > Settings > Report settings (service) or File > Options and settings > Options > Current File > Report settings (Desktop) to give consumers a focused list of fields. Matrix auto-expand now adjusts which fields are expanded when you add or reorder fields, so new fields are visible. Users can show or hide totals on table and matrix visuals from the Explore toolbar. Formatting applied to a matrix during report authoring now carries over when users launch an exploration from it.

A Summarize button on the report ribbon generates a report-wide summary in the Copilot pane, surfacing key trends, highlights, and changes across pages and visuals. A Copilot summary option on the visual header produces a visual-level summary, calling out trend shifts, category differences, and key drivers. For more information, see Summarize a report with Copilot.

The Copilot narrative visual now supports the app-owns-data embed scenario. Embed the narrative visual in customer applications where report consumers don't need to sign in. For more information, see Create a narrative visual with Copilot for Power BI.

A new Copilot tooling format for Prep data for AI stores Copilot metadata in a Git-friendly, documented structure designed for direct editing and source control. It decouples Copilot from Q&A, moves metadata outside the Analysis Services database to reduce memory load, and improves performance for metadata-only scenarios. Support rolls out incrementally. For more information, see Prepare your data for AI.

Reporting

Visual calculations let you add running sums, moving averages, percent of parent, and other calculations directly to a visual without adding DAX measures to your semantic model. Calculations operate on aggregated data within the visual and are aware of row position. Select a visual and choose New visual calculation from the ribbon or context menu. For more information, see Using visual calculations.

Custom totals now include None and Average options in addition to Sum, Min, Max, Count, and Count Distinct. Right-click a numerical column in a table or matrix and select Customize total calculation to choose the aggregation for the total row. For more information, see Custom totals.

Designate any page as the landing page so report viewers always start on the right page. Right-click a page tab and select Set as landing page, or use the page formatting pane. For more information, see Set a landing page.

Set a format string locale in report settings so all locale-aware format strings use your specified locale instead of the viewer's browser language. Note: this affects how values display in visuals only—USERCULTURE() and metadata translations still use the viewer's browser locale. For more information, see Default format string locale.

Reports pre-populate input fields with default values from user data functions, and users can submit task flows without specifying every parameter value. For more information, see Understand translytical task flows.

The input slicer now supports numeric fields and a numeric input mask. Use operator syntax to filter values with expressions like ranges (1-2), greater than (>2), less than (<2), not equal (<>2), or wildcards (2*). Without a data column, limit the slicer to numeric entry for translytical task flows. Invalid input now shows clear feedback. For more information, see Create and use an input slicer.

Set a property so fields added to a matrix visual automatically expand in embedded scenarios, displaying expanded hierarchies without requiring user interaction. Explore also auto-expands as fields are added to the matrix. For more information, see Power BI embedded analytics playground.

Lists in text boxes now preserve formatting correctly, indentation renders as expected, and you can paste bulleted lists directly from Word. For more information, see Add text boxes, shapes, and smart narratives to Power BI reports.

The Azure Maps visual has an updated formatting pane aligned with the modern formatting experience in Power BI, with map settings, layer options, and style controls in a consistent layout. For more information, see Get started with Azure Maps Power BI visual.

The theme dropdown has an updated look. The Remove custom theme tile is renamed Reset to default for clarity. Fixes include: the slicer now defaults to dropdown mode without issues, and the first page correctly uses the new canvas size. For more information, see Visual defaults in Power BI reports.

Specify exact pixel widths for columns in table and matrix visuals in the formatting pane, or set a default width for all columns. Auto-size behavior now includes a Fixed width option in addition to fit to content and grow to fit. Widths can be set differently for desktop and mobile views. For more information, see Table and matrix visual overview.

Create and manage email subscriptions for Power BI reports directly within org apps. Consumers viewing reports in an org app can stay informed with scheduled email snapshots and links back to the report, just as they would when accessing reports elsewhere in the Power BI service. For more information, see Email subscriptions for reports in org apps.

Modeling

Users with edit permissions on semantic models are now taken directly to the web modeling experience instead of the model details page. Most actions from the model details page are now integrated into model view. For more information, see Edit semantic models in the Power BI service.

Data connectivity

A redesigned Power Query Get Data experience provides a unified place to discover and connect to data sources. Key highlights include improved data source discovery, a streamlined connection flow, and built-in accessibility features such as keyboard navigation and dark mode. Brings greater consistency to Power Query across Microsoft Fabric, Power BI Desktop, and Microsoft Excel. For more information, see New Get Data experience.

Resources

Want to learn about Power BI through videos and other engaging content? Check out these video sources and content:

• See all Power BI playlists on YouTube.
• Power BI YouTube channel: Official Microsoft Power BI channel.
• Follow Power BI on X: @MSPowerBI.
• Go to the Power BI forums in the Microsoft Fabric Community.

Note

Some resources use earlier versions of Power BI Desktop or the Power BI service.

If your organization needs an earlier version, download it. Use the most recent version of Power BI Desktop when possible. Earlier versions have these limitations:

• Previous releases of Power BI Desktop aren't serviced. Use the most recent release for the latest features and updates.
• Previous versions can't open files created or saved in newer releases of Power BI Desktop.
• If you load a report from a newer release, get a warning, and then save it in a previous version, you lose information related to new features.
• Only English versions of Power BI Desktop are archived.

Download the May 2026 version of Power BI Desktop.

Find previous monthly Power BI updates in the Power BI monthly updates archive.

Original source

How to identify affected semantic models

Excel and CSV files remain a valid data source for Power BI semantic models and reports. You can use them to create reports in the Power BI service from the Create page or in Power BI Desktop.

This blog post is for users who created reports using the old experience to import an Excel or CSV file from the Create page in the service. There's no impact for users who created semantic models in Power BI Desktop.

Semantic models created by the old import files experience can't be edited in the browser or downloaded and have no options to schedule a refresh. If your semantic model can be edited in the browser, downloaded, or has a scheduled refresh option, you're already using the new Excel and CSV connectors, and there's no impact on your semantic model and reports.

