Microsoft Fabric Updates Blog

Decoupling Semantic Model for Mirroring Customers

Overview

Semantic models are evolving to work more seamlessly with Mirrored artifacts—giving you greater flexibility, control, and transparency when working with mirrored data.

Why are we decoupling semantic models from Mirroring Artifacts?

Historically, Mirrored artifacts were created with an automatically ‘coupled’ semantic model.

While convenient out-of-the-box, this approach limited how you could shape, interpret, and manage business rules—or coordinate changes between your data storage and its semantic definitions.

With the needs of our diverse customers in mind, we’re moving to a decoupled model—putting the power to define, version, and manage semantic models into your hands.

How will the decoupling process work?

The shift to decoupled semantic models will happen in two phases. Phase 1 has recently been launched. Phase 2 will be underway within the upcoming months.

Phase 1 – New Mirrored artifacts will no longer be created with a default, coupled semantic model. When you create a new artifact, you’ll have the ability to design and manage your own semantic model using the tools and frameworks that best meet your analytics needs. We are pleased to announce that this has already launched.

Phase 2 – Existing Mirrored artifacts will have their default semantic models decoupled. More details will be released in the near future.

How will decoupling impact customers?

Decoupling brings multiple advantages for Mirroring’s Customers.

With the decoupling of semantic models, you’ll experience a clear separation between your raw data storage and the business definitions applied to that data. This clean break means each layer—storage and logic—can evolve independently, giving you more flexibility to grow and adapt your workflows and reporting over time.

Decoupling also empowers you to build tailored semantic models that precisely fit your organization’s unique needs. Whether your requirements cross different business units, analytics teams, or specific use cases, you can now craft semantic models that best represent how you want to interpret and interact with your data, all while maintaining a reliable single source of truth.

Additionally, this approach unlocks the ability to layer multiple semantic models on top of the same dataset. This means you can serve a wide range of reporting, business intelligence, and data science needs—all in parallel—without impacting each other’s work or creating conflicts between different perspectives on the data.

Having the semantic layer decoupled also brings you fuller access to your raw mirrored data. You’re free to query it directly using SQL, BI tools, or data science platforms, taking advantage of advanced features like complex joins, window functions, and custom aggregations. This level of access removes previous limitations, allowing for more sophisticated analyses and reporting.

Finally, decoupling supports robust versioning for your semantic models. You can develop, test, and iterate on your business logic as needed, without affecting others or worrying about version conflicts with the data. This makes it much easier to manage changes, experiment with new definitions, and maintain consistency across your organization as your analytics needs evolve.

To learn more, feel free to explore Sunsetting Default Semantic Models – Microsoft Fabric blog post.

Related blog posts

Decoupling Semantic Model for Mirroring Customers

November 5, 2025 by Pam Spier

There’s no AI without data. Are you ready to invest in your future? Fabric Data Days offers 50+ days of immersive learning designed for data professionals at every level and students alike. Don’t miss your chance to gain practical experience, earn free certifications exam voucher, and connect with a global community of experts. Here’s what … Continue reading “Advance your career in Data & AI with Microsoft Fabric Data Days”

November 4, 2025 by Misha Desai

We’re introducing a set of new enhancements for Data Agent creators — designed to make it easier to debug, improve, and express your agent’s logic. Whether you’re tuning example queries, refining instructions, or validating performance, these updates make it faster to iterate and deliver high-quality experiences to your users. New Debugging Tools View referenced example … Continue reading “Creator Improvements in the Data Agent”