Поделиться через


Overview of DAX Copilot data transformation (preview) in healthcare data solutions

Important

  • This is a preview feature.
  • Preview features aren't meant for production use and might have restricted functionality. These features are available before an official release so that customers can get early access and provide feedback.
  • To review the terms of service, see Healthcare data solutions in Microsoft Fabric.

Healthcare data solutions in Microsoft Fabric let you analyze Dragon Ambient eXperience (DAX) Copilot raw data using data insights from DAX Copilot. You can ingest and store DAX Copilot artifacts such as transcript data, audio files, and AI summaries in Fabric lakehouses with the DAX Copilot integration with Microsoft Fabric feature. With the DAX Copilot data transformation (preview) capability, you can now ingest and transform raw transcript data across different lakehouse levels and structure it for analytics, AI/machine learning modeling, and Power BI visualization. This capability helps healthcare professionals, researchers, and analysts use ambient clinical intelligence for advanced analytics, supporting collaboration, research and development, and AI innovation in healthcare.

DAX Copilot data transformation (preview) is an optional capability under healthcare data solutions in Microsoft Fabric. You have the flexibility to decide whether to use it, depending on your specific needs or scenarios.

To learn how to deploy, set up, and use this capability, see:

Conceptual architecture

The DAX Copilot data transformation (preview) capability uses the medallion lakehouse architecture explained in Data architecture and management in healthcare data solutions. Here's how this framework organizes and processes DAX Copilot data across the different lakehouse layers:

  • Landing zone: DAX Copilot data files land in the bronze landing zone. You must first Integrate DAX Copilot with Microsoft Fabric and set up the required configuration as a prerequisite. The files in the landing zone are stored in a complex folder structure that includes a date-based hierarchy and a type-based hierarchy.

  • Bronze: Also called the raw zone, this layer stores the source conversational data from DAX Copilot. The healthcare#_msft_dax_bronze_ingestion notebook ingests this data into a bronze table called DaxCopilotData. This table stores the DAX file content and metadata.

  • Silver: In this layer, the healthcare#_msft_dax_silver_ingestion notebook builds the silver model based on the bronze data. DAX Copilot transcripts are stored here in the DAXTranscripts table. This table stores the transformed transcript content for further AI-based enrichments.