In My Desk, previously, users selected a governance role such as Owner, Steward, or Custodian from a dropdown, and responsibility data was displayed across multiple tabs such as Databases, Tables, Table Columns, Files, File Columns, Reports, Report Columns, Terms, APIs, and API Attributes. This required navigation across tabs and did not provide a single consolidated view.
In the current version, a Summary tab is introduced as the default view, which displays responsibilities across modules such as Data Catalog, Business Glossary, Tags, ROPA, Service Desk, Reference Data Management, Governance Apps, Policy Manager, Governed Data Query, Projects, and Data Quality in a single screen. The dropdown selection is removed from this view, and all sections are expandable for detailed visibility. Existing detailed tabs remain available for module-level access. Additionally, My Desk is now enabled for all license users by default, and Application Security is updated to correctly structure My Desk under My Resources.
Affected Users: Administrators, Data Stewards, Business Users, and Analysts.
π For more details, see My Desk | Dashboard for User Responsibility & Access.
In Elasticsearch, previously, when users performed a search and applied filters that resulted in no data, clearing the search term using the cross mark removed the search input but did not restore or display the previously applied filters. This caused confusion as users could not view or reuse their earlier filter selections.
In the current version, filter state is retained even when search results return no data. When the search term is cleared, previously applied filters are displayed correctly, allowing users to continue refining results without reapplying filters. This ensures consistent filter behavior and improves usability during search operations.
Affected Users: Administrators, Business Users, and Analysts.
π For more details, see Search | Filters | Filter Visibility Issue
In Browser Extension, previously, when Guest access was enabled, all users were first directed to the Guest experience after entering the URL and Email. Licensed users then had to navigate through the Guest interface and select an additional option before reaching the credential screen for full access. This created unnecessary steps and delayed access for licensed users.
In the current version, after entering the URL and Email, the extension evaluates whether Guest access is enabled. If enabled, users can choose to continue as a Guest or as a Licensed User. Licensed users can now directly access the credential screen without navigating through Guest features. If Guest access is disabled, the credential screen opens directly. If an active session already exists, full features load automatically.
Affected Users: Licensed Users, Guest Users, Administrators.
π For more details, see Enhanced Login Experience for Guest and Licensed Users
In askEdgi, previously, users could not download data from workspace objects, and download behavior was inconsistent across catalog and non-catalog objects, analysis results, and chat views.
In the current version, download is enabled for all workspace objects, including catalog tables, non-catalog tables, AI-generated tables, uploaded files, and derived datasets. Access is controlled through role-based configuration using the key askEdgi.workspace.download.allowed.roles. The system evaluates user roles to allow or restrict downloads, and the behavior is applied consistently across the workspace object list, analysis results, chat responses, and recipe execution outputs. Downloaded data reflects the final evaluated dataset as displayed in the interface.
Affected Users: Administrators, Authors, Viewers
π For more details, see Workspace | Enable data download for all workspace objects.
In Workspace, previously, logs were displayed in chronological order, with older entries at the top and latest entries at the bottom, requiring users to scroll to view recent activity.
In the current version, logs are displayed in reverse chronological order, where the latest log entries appear at the top by default. New logs also appear at the top during refresh or updates, while log content, formatting, and metadata remain unchanged, ensuring consistency across reloads.
Affected Users: Administrators, Support Teams, and Analysts.
π For more details, see Workspace Logs.
AI Prompt Questions module, previously, the AI Prompt Questions module allowed administrators to configure contextual prompt questions, but there was no centralized mechanism to manage reusable AI instruction templates. As a result, generated descriptions were often not in the expected format, lacked consistency, and required users to manually provide additional prompts each time.
In the current version, a new configuration layer called AI Instructions is introduced within the AI Prompt Questions screen. This enables administrators to create, manage, and control reusable instruction templates that are applied during AI description generation. The AI Description Generator now uses only active instructions, ensuring descriptions are generated in a consistent, clear, and governed format without requiring manual input from users.
Affected Users: Administrators, Authors, Viewers
π For more details, see AI Generative Description | AI Instructions configuration.
