When I first learned how to use Power BI in 2019, I was an inexperienced college intern with only a passing knowledge of BI and data management. The process of creating and publishing a Power BI report was very simple. I imported data from SQL, did some simple data prep in Power Query, wrote a few simple CALCULATE statements, and dragged them into visuals. Once I was done, it was as simple as publishing to a workspace and my job was done.
It was not easy to wrap my head around some of the more advanced bits of Power BI, such as M code and variable DAX, but overall Power BI was a straightforward, drag-and-drop-heavy tool that I felt I could teach to anyone. There was no need to understand DevOps, data governance, or enterprise architecture. The service layer existed, but I mostly felt that it was there to share and manage reports. The tool rewarded people who could import a dataset, clean it up, and communicate it with insightful visuals. Back then, Power BI felt like a tool I was learning. Today, it feels more like a vast, evolving ecosystem.
If I were that same intern entering the world of Power BI today, I would have a much different experience. The biggest shift in Power BI was not the introduction of new features or quality-of-life updates; it was the pivot by Microsoft to Fabric. Thanks to Fabric, Power BI changed from a standalone BI tool to a pivotal piece of a broader analytics management platform. Tools such as OneLake and DirectLake fundamentally changed the data context in which Power BI operates. Performance tooling improvements like Performance Analyzer have made it easier than ever to monitor and debug slow or broken reports. Deployment Pipelines have formalized the previously tedious Dev/Test/Prod workflows. Now, many of the BI concepts I studied in college have become much more accessible in Power BI, such as lakehouses, warehouses, and data pipelines.
This was not just a simple upgrade for Microsoft; it was a total shift of identity. Instead of thinking about how to just build a report, I now think about where the data should live in Fabric, how transformations should be handled upstream, how reusable my semantic model needs to be, and how my work exists within my employer’s data governance policies. I have access to many new tools that would allow me to do my work with faster data, and I can be more creative with how I communicate data. In a sense, I got to grow up as a developer alongside Power BI.
There have also been plenty of advanced features that have been added over the years to make development easier and more rewarding. In 2019, if I wanted to create time intelligence variations such as Month-to-Date and Year-to-Date calculations, I had to do a lot of repetitive copying and pasting while changing very little in the DAX. Today, calculation groups make this a breeze. Field parameters have made dynamic axis switching and measure selection much easier. Visual calculations have made it easier than ever to add simple, one-time-use calculations to your report without cluttering your model with more DAX measures.
Perhaps the biggest change of all has been the introduction of the Power BI Project format, or PBIR. Version control was always a pain in Power BI, with your PBIX file being its own monolithic data world. Working collaboratively often included sending different versions of a report back and forth several times to a coworker. Now PBIR changes your report to be a structured, inspectable file that enables Git-based collaboration. Now we have access to true version control, line-by-line change tracking, and deeper integration into DevOps workflows.
This change brings Power BI closer to the world of software development, which is what has caused me to stop and think about who the audience is for Power BI in the modern day. PBIR is another sea change that shows Microsoft knows that the average user of the Power BI tool has changed. No longer is a single user taking a legacy Excel report and elevating it. Now some BI developers are operating like software developers and pushing changes that are affected by a vast Fabric ecosystem. But there are signs that even that system of development may only be temporary.
While the PBIR file format may make developers’ lives easier, the real advantage it has is that it allows AI to operate on reports more smoothly. Anyone reading the monthly Power BI updates over the past year knows that Microsoft is heavily investing in and pushing the use of CoPilot in Power BI and in all its other tools. Natural language querying, automated visual generation, and measure suggestions are just a few examples of how Microsoft is shaping its tool to be at the forefront of AI and BI. The metadata and inspectable components of the PBIR format will unlock a whole new level of AI capabilities.
PBIR is the perfect representation of where Power BI is at today. On the one hand, it is a new tool for advanced developers to improve their BI ecosystem. On the other hand, it enables AI to further ease the burden of new users and lowers the barrier to entry for them. It perfectly embodies how over the 2020s Microsoft has been able to simultaneously lower the floor and raise the ceiling of Power BI development and create a tool usable by almost any user. The interns of tomorrow will not be stuck copying formats across files. That will be done in seconds by CoPilot. Their first tasks may involve using features that took my years to master. That is a testament to not only the growth of Power BI, but the broader BI world. The new users of today are not stepping into just a self-service reporting tool, they are stepping into an entire analytics ecosystem. From there, Power BI can take them as little or as far as they want.
Not sure on your next step? We'd love to hear about your business challenges. No pitch. No strings attached.