What is changing?

Reports created using the old import Excel or CSV experience stop refreshing from their source file as of July 31, 2026. There's no workaround or fix to keep them updated. These reports need to be recreated to continue updating. For guidance on creating reports from Excel and CSV files, refer to Get data from Excel workbook files.

These reports can still be consumed and edited, even past this date, but the data no longer stays in sync with the source Excel file.

We're removing the old Excel and CSV import experience in Power BI service, which is currently accessible through the Create page, by May 31, 2026. The existing options for Excel and CSV, as well as Get data, continue to be available. We recommend you start using the other options for Excel and CSV files.

Figure 1: The legacy / old Excel and CSV import option is outlined in red at the bottom center of the Power BI Service Create page.

Key dates to remember

The legacy Excel and CSV import experience will be deprecated in phases:

• May 31, 2026: You will no longer be able to create new semantic models using the legacy import experience.

• July 31, 2026: Existing semantic models created with the legacy experience will stop refreshing. Reports based on these models will still open, but their data will no longer update. To keep your reports current, recreate them using a supported import method.

• August 31, 2026: Existing semantic models created with the legacy experience will no longer be available for querying. Reports bound to these semantic models will no longer open and will display an error message.

For more details on identifying these semantic models and options to recreate your reports, refer to the Get data from Excel workbook files documentation.

Next steps

Learn more: Review the documentation for guidance on using the current Excel and CSV connectors.

Provide feedback: Share your thoughts in the Power BI Community.

Submit ideas: Suggest new features at Power BI Ideas.

Thank you for being part of the Power BI community and for your continued feedback as we evolve the platform.

Original source

Welcome to the April Power BI update!

Power BI’s April 2026 update is here, bringing continued improvements across Copilot and AI, reporting, visuals, and modeling. This release includes more flexibility when working with layouts and visuals, expanded Copilot experiences—especially on mobile—and several preview features that continue to enhance performance and authoring workflows. You’ll also find important announcements and deprecation notices to keep in mind as you plan ahead. With FabCon fresh on our minds, now is a great time to dive into what’s new and see what’s coming next.

Contents

• April Power BI Video
• https://youtu.be/Nn19PQF59MM
• April Power BI Desktop
• Power_BI_April_2026_Feature_Summary
• Version number: v: 2.153.910.0
• Date published: 21/4/2026

Events and Announcements

The FabCon SQLCon live recap series starts April 14th

Whether you were there or not, this is your fast track to staying in the loop. In this series, we break down keynotes and corenotes into clear insights and standout demos you can use right away. If it mattered at FabCon, you’ll find it here.

Join us, get inspired, and stay connected to what’s happening across the Fabric community. Register now.

Could you be the next Power BI Dataviz World Champion?

During the recent world championships, one of the finalists confessed that they almost didn’t enter the contest. When we say, “You’ll never know unless you enter!” we really mean it!

Don’t miss your chance to enter the next one. Let us know you are interested and we’ll let you know when it starts!

General

Deprecation of Old File Picker Experience in Power BI Desktop

Starting in April, as part of the SU04 release, users will no longer be able to access the old file picker experience in Power BI Desktop. Last January, we announced an updated file picker experience that provides users with a more intuitive, straightforward way of navigating between files and folders. As of SU04, we are moving the updated experience out of preview and making it the default experience in Power BI Desktop. With this change, users will no longer be able to toggle between the old and updated experience.

Note: No action is required from users as part of this deprecation; this is simply an informational announcement.

Deprecating the Built-In Netezza ODBC Driver

The IBM Netezza ODBC driver has been Generally Available for several weeks, and we are beginning the deprecation from the previously built-in ODBC driver to the newer version.

Customers do not need to install the new connector; you may reuse your existing connector but will need to install the new ODBC driver. We encourage customers to do this as soon as possible to ensure a smooth transition.

IBM_ODBC_Netezza_driver_that_will_be_deprecated_from_Fabric

Image: IBM ODBC Netezza driver that will be deprecated from Fabric.

Refer to the IBM Netezza ODBC documentation for more information.

Copilot and AI

Copilot in Power BI mobile now offers expanded features

In-report Copilot in Power BI Mobile apps just got a major upgrade: you can now have a full, back-and-forth chat with your report, right from your phone or tablet. Instead of stopping at summaries and prebuilt prompts, in-report Copilot in the mobile app now supports open-ended questions and follow-up conversations, all grounded in the specific report you’re viewing.

Ask about a metric or KPI, dig into what’s driving the numbers, and even get AI-generated visualizations to take your analysis further.

Every answer includes citations back to the exact visuals Copilot used, so it’s easy to validate insights and keep exploring with confidence on the go.

On iPhone and iPad, voice dictation makes it even faster to get hands-free answers while you’re prepping for a conversation or reviewing results away from your desk.

Two_mobile_phone_screenshots_display_a_sales_report_for_Contoso_showing_revenue

Figure: Using in-report Copilot on the Power BI Mobile app to ask natural‑language questions and explore insights directly within a report.

To learn more, refer to the documentation In-Report Copilot in Power BI Mobile Apps

Reporting

Modern visual defaults and customize theme improvements (Preview)

This month's update adds a base theme switcher to the Customize current theme dialog (View ribbon > Themes). If your custom theme doesn't yet work with the new modern defaults, you can use the base theme switcher to revert to the previous base theme until you've had a chance to update your custom theme. You can also use the base theme switcher to update an existing report created with an older base theme to the latest base theme.

Screenshot_of_the_Customize_current_theme_dialog_showing_the_base_theme_switcher

Figure: Base theme switcher in Customize current theme lets you revert to the previous base theme or update an older report to the latest modern defaults.

This update also includes common page sizes for each aspect ratio type in the Canvas settings > Size drop-down. Custom sizing remains available for any dimensions you need.

Canvas_settings_panel_in_Power_BI_showing_Type_dropdown_set_to_16_9_and_Size_dro

Figure: Canvas settings with preset size options for 16:9 aspect ratio reports in Power BI Desktop.

