Skip to main content

v20260605 (Release)

· 4 min read

This release delivers major new agentic and AI capabilities across workflows, data management, and model building along with extensive stability and UX improvements throughout the platform.

New Features & Major Enhancements 🚀​

Matrix Editor Redesign​

  • Spaces are highlighted on the matrix for clearer visual feedback when selected.
  • When selecting a space, the relevant information (Data, Row Headers, Column Headers, Corner Headers) is shown in the matrix editor.
  • Row numbers and column letters are now available in edit mode to make matrix setup easier.

Dashboard​

  • When configuring a filter, a default value can be set for the it. This allows the data in the widgets to load immediately when opening a dashboard.

Workflow​

  • A new AI Helper is now available on the Workflow page beside the toolbox, enabling users to build the structure of their workflows through natural language, reducing setup time.
  • Users can now run or stop a workflow from view mode, without switching into edit mode.
  • Azure Fabric pipeline integration is now available as a workflow data source and connection type.
  • A new ERP Integration activity enables workflows to interact directly with connected ERP systems as part of an automation sequence.

MCP & Agentic Capabilities​

  • Agentic ERP integrations are now supported, with incremental sync and dimension updates from NetSuite connections.
  • Una now supports MCP authentication using OAuth 2.1, including multi-tenant support, enabling secure connections from external AI tools i.e. Claude.
  • The MCP client configuration is available from the tenant modal making it faster to connect AI clients.

Staging Tables Enhancements​

  • Users can now generate a staging table directly from an Excel file, streamlining data setup.
  • Users can download an import template and import a flat file into a staging table from the UI.

Model Builder Enhancements​

  • Reserved Tags are now supported in the Model Builder, automatically applying predefined tags to models so users no longer have to add them manually after a model is built.

Multiple Time Hierarchies​

  • The Multi-Dimensional Engine now supports multiple time hierarchies in a single model query, enabling more advanced time-based analysis.

Relational Planning​

  • Pivot tables now support data write-back, allowing users to edit and save cell values directly back to the source. Write-back applies to non-aggregated data, and users can control whether data writing is enabled or disabled depending on what is displayed.

Notifications​

  • Users will now receive an in-app notification when new actions are assigned to them, so nothing falls through the cracks.
  • Added ability to toggle disable/enable push notifications.

AI Widget​

  • AI widgets can now refresh when data is saved to a matrix. This allows them to be used in data input canvas rather than just for reporting.

Layout Engine (Beta)​

  • Users can now export a dashboard to a PDF as requested.

Fixes & Improvements 🛠️​

Matrix​

  • Fixed internal hyperlinks being broken and unable to open the file path explorer.
  • Fixed copy/pasting JSON code from the Advanced editor in one matrix to another resulting in "no rows to show."
  • Fixed Drill Up on a parent column not correctly collapsing nested drill-downs below it.

Hierarchy Editor​

  • Fixed Cut + Paste as Child onto self deleting the member instead of moving it.
  • Fixed Alternate Hierarchy Operator and Tag fields not being independently editable from the Primary Hierarchy.
  • Fixed the "Name already exists" error appearing incorrectly in the Alternate Hierarchy editor.
  • Default operator is now + when creating a new dimension member.

Datatable​

  • Fixed Drill Up not working in the Datatable widget.

Staging Tables​

  • Fixed adding columns to a staging table not updating the corresponding import definition history table.

Import Definitions​

  • A client-facing warning is now shown when a slicer is set to "Everything" and when a dimension slicer is not configured in Import Definitions, reducing the risk of unintended data selection.