In Projects, previously, when a parent task was moved to a Done status, its child tasks stayed in their current statuses. This could make the parent task appear complete even though some child tasks were still open. Users had to update each child task manually.
Now, when a parent task is moved to a Done status, the system asks whether the related child tasks should also be moved to the same status. If confirmed, all child tasks are updated automatically. If canceled, only the parent task is updated.
Affected Users: Administrators, Authors, Viewers
π For more details, see Option to Update Child Tasks When Parent Task Is Moved to Done.
In Data Quality Rules, previously, the rule names allowed multiple special characters. This created inconsistencies with naming validations used in other modules and could affect search behavior, sorting, and overall standardization.
In the current version, unsupported special characters are restricted in Data Quality rule names, while the period (.) character remains supported for naming flexibility. Validation rules are now aligned with naming conventions used across the application. Existing rule names created before this change continue to work as expected, while new or updated rule names must follow the revised guidelines.
Affected Users: Administrators, Authors, Data Quality Rule Owners, Data Stewards
π For more details, see Updated Naming Rules for Data Quality Rule Names.
In Data Anomaly Detection settings, previously the lower-level objects did not inherit from parent objects. Tables continued to use the connector or global settings even when schema-level custom settings existed, and table columns did not inherit table-level settings. In addition, anomaly result pages did not display complete object hierarchy details, and job logs did not clearly indicate which configuration source was used during execution.
In the current version, anomaly detection settings now support hierarchical inheritance across object levels. Tables inherit settings from Schema, Connector, or Global configurations based on precedence, while table columns inherit settings from Table, Schema, Connector, or Global configurations. Configuration pages now display clearer parent setting indicators. Anomaly result pages include separate columns for Connection Name, Schema Name, Table Name, and Table Column Name, and job logs now clearly identify the configuration source used during execution.
Affected Users: Administrators, Data Stewards, and Data Governance Users
π For more details, see Data Anomaly Detection | Enhanced Inheritance, Object Visibility, and Job Log Details.
In Question Wall, previously, AI descriptions were generated only once after all assigned questions were answered. If a new AI prompt question was added later, answering that question did not trigger regeneration. Users had to manually select Generate Description to refresh the content, which could result in outdated descriptions that did not reflect the latest responses.
In the current version, when a new AI prompt question is added and answered, the system re-evaluates whether all assigned questions are completed. If all assigned questions are answered and auto-generation is enabled, the description is automatically regenerated using the metadata configured in ai.generatedesc.object.metadata.config. The newly generated description replaces the previous version automatically, with no manual intervention required.
Affected Users: Administrators, Authors, Stewards, and Business Users
π For more details, see Automatic Regeneration of AI Descriptions After New Question Responses.
In the Query Sheet, previously, query result downloads were processed via a single API call. For large query results, this caused latency issues and delayed downloads.
Now, query result downloads are handled by a background job, improving performance, reducing latency, and providing a more reliable download experience for large result sets.
Affected Users: Administrators
π For more details, see Background Job Support for Query Sheet Downloads.
In the Impact Analysis summary, previously, when a user added or deleted a source object, the Last Updated By field was cleared and displayed as blank. This made it difficult to identify who performed the latest update.
Now, when a user adds or deletes a source object in Impact Analysis Summary, the Last Updated By field is updated correctly to display the user who performed the latest change.
Affected Users: Administrators, Authors
π For more details, see Corrected Last Updated By Field.
In the Job Workflow, previously, users could execute only one Job Workflow at a time. When multiple Job Flow Names were selected, the Execute option was disabled.
Now, users can select multiple Job Flow Names and execute them together in a single action.
Affected Users: Administrators, Authors
π For more details, see Bulk Execution Support for Multiple Job Workflows.
In Service Desk Templates, previously, when no template was mapped to the CTA button, the system displayed a generic error message. Now, the system shows a detailed message: βThere is no Template mapped to this Call to Action (CTA) button. Connect with the System administrator to map at least a global template to this CTA button.β
Affected Users: Administrators, Authors, Data Quality Rule Owners, Data Stewards
π For more details, see Improved Error Handling for Unmapped CTA Templates.