Table and matrix built-in styles have been fixed, with banded rows now enabled by default, as well as default +/- buttons for matrix visuals. Axis colors now also use the correct structural color, fixing issues with the Innovate and Orchid custom themes.

Additionally, the built-in theme tiles have an updated look and the Reset to default tile is distinguished from the other built-in theme tiles. The behavior of the tile is not new; it simply clears any custom theme applied, leaving any formatting changes to individual visuals untouched until the visual itself is reset to default in the formatting pane.

Power_BI_theme_gallery_showing_eight_built-in_theme_options_as_thumbnail_preview

Figure: Built-in theme gallery in Power BI Desktop with the "Reset to default" option to remove any custom theme and update report to latest base theme.

To get the latest version in an existing report, go to the Customize current theme dialog and select Update theme. Or choose the Reset to default tile in the Theme drop-down.

Thank you for your feedback during the preview. You can continue to provide feedback and learn more about Visual defaults in Power BI reports.

Fixed size layout for card, button slicer, and list slicer visuals

Card, button slicer, and list slicer visuals now support a Fixed size option in the Layout section of the format pane. Instead of specifying how many items to display, you can define the exact pixel dimensions for each card, button, or list item. When the visual container isn't large enough to display all items at the specified size, scroll bars appear automatically.

This update also renames Autogrid to Fit to space for clarity. When Fit to space is on, items grow or shrink to fill the visual container based on the items present. When Fit to space is off, the visual reserves space for the specified number of items, even when fewer items exist. For list slicer, the Fixed number of buttons option, equivalent to Autogrid off, is also renamed to Fit to space for consistency, especially when changing between visual types.

Fixed size gives you precise control over each item's dimensions. As you resize the visual container, items maintain their specified height and width rather than scaling proportionally. This behavior is useful for creating consistent layouts across the report page or ensuring uniform button, list, or card sizes across multiple visuals. For list slicers specifically, fixed size provides a more natural experience when working with hierarchies. Expanding and collapsing hierarchy levels causes the number of visible items to change dramatically, and fixed-size items ensure consistent spacing as you navigate through the data.

To use fixed size:
Select your visual.
Expand the Layout section in the Format pane (found under Multi-card layout or Multi-button layout and toggle Fixed size to On.

For Vertical arrangement, set the Height value in pixels.

For Horizontal arrangement, set the Width value in pixels.

For Grid arrangement, set both Height and Width value in pixels.

When Fixed size is enabled, Fit to space is disabled since dimensions are now controlled explicitly.

Screenshot_of_Power_BI_Desktop_showing_a_list_slicer_visual_with_the_Format_pane

Figure: List slicer with Fixed size enabled and Height set to 30 pixels.

To learn more, refer to the documentation for card visuals, button slicers, and list slicers.

Card visual: Category interactivity and formatting updates

The card visual now provides clear visual feedback when you select a category header—the selected card appears highlighted while others dim, making it easy to see your current selection. When you add multiple data columns to the category field, the values concatenate in the category header for a cleaner display. You can also use Edit interactions to control which visuals the card filters, giving you more flexibility in how your report responds to selections. Additionally, top-level images display correctly when the image data is base64 encoded, so you can use images from your data without extra conversion steps.

Power_BI_card_visual_with_a_category_header_selected_the_selected_card_is_highli

Figure: Card visual showing category header selection highlighting and dimming.

For more information about the card visual, refer to Create a card visual in Power BI.

Azure Maps visual: Map style picker sync

When you change the map style using the in-visual style picker, the selected style now persists with the Format pane. Your styling choices stay consistent across both the style picker and the Format pane, giving you predictable behavior as you design your reports.

Azure_Map_visual_in_Power_BI_showing_North_America_with_blue_bubble_markers._An

Figure: Azure map updating the format pane style when the report creator adjusts the style on the visual itself.

For more information about map styling options, refer to Get started with Azure Maps Power BI visual.

Easily identify preview visuals in the Visualizations pane

We appreciate your continued feedback on visuals and have finished addressing the concern that preview visuals weren’t always clearly notifying their preview status. Preview visuals now display (preview) after their names in the Visualizations pane, making it easier to identify which visuals are still in preview. Additionally, preview visuals now appear below the divider in the pane (alongside custom and unpinned visuals), clearly separating them from generally available visuals. These changes help you quickly understand when you’re working with a preview feature.

Power_BI_Visualizations_pane_showing_the_Build_visual_tab_with_visual_type_icons

Figure: Visualization pane when editing a report in Power BI Desktop showing preview visuals below the divider line and with (Preview) after their name.

For more information about visuals, refer to Visualizations overview in Power BI.

Narrative Visual Default Type Update

The Narrative visual currently offers two modes: Copilot and Custom, giving report authors flexibility in how they generate and customize summaries. Previously, authors needed to explicitly choose which mode to use when creating the visual.

Power_BI_April_2026_Feature_Summary

We’ve recently improved this experience by introducing a smarter default. If a user has a Copilot license, the Narrative visual now opens in Copilot mode by default, making it quicker to get AI‑powered insights right away. We also increased the character limit to 10,000, enabling richer prompts and more detailed narratives. Authors can still easily toggle between Copilot and Custom modes at any time, ensuring full control over their storytelling workflow.

Power_BI_April_2026_Feature_Summary

This improved default is available now, and learn more with the Create Smart Narrative Summaries documentation.

Modeling

Direct Lake calculated columns and tables (Preview)

Direct Lake storage mode accelerates time to insights by unlocking incredible performance directly against OneLake, without the need to manage costly, time-consuming refreshes for large volumes of data.

Calculated columns (unmaterialized) on Direct Lake on OneLake tables is in the process of deployment and will be available in the service in the next few weeks.

Calculated tables referring to Direct Lake on OneLake columns.

These features are particularly helpful when adding columns and creating tables upstream isn’t feasible, such as when data preparation in OneLake is owned by another team. Refer to the Create calculated columns in Power BI Desktop documentation for more information on calculated columns.

Screenshot_of_DAX_expression_to_define_a_calculated_column_to_define_customer_ag

Figure: To use the feature in Power BI Desktop, you must enable the Direct Lake calculated columns (unmaterialized) preview feature switch.

