This month’s episode of Fabric Threads is a jam-packed mega-episode that covers updates from the past two months, as well as the biggest announcements from FabCon Europe.

Our team is back from Stockholm and still running on the excitement of being around so many other Fabric enthusiasts. While we were there, we were constantly on the lookout for the latest and greatest topics to discuss on this show, and we definitely didn’t come home empty-handed.

We’ve got a great round-up in store for you today.

Preview of Copilot for Semantic Models

One of the most exciting features discussed this month is the ability to ask Copilot questions directly about your semantic model in Power BI Desktop. If you’re familiar with tools like ChatGPT, this experience is similar but designed specifically for exploring data models within Power BI.

Copilot has already been available within Fabric and Power BI, allowing users to ask questions about their semantic model within the Power BI service or ask questions about your report, your lake house, or your warehouse within Fabric.

And now what this feature allows you to do is actually ask questions and generate insights on top of your semantic model, but within Power BI Desktop. You can interact with your data using natural language questions, significantly lowering the barrier for non-technical users.

Think of your marketing or sales teams—they can ask straightforward questions without needing to know how to write complex queries. This feature is designed to provide answers in the form of visualizations, not just raw data, making insights more accessible than ever before.

I’ve personally been using it every day and having a lot of fun with it.

It’s worth noting that since this feature is still in preview, it might not handle all types of questions perfectly right away. Users should expect a bit of a learning curve, especially with more complex data models, and stay tuned for improvements as Microsoft continues to refine this tool.

Public Preview of AI Skill

Moving on to what we believe is the best feature of the last two months—the AI skill feature is now in public preview. This new tool allows you to develop custom generative AI models directly within Fabric, offering a more flexible and personalized approach to AI.

Previously, users relied on the out-of-the-box Copilot AI model for their needs. Now, with AI skills, you can customize your AI models to answer specific domain-related questions. This means you can provide examples and instructions that guide the AI on how to interpret and respond based on your unique business context.

Imagine having data from your lakehouse or warehouse and setting up your AI skills to analyze marketing trends or operational efficiencies. This feature is particularly valuable for data scientists and developers who need to create AI solutions that go beyond generic responses and dive into the nuances of your organization’s data.

Of course, you can’t expect it to understand everything perfectly right away, and it will require ongoing training and prompt engineering to create the best version of your model.

Security remains a top priority here. AI skills operate using the developer’s credentials, ensuring that data governance is respected and that only authorized data is accessed. Given that this feature is still in preview, expect ongoing updates that will continue to hone its capabilities over the coming months.

Share Feature in Fabric AI Skill

The AI skill we just talked about above was released in August, and the very next month, Microsoft introduced a feature that allows you to share your custom AI models with other people in your organization.

This feature is a game-changer for teams looking to collaborate on AI projects, making it possible to work together seamlessly by sharing AI models and configurations.

Previously, sharing an AI skill was inconvenient and limited to more advanced users. Now, by simply configuring the appropriate permissions, data scientists and AI developers can easily distribute their models to business analysts or other stakeholders who need access. This advancement speeds up the development cycle and ensures that your entire team can contribute to refining and using these AI models.

Just a word of caution—ensure your permissions are set correctly when sharing these skills. You want to make sure that the people you’re sharing skills with have access to the underlying data that supports the infrastructure. Just like anything else Fabric and AI-related, it’s important to have your security in place before you start developing.

FabCon Europe Recap

There were so many incredible announcements dropped at FabCon Europe in Stockholm. One thing that we’ve realized in conversations with clients, though, is that sometimes it’s not immediately obvious the value these announcements may have.

So, we wanted to highlight the top 3 announcements we’re most excited about and give you some insight into the true business value they unlock.

Enhanced Real-Time Data Integration

Fabric now provides more real-time data ingestion, processing, and visualization features.

This new capability is designed for organizations that need real-time insights to make data-driven decisions on the fly. It’s particularly beneficial for industries where real-time insights are important, like retail and finance.

With improved data ingestion and processing, businesses can connect their ERP, financial, and operational systems to Microsoft Fabric, enabling instantaneous data visualization.

Ultimately, this helps organizations to make reliable, informed decisions in a split second. Streamlining the flow of live data helps them quickly respond to trends, mitigate risks, and seize opportunities more immediately.

Expanded AI and Copilot Capabilities

Microsoft continues investing heavily in AI, integrating Copilot across all areas of Fabric and Power BI. You can clearly see the proof in the updates we mentioned earlier in this podcast since they are all AI-related.

This means more powerful AI tools are now at the fingertips of users, making it easier than ever to generate insights and automate complex tasks without extensive technical know-how. Now, more employees are empowered to engage with and leverage data for innovation, increasing overall productivity and reducing reliance on specialized data teams.

GitHub Integration for DataOps

The Fabric team announced the integration of GitHub for DataOps in Fabric. This significantly enhances version control and collaboration capabilities of data projects.

This is definitely a highly technical feature that may not be applicable to every single Fabric developer, but it’s huge for organizations that are truly taking advantage of Fabric as their go-to data storage, ETL, analysis, and business tool.

Data teams that manage large-scale, iterative projects should be particularly excited about this update because it allows them to manage and track changes in their data pipelines and models more efficiently and reliably.

This will also enhance collaboration across teams and you’ll be able to roll out changes to Fabric models and components much faster and more consistently.

Looking Ahead

With so many new features and enhancements, it’s clear that Microsoft Fabric is positioning itself as a powerhouse for AI-driven data solutions. The focus on low-code and no-code functionalities, combined with deep integration with AI tools, is democratizing data analytics like never before.