Power BI News Today September 2025

This September 2025, Microsoft Power BI delivered one of its biggest updates yet, with a strong focus on modeling, AI, and productivity. The release introduced DAX User Defined Functions (UDFs) in public preview, letting report builders encapsulate logic and reuse it across models, a major step toward more modular and maintainable analytics.

Alongside this, the long-awaited semantic model editing in the Power BI Service is now generally available, meaning users can shape and refine data models directly in the browser without relying only on Desktop. The Tabular Model Definition Language (TMDL) view also became GA, opening the door for code-first editing and source-controlled workflows.

DAX, short for Data Analysis Expressions, is the formula language that powers calculations and measures inside Power BI. It’s often described as the “engine under the hood” that lets analysts create everything from simple sums and averages to highly complex time intelligence expressions. Up until now, developers often had to copy and paste the same pieces of logic into multiple measures or models, leading to redundancy and more room for error. With the new DAX User Defined Functions (UDFs), Power BI takes a major step toward cleaner and more maintainable modeling. UDFs allow you to write a calculation once, give it parameters, and then reuse it across multiple places, just like traditional programming functions. For example, instead of rewriting a rolling average formula dozens of times, you can define it once as a UDF and call it with different columns or time windows as needed. This not only saves time but also helps make sure that business logic stays consistent across reports and teams.

Beyond the convenience, UDFs open the door for more collaboration and standardization. In enterprise environments, entire teams rely on shared semantic models, and until now, keeping calculations consistent meant enforcing conventions manually. With UDFs, organizations can effectively “package” their best practices into reusable formulas, making it much easier for both beginners and advanced developers to build on top of a solid foundation. Pair this with the new semantic model editing in the Power BI Service and the TMDL view, and Power BI is becoming much more than a drag-and-drop report builder — it’s evolving into a robust modeling environment where structured, repeatable, and code-friendly workflows are possible. Together, these changes transform DAX from a tool that required manual discipline into one that actively supports modular design, reusability, and long-term maintainability.

On the AI front, Microsoft expanded Copilot capabilities: a new standalone Copilot experience is rolling out by default, with smarter search, descriptive captions, and the ability to save explorations into Pro workspaces. Reporting received a boost as well, with Performance Analyzer in web editing to diagnose slow visuals, and enhanced time intelligence functions like custom calendars and new totals. The update also marked a shift in visuals and integrations — Bing Maps is being phased out in favor of Azure Maps, NFC tag support on mobile went GA, and Power BI in Teams now opens in a dedicated window. Altogether, this release signals Microsoft’s push to make Power BI not just a reporting tool, but a deeply integrated analytics platform where modeling, AI, and collaboration converge.

Building on these advancements, the new Copilot experience is designed to be more than a simple assistant—it is becoming an active partner in analysis. Rather than limiting users to asking plain-language questions, Copilot now provides iterative, conversational guidance. Analysts can refine their requests step by step, with Copilot adjusting outputs in real time. For example, if a user asks for a sales breakdown by region, Copilot can generate the chart and then, when prompted, highlight only top-performing markets or overlay growth trends. This conversational loop helps both technical and non-technical users move from broad questions to precise insights without leaving the Power BI interface.

Another important enhancement is the ability to save and share Copilot-driven explorations directly into Pro workspaces. What used to be ad hoc, temporary analysis can now be preserved as part of the organization’s collective knowledge. This shift makes AI-generated insights not just useful in the moment but also repeatable and auditable. Coupled with descriptive captions and improved search logic, Copilot’s outputs are easier to understand and validate, which addresses one of the biggest hurdles in adopting AI for analytics—trust. Teams can now collaborate around these saved explorations, using them as a starting point for deeper modeling or reporting, rather than treating them as one-off experiments.

Together, these improvements mark a transition from Copilot being a novel productivity aid into a strategic analytics enabler. By weaving AI more tightly into the workflow and connecting its results to enterprise models and shared workspaces, Microsoft is moving toward a vision where Copilot isn’t just answering questions—it is actively shaping how organizations discover, refine, and communicate insights. This integrated approach makes it far easier to balance speed with governance, allowing businesses to move quickly while still maintaining consistency in their reporting standards.