User context aware calculated columns (Preview)

We are introducing the ability to make calculated columns user-context aware by dynamically responding to DAX functions including UserCulture(), UserPrincipalName(), CustomData(). This enables new scenarios like data translations, and we’re excited to see the creative ways the community will use this!

User-context-awareness can be set for calculated columns on Direct Lake on OneLake, Import and DirectQuery tables using the Expression Context property. Direct Lake on OneLake is in the process of deployment and will be available in the service in the next few weeks. Refer to the Create calculated columns in Power BI Desktop documentation for more information on user-context-aware calculated columns.

Screenshot_of_a_DAX_expression_in_the_formula_bar_to_dynamically_return_a_column

Figure: A DAX expression in the formula bar to dynamically return a column with translated values based on the USERCULTURE() function.

In the following example, a multi-lingual semantic model and report uses both data and metadata translations. By changing the language URL parameters to simulate a different browser locale, everything is displayed in Portuguese, including product names from the Product table.

Animated_GIF_of_a_sales_report._The_browser_locale_is_overridden_to_Portuguese_a

Figure: To use the feature in Power BI Desktop, you must enable the User-context-aware calculated columns preview feature switch.

DAX user-defined functions (Preview)

Alongside our ongoing preview of DAX user defined functions, we’ve enhanced the DAX NAMEOF function to give you much finer control over how object names are returned. NAMEOF now supports optional parameters that let you choose exactly which part of a table, column, measure, or calendar name to return, and control how that name is formatted for display. The new function signature is:

NAMEOF ( <object> [, <component> [, <escaped>]] )

This makes it possible to programmatically reference just the table, just the column, or just the measure and improve usability for display scenarios, while keeping existing behavior unchanged for current models.

Visualizations

Date Picker by Powerviz

The Powerviz Date Picker offers a modern calendar view, Presets, Pop-up mode, Smart Button Label, Custom Preset Title, and more, making it a must-have date slicer for Power BI reports. Its rich formatting options help with brand consistency and a seamless UI experience.

Key Features

• Smart Button Label: Make the Pop-up button instantly understandable by showing the selected range, preset name, or both as per your choice.
• Custom Preset Title: Localize preset names to your users’ language. Fully customize what each preset displays across the report.
• First Day of the Week: Customize your calendar experience by choosing which day your week starts — perfect for global teams and regional preferences.
• Display Mode: Choose between Pop-up and Canvas modes.
• Presets: Many commonly used presets like Today, Last Week, YTD, MTD, or create your preset using field.
• Default Selection: Control the date period selected when the user refreshes or reopens the report.
• Filter Type: Choose between Range and Start/End types.
• Themes: 15+ pre-built themes with full customization.
• Holidays and Weekends: Customize holidays/weekends representation.
• Month Style: Select single- or double-month date slicer.
• Other features included are Multiple date ranges, Import/Export Themes and more.
• Try Date Picker visual for FREE from AppSource
• Check out all features of the visual
• Step-by-step instructions
• YouTube Video
• Learn more about visuals
• Follow Powerviz
• Power_BI_April_2026_Feature_Summary Power_BI_April_2026_Feature_Summary

Drill Down Waterfall PRO by ZoomCharts

With powerful customization options and intuitive interactions, Drill Down Waterfall PRO helps report creators present financial and operational data in a way that is easy to explore and understand. The latest update expands the functionality of Change thresholds with a new Automatic mode that detects subtotals and calculates the difference between consecutive subtotal segments, enabling clearer storytelling in multi-period reports.

Key features include:

• Automatic change thresholds: Automatically detect subtotal columns and calculate the change between each consecutive subtotal and the first and last subtotal.
• Custom Sequence: Have full control over the column order with the Sequence field.
• Drill Down: Use multiple categories to enable drill down directly on the chart.
• Automatic Subtotal Calculation: Automatically calculate subtotals if their fields are empty.
• Annotations: Display markers and comments on the chart from Comment and Comment Marker fields.
• Customization: Customize X and Y axes, legends, tooltip content, and adjust the appearance settings for positive, negative and total columns separately.
• Thresholds: Display up to four constant or dynamic thresholds as lines or areas.
• Cross-chart filtering: Dynamically filter data across multiple visuals to create interactive, insightful and intuitive reports.
• Get Drill Down Waterfall PRO on AppSource
• Power_BI_April_2026_Feature_Summary Power_BI_April_2026_Feature_Summary

Closing

That’s a wrap for the April 2026 Power BI update. This month’s release builds on recent investments across reporting, modeling, visuals, and Copilot, while introducing new previews for you to explore. As always, we appreciate your feedback—especially on preview features—and encourage you to continue sharing your input as we work on future updates.

Power_BI_April_2026_Feature_Summary

Original source

Translytical task flows

Translytical task flows take interactive Power BI reports to the next level where users can update records, add annotations, and trigger actions in external systems—all without leaving the report.

With translytical task flows, report consumers become active participants. Instead of viewing data and then switching to another application to take action, users can complete their entire workflow within the Power BI interface.

Translytical task flows connect Power BI reports to Fabric User Data Functions. When a user interacts with a report—selecting a record, entering a value, and clicking a button—the report passes the context of what they have selected to a function to execute the requested action.

Translytical_Task_Flows_Generally_Available

Figure 1: Example of a translytical report experience, where users can review project status and add notes directly in Power BI.

Capabilities

Adding data: Insert new records into your database directly from the report. For example, a customer service representative can add a new customer record while reviewing existing accounts and see it immediately in the report.

Updating data: Update existing records without leaving the report. A logistics coordinator can change an order status or add notes to a shipment record as they work through their queue.

Deleting data: Remove records that are no longer needed. An inventory manager can delete discontinued products from active lists while reviewing stock levels.

Calling external APIs: Trigger actions in other systems through API requests. A sales representative can request a discount approval that posts directly to Microsoft Teams, where a manager can review and respond. Not only can they trigger these but capture and pass on a message too.An_adaptive_card_requestion_an_update_in_a_Teams_channel

Figure 2: An adaptive card in a Teams channel requesting an update to a status with link to the Power BI report to update it using translytical task flows.

These capabilities can work together too. For instance, a sales opportunity report can include a discount request form. When a user selects opportunities from a filtered table, enters a discount percentage, and adds a justification, clicking the submit button sends all that context to a function. The function then processes the discount and posts the request to Teams with the relevant details, creating a complete audit trail from insight to action. The updated data is also immediately visible in the same report!

Other scenarios for translytical task flows include:

• Data annotation and quality management: Field teams often discover data issues while working in reports. With translytical task flows, they can correct a misspelled customer name, update an outdated address, or add contextual notes to records immediately. This approach improves data quality at the point of discovery rather than routing corrections through separate processes.

• Workflow automation: Business processes frequently require approvals, notifications, or ticket creation. Translytical task flows can trigger these actions based on report context. A procurement analyst reviewing vendor performance can flag a supplier for review, automatically creating a task in the appropriate system with all the supporting data attached.

• AI-assisted decision making: Use AI functions in your extract load and transform process with Fabric Notebooks to categorize or summarize your data. These can then be reviewed and updated in your Power BI report.

For data write-back scenarios, User Data Functions currently have native connection management for the following Fabric data sources:

• Fabric SQL databases
• Fabric warehouses
• Fabric lakehouses (for files)

Once you have your scenario in mind, building your first translytical task flow involves these main tasks:

• Storing your data in a Fabric data source.
• Developing a User Data Function to handle the action.
• Creating a Power BI semantic model to use this data.
• Building a Power BI report with interactive elements to capture the user's input and call the function.

Copilot in Microsoft Fabric is built into many Fabric workflows and GitHub Copilot in Visual Studio Code with MCPs to access learn documentation and connect to Power BI semantic models can help you accelerate this process too.

Translytical task flows represent a shift in how organizations can use Power BI. Reports are no longer only for analysis—they enable you to take action. By connecting insight directly to execution, teams can reduce the time between identifying an opportunity and acting on it.

Next steps

Learn more: Refer to the Translytical task flows documentation for detailed guidance on building your first task flow.

Explore User Data Functions: Visit the User Data Functions in Fabric documentation to understand all the capabilities it supports.

Submit ideas: Share your feedback and feature requests through the Power BI ideas forum.

Original source

Updated on 20 April 2026: Thanks so much for all the engagement on this topic since the original publishing of this blog post on 2 April 2026. We’ve seen several recurring questions, and we’d like to clarify our position on a few key points.

Dataflow Gen1 remains supported, but it is in a legacy state and won’t receive future innovation. We understand many customers have built important, business‑critical solutions on Gen1, and those existing workloads can continue to run while future investments are planned around Dataflow Gen2.

Primary call to action for Premium customers: For Premium customers who already have access to Fabric, Dataflow Gen2 is the recommended path to take advantage of the latest improvements across functionality, productivity, performance, pricing, scale, reliability, and more. Many teams find it helpful to start by evaluating Gen2 for new or evolving workloads where these benefits can be realized quickly.

Guidance for Pro and Premium Per User (PPU) customers: Many customers rely on Dataflow Gen1 in Pro/PPU today, and it can continue to be the right choice depending on the scenario. If Gen1 best fits your current use case, it remains supported and existing workloads can continue to run as-is. As we introduce new Dataflow Gen2 paths for Pro/PPU scenarios, we’ll share clear guidance and recommended steps to help with a smooth transition.

GCC support for Premium customers: For customers using Dataflows Gen1 in GCC environments on Premium, Dataflow Gen2 support in GCC will be available before any future transition milestones, ensuring a supported upgrade path.

More granular Fabric controls: Customers have asked for finer‑grained enablement than today’s all‑or‑nothing Fabric switch. Work is underway to provide more granular administrative controls, including the ability to enable only Dataflow Gen2, so capabilities can be rolled out incrementally with the right governance.

Original post (2 April 2026)

For more than eight years, customers have relied on Power BI Dataflows (Gen1) as a core part of their analytics solutions. We’re grateful for the trust you placed in the Power Query experience to build reusable, low-code data preparation pipelines that power reports, semantic models, and downstream analytics.

Now, we’re sharing an update on the future of Dataflows.

Dataflows Gen2 builds on everything you know from Gen1—preserving the familiar Power Query authoring experience—while delivering major improvements in scale, flexibility, cost efficiency, and manageability. Going forward, all new Dataflow innovation will land only in Dataflows Gen2.

The future of Dataflows Gen1

Power BI Dataflows Gen1 has reached the end of active innovation and is moving into a Legacy state:

Existing Gen1 dataflows will continue to work for the foreseeable future. However, specific retirement dates for Gen1 are being finalized, and we’ll share details as plans progress. For customers running Gen1 at Premium capacity, we will provide at least 12 months’ notice before Dataflow Gen1 is retired. (If you’re using Gen1 on Pro or Premium-Peruser, we still strongly recommend planning a move to Dataflows Gen2 to take advantage of the latest investments.)

In the meantime, Gen1 artifacts will remain available and will be clearly marked as Legacy in product experiences, including the New Artifact menu.

No new features are planned for Dataflows Gen1. Support will be limited to a narrow set of high impact issues, where changes can be delivered safely within the existing architecture. Many of the remaining Gen1 limitations would require significant architecture changes and are best addressed by moving to Dataflows Gen2, which was designed to solve these scenarios more comprehensively.

The previous statements apply to all Dataflows Gen1—including usage on Pro, PPU, and Premium licenses.

What’s new in Dataflows Gen2

Dataflows Gen2 retains the familiar Power Query experience while introducing substantial platform-level improvements to scale, performance, governance, and cost efficiency. The following updates provide additional details on each benefit area to help customers understand the full scope of Gen2 enhancements.

More flexible destinations

Dataflows Gen2 supports a significantly broader range of output destinations, enabling alignment with diverse data architectures across business, departmental, and enterprise scenarios.

SharePoint and OneDrive are ideal for business users who need refreshed files (CSV, Excel) for downstream workflows, Office automation, or integrations with Power Automate and Teams.

Azure Data Lake Storage is a great fit for customers building scalable ingestion pipelines for data science, machine learning, or lakehouse scenarios.

Azure SQL Database / SQL MI enables operational reporting, standardized relational storage, and hybrid analytics scenarios.

Microsoft Fabric Lakehouse / Warehouse / SQL analytics endpoints are the most seamless destination option for customers aligning with the Fabric vision. Gen2 integrates deeply with Fabric runtimes, unlocking better performance, more consistent semantics, and governance alignment.

Snowflake and other cloud databases support multi‑cloud architectures and reduces friction for enterprise customers already standardized on multiple warehouse technologies.

This broader range of destinations allows teams to use dataflows as a general-purpose low-code data ingestion and transformation layer, enabling many more scenarios never supported before with Dataflow Gen1.

Improved performance and scale

Dataflows Gen2 is built on the Fabric runtime and a modernized execution engine that delivers a step‑change in performance, reliability, and scalability.

Modern query evaluation leverages Fabric's elastic compute layer to automatically manage scaling behavior, minimizing the need for manual optimization or workload tuning.

Fast Copy technology supports high-throughput ingestion into Fabric destinations, enabling sustained data movement measured in gigabytes per minute.

Parallelized execution enables multiple partitions and transformation steps to be processed concurrently, significantly reducing refresh durations compared to Gen1.

Enhanced support for large datasets includes improvements in memory handling, high-cardinality data processing, and unbounded ingestion patterns.

Predictable refresh behavior ensures more consistent performance under varying workload conditions.

Together, these improvements establish a more durable and high-performance engine for data preparation, especially in enterprise environments that rely on large or frequently refreshed dataflows.

Built-in AI assistance

Dataflows Gen2 introduces integrated AI capabilities designed to accelerate development, improve quality, and reduce the learning curve for users working with complex transformations.

Copilot-assisted authoring converts natural-language instructions into Power Query logic, improving productivity and lowering the barrier to entry for users with limited M expertise.

Code explanation capabilities translate complex or legacy M scripts into easy-to-understand natural language descriptions, improving maintainability and simplifying onboarding.

Automated performance and foldability recommendations help users align transformations to foldable patterns, resulting in faster load times and lower compute consumption.

AI-powered data quality insights assist with identifying semantic types, outliers, and join keys and detecting common data issues early in the pipeline.

These capabilities provide consistent guidance across the authoring lifecycle and support organizations with diverse skill levels.

Richer diagnostics

Gen2 introduces a more comprehensive set of diagnostics designed to improve traceability, troubleshooting, and operational reliability.

Detailed refresh history now includes detailed timing information to clarify where time is spent during execution.

Expanded logging and instrumentation provide visibility into foldability decisions, connector behaviors, authentication flows, and network operations.

More consistent refresh semantics across all destinations ensures uniform behavior regardless of the target storage system.

Greater operational transparency supports root-cause analysis and reduces the time required to identify and resolve failures.

These diagnostic improvements help teams manage dataflows more efficiently and maintain higher levels of operational readiness.

Tiered pricing and potential cost savings

The introduction of a tiered pricing model is a major advantage for customers transitioning from Gen1.

Gen2 decouples Dataflows execution from Premium capacity consumption, aligning compute usage with Fabric’s capacity unit (CU)–based architecture.

This model allows customers to pay only for the compute resources required rather than maintaining always-on Premium capacity for workloads that may be intermittent or variable.

Elastic scaling ensures that high-volume or burst workloads can consume proportionally more compute during peak times, while lighter workloads incur lower costs.

Organizations with unpredictable or seasonal refresh patterns may see material cost reductions compared to Gen1 running on Premium capacity.

This flexible cost model provides more predictable and efficient resource utilization, particularly for enterprises with diverse or rapidly evolving data refresh patterns.

Our recommendation: Start planning your upgrade

If you’re starting a new project, we strongly recommend using Dataflows Gen2. It offers better performance, richer diagnostics, built‑in AI, broader destination support, and a more flexible cost model.

For existing Gen1 dataflows, now is the right time to begin planning your upgrade:

• Use Save as Dataflow Gen2 for quick, low effort upgrades of individual dataflows.
• For larger migration scenarios, the Save As API enables bulk migration and automation, supporting CI/CD workflows.
• For programmatic or large‑scale upgrades, refer to the Migrate to Dataflow Gen2 (CI/CD) guidance.

Starting now allows you to modernize incrementally, beginning with a subset of your portfolio while keeping all your Power Query skills fully transferable.

Thank you

Thank you again for your long-standing investment in Dataflows and Power Query. We’re excited to support your transition to Dataflows Gen2 and to help you unlock new capabilities for the next generation of analytics solutions.

Original source

If you haven’t already, check out Arun Ulag’s hero blog “FabCon and SQLCon 2026: Unifying databases and Fabric on a single, complete platform” for a complete look at all of our FabCon and SQLCon announcements across both Fabric and our database offerings.

The next evolution of in-report Copilot on the mobile app is now available in preview: a full conversational chat experience that lets you summarize, inquire, and analyze report data directly from your phone and tablet.

Previously, in-report Copilot on the mobile app supported summaries and predefined prompts. This made it easy to get a quick overview but limited how far users could go when they wanted to explore their data more deeply. With this update, in-report Copilot on the mobile app moves to a full chat experience, bringing the same chat‑based capabilities you know from Power BI service, optimized for mobile and grounded in the report you’re viewing.

In_a_Contoso_sales_report_in_Power_BI_Mobile_a_user_opens_inreport_Copilot_from

Figure: Using the in-report Copilot in mobile app to answer on-the-go data questions.

Getting started

Every report that meets Copilot requirements in Power BI has the Copilot button in its header, ready to assist you with just a single tap.

In the chat pane you can:

• Generate a summary of the current report
• Ask a question about the report’s data
• Ask a follow up question to explore further
• Use the prompt gallery to help get started with sample prompts

Two_mobile_phone_screenshots_display_a_sales_report_for_Contoso_showing_revenue

Figure: Using in-report Copilot on the Power BI Mobile app to ask natural‑language questions and explore insights directly within a report.

Just like in Power BI service, in-report Copilot on the mobile app analyzes the report content for you, surfacing insights that would otherwise take time and effort to uncover grounded on the report data.

Discover more

When you ask about a metric or KPI in a report, in‑report Copilot often includes a visualization along with its text response. Tap the visual to open a pane where you can interact with it and explore the details.

If you ask in-report Copilot to summarize a report, the response will include a summary with relevant citations. You can tap on a citation to open the visual in focus mode to view and interact with the visuals.

Keep chatting with your data. Ask in-report Copilot follow-up questions, refine your queries, and dive deeper into report insights.

To share data insights with your team, copy the response using the Copy action or tap Share from the visual pane.

Voice support: On iPhone and iPad, in-report Copilot supports dictation, so you can ask hands-free questions. Use Read aloud to listen to the response.

The new in-report Copilot chat experience on the Power BI mobile app makes it easier to analyze report data in context—whether you’re reviewing results on the go or preparing for a discussion around a specific report.

Share your feedback

Stay tuned as we enhance your experience with more capabilities in the coming releases.

We invite you to chat with in-report Copilot on the mobile app and share your thoughts in the comment section Your feedback will help us shape the future of chat on the Power BI mobile app.

Original source

If you haven’t already, check out Arun Ulag’s hero blog “FabCon and SQLCon 2026: Unifying databases and Fabric on a single, complete platform” for a complete look at all of our FabCon and SQLCon announcements across both Fabric and our database offerings.

TMDL View on the Web will be available in the coming weeks, providing a powerful code-first semantic modelling experience directly to your browser.

With this new capability, Power BI developers can script, modify, and apply changes to semantic model objects using the Tabular Model Definition Language (TMDL)—a human-readable code format that describes your entire semantic model as code—without switching to Power BI Desktop or downloading model files.

Watch the TMDL View on the Web demo in the March 2025 Update video.

What is TMDL View on the Web?

TMDL View on the Web is a new feature in Power BI that enables developers to view and edit semantic models as code directly in the browser using TMDL. It expands the Power BI web modelling experience by introducing a rich code editor for working with TMDL scripts—giving pro developers full transparency into the semantic model code and enabling more efficient workflows through code editing.

TMDL View was initially introduced in Power BI Desktop in January 2025 and became generally available in September 2025. Now, we’re introducing the same code-first modelling experience to published semantic models in the workspace—directly in your browser.

No downloads. No switching tools. Just seamless modelling on the web.

Key capabilities and benefits of TMDL View on the Web

TMDL View on the Web provides the following capabilities:

Explore your published semantic model metadata

Get full visibility into all objects and properties within your semantic model, including advanced properties not exposed in the standard UI.

To view the TMDL definition of any object, simply drag and drop it into the editor, or open the context menu and select “Script TMDL to Script tab” or “Script TMDL to Clipboard”.

This makes exploring and understanding your model structure faster and more efficient.

TMDL_View_on_the_Web_showing_the_TMDL_definition_of_a_semantic_model_object_incl

Figure: Exploring published semantic model metadata in TMDL View on the Web by scripting the TMDL definition of an object, including advanced properties not exposed in the standard modelling UI.

Enhanced development efficiency

TMDL View on the Web includes a modern code editor designed to boost productivity with built-in IntelliSense, multiline editing for bulk updates, search and replace capabilities and more. For example, you can use IntelliSense and multiline editing to assign a display folder to all Sales-related DAX measures in a single operation—reducing repetitive work and improving consistency.

You can also take advantage of AI-powered tools such as GitHub Copilot to assist with authoring TMDL scripts. For instance, you can script your model as TMDL, use an AI assistant to generate or modify code, and then paste it back into the editor to preview and apply changes—streamlining your modelling workflow even further.

TMDL_View_on_the_Web_code_editor_showing_IntelliSense-assisted_multiline_editing

Figure: Using multiline editing and IntelliSense in TMDL View on the Web to assign a display folder across multiple measures in a single operation.

Modify any semantic model property/object

Edit properties and objects directly in the browser—including advanced settings such as partition definitions or properties like isAvailableInMdx, which are not exposed in the standard modelling interface. This capability gives developers full control over their semantic models, enabling advanced configurations without relying on external tools or downloading the model.

TMDL_View_on_the_Web_showing_a_TMDL_script_where_the_isAvailableInMdx_property_i

Figure: Editing isAvailableInMdx property in TMDL View on the Web, with a side‑by‑side preview showing the before‑and‑after impact on the TMDL definition.

Increased reusability and collaboration

Easily share and reuse semantic model objects by sharing TMDL scripts.

For example, to reuse a Calendar table from another semantic model or from a centralized gallery such as TMDL gallery, copy its TMDL script, paste it into your target model, preview the changes, and apply them.

TMDL_Gallery_page_showing_a_Calendar_table_example_with_its_TMDL_definition_avai

Figure: TMDL Gallery page showcasing a Calendar table shared as a reusable TMDL script for reuse across semantic models.

TMDL_View_on_the_Web_displaying_a_TMDL_script_from_the_TMDL_Gallery_pasted_into

Figure: Reusing a TMDL script from the TMDL Gallery by pasting it into another model in TMDL View on the Web, previewing the changes, and applying them.

Key differences between TMDL View in Desktop and TMDL View on the Web

To help you understand how the web experience differs from the Desktop version, here’s a side-by-side comparison of key functional differences between TMDL View in Power BI Desktop and TMDL View on the Web:

Key Difference

TMDL View in Power BI Desktop

TMDL View on the Web

• View mode and Edit mode

  No distinct modes—changes can be made and applied to the model at any time

  Introduces two modes: View mode (to script and preview changes) and Edit mode (to apply them to the model), enabling safer experimentation before committing changes.

• Script persistence

  TMDL scripts are saved as part of the semantic model. A model may contain previously saved scripts.

  Scripts do not persist. They are discarded when the semantic model or browser is closed. Previously saved scripts were not displayed.

• Version history support

  Not available.

  Leverages workspace version history to restore previous versions of the semantic model if needed.

• Write permissions

  Not applicable. Desktop authoring does not rely on workspace permission levels.

  Requires write permissions on the semantic model to open and use the experience.

TMDL View on the Web will begin rolling out in preview over the coming weeks. Here's how to prepare:

Watch the demo: March 2025 Update video

Explore the documentation:

Tabular Model Definition Language

TMDL View

We look forward to seeing how you use TMDL View on the Web to accelerate your semantic modelling workflows and bring code-first development practices to your Power BI projects.

Questions? Leave a comment!

Original source

If you haven’t already, check out Arun Ulag’s hero blog “FabCon and SQLCon 2026: Unifying databases and Fabric on a single, complete platform” for a complete look at all of our FabCon and SQLCon announcements across both Fabric and our database offerings.

Power BI is the decision layer for millions of users because it doesn’t just visualize data—it standardizes meaning. Semantic models capture the definitions that businesses run on the measures people trust, the relationships that provide context, and the governance that keeps answers consistent. Microsoft Fabric IQ builds that same semantic layer, extending trusted Power BI definitions into a broader, connected context that can support analytics and AI with fewer gaps and less guessing.

Here’s how customers are already putting that vision into action:

“Power BI Copilot coupled with trusted data products have become the common language of insight across the enterprise — connecting teams, data, and decisions through a single, trusted analytics platform.”
Paul Wellman, Vice President, Enterprise Data & Analytics Platforms, TD Bank Group

We’re announcing new capabilities that make Power BI more open, more powerful, and more deeply integrated into the way you work while strengthening the semantic foundation that AI relies on to deliver consistent, trustworthy answers

What’s new in Power BI

Translytical Task Flows (Generally Available)

Power BI has evolved from a reporting tool to an operational surface. With translytical task flows, users can act directly from reports: update records, trigger workflows, or resolve data issues in real time. There’s no need to move into another system or submit a request and wait. Insight and execution now happen in the same place, at the same time. To learn more, visit our translytical task flows documentation.

Animated_demo_of_Translytical_Task_Flows_in_Power_BI_showing_a_user_taking_an_ac

Figure: Translytical task flows in Power BI—take action directly from a report to update records and move work forward without leaving the analytics experience.

Modern visual defaults (Preview)

With Modern visual defaults, new reports start out looking polished and consistent—so teams spend less time tweaking formatting and more time surfacing insights. The updated base theme aligns with Fluent 2 design, creating a professional visual style across charts, slicers, buttons, and tables right out of the box. Charts, slicers, and buttons also have style presets available, for quick style changes in a couple of clicks, ensuring everyone can present data with clarity and impact from day one.

Screenshot_of_a_Power_BI_report_page_using_Modern_visual_defaults_showing_KPI_ca

Figure: Modern visual defaults (Preview) in Power BI Desktop—updated Fluent 2 styling and a refreshed base theme give new reports a polished look by default.

Report Copilot for Mobile (Preview)

Whether you’re in a meeting or on the move, you can now ask questions using voice or text in the Power BI mobile app and receive instant answers or visuals from Copilot.

Screenshot_of_a_mobile_app_displaying_AI-generated_sales_data_for_top_managers_b

Figure: Copilot in the Power BI mobile app answers a question and generates a visual.

TMDL View in the Web (Preview)

Tabular Model Definition Language (TMDL) View on the Web is launching in preview soon, bringing a code‑first semantic modeling experience directly to the browser. Developers will be able to view, edit, and apply changes to all semantic model metadata using TMDL, enabling greater transparency, efficiency, automation, and more consistent model development.

Screenshot_of_a_data_modeling_interface_in_Power_BI_DAX_code_for_calculating_sal

Figure: Editing semantic model metadata in a code-first TMDL experience in the browser.

Direct Lake on OneLake (Generally Available)

Reduce refresh overhead and keep data in open formats. Power BI is standardizing open-data formats by adopting Delta Lake and Parquet to help you avoid vendor lock-in and reduce data duplication. Direct Lake storage mode accelerates time to data-driven decisions by unlocking incredible performance directly from OneLake.

Screenshot_of_a_software_interface_showing_a_New_semantic_model_dialog_box_for_c

Figure: Creating a semantic model with Direct Lake on OneLake.

Direct Lake on OneLake, now generally available, provides compatibility with OneLake security, more modeling features, and faster query performance. Refer to Direct Lake on OneLake documentation to learn more.

Table Visual Updates (Generally Available)

We’ve added polish where it matters most. You now have greater control over totals in table visuals, along with cleaner, modern default styles that improve readability. These enhancements may seem small—but across thousands of reports, they save time and elevate the user experience at scale.

Direct Lake calculated columns (Preview)

You asked, and we listened. Calculated columns for Direct Lake tables will soon be available in preview. When adding columns upstream isn’t feasible (such as when data preparation in OneLake is owned by another team) you can extend Direct Lake tables by creating calculated columns. We are also introducing the ability to make calculated columns user-context aware by dynamically responding to DAX functions including UserCulture(), UserPrincipalName(), CustomData(). This enables new scenarios like data translations, and we’re excited to see the creative ways the community will use this!

Screenshot_of_a_data_modeling_interface_showing_a_formula_to_calculate_Age_using

Figure: Creating a calculated column for a Direct Lake table in the semantic model

We’re just getting started—join us!

As you explore these updates, continue to follow this space for more deep dive blogs on these feature releases over the next couple of weeks.

We’d love to keep the conversation going with you. If you’re joining FabCon Atlanta 2026, build your agenda and attend the Power BI and Fabric sessions that map to your role. Come find us in the Community Lounge or at Ask the Experts to share what’s working and what else you want to see in the product.

If you’re not onsite, you can still participate. Post questions and feedback in the Power BI Community forums and influence the product roadmap by sharing and upvoting ideas.

Let’s shape the future of data and AI together.

Original source