Microsoft has added a preview that uses Copilot in Fabric to generate Real-Time Dashboards from a natural-language description. The feature works from a selected database in an Eventhouse, and the documentation also describes creation from a Kusto Query Language (KQL) Queryset. For teams monitoring streaming and historical data, it changes the starting point from an empty dashboard to a proposed operational view that still needs review and editing.
The feature sits in the Real-Time Intelligence workload. Copilot already helps write KQL queries and edit individual dashboard tiles. The preview adds a higher-level workflow in which the user describes the dashboard goal, the data to prioritise and the preferred organisation of the view.
From prompt to dashboard
In the Eventhouse workflow, the user opens New Real-Time Dashboard and selects Generate with Copilot. In the Copilot pane, the user describes the intended view and can add details about relevant signals, comparisons, tables and layout. Copilot then shows how it has interpreted the request, allowing the user to refine the scope before creating the dashboard.
When the user selects Start dashboard creation, Fabric asks for a dashboard name and workspace location. The created item is a Real-Time Dashboard that can be edited, rather than a static chart proposal. Users can change visual types, modify the KQL queries behind tiles and continue using Copilot to help with query or visual edits.
Microsoft’s example uses stadium operations. The prompt points Copilot to transit, gate, stadium zone and match-event data. The generated dashboard combines summary tiles, time comparisons, passenger breakdowns by station and transit line, gate and zone comparisons, and a table showing current gate load and queues.
That example also shows the practical boundary of the feature. The generated dashboard depends on the data model, the prompt and the measures Copilot creates. Microsoft says users should verify queries, relationships between data and the meaning of measures before relying on a generated dashboard for operational decisions.
Manual editing remains part of the work
Real-Time Dashboards still use tiles backed by KQL queries. A tile supports one visual representation, and the editor lets users choose between Copilot and manual KQL query authoring. Users can switch between those options.
Copilot for KQL can translate natural-language questions into queries and support follow-up prompts. It does not generate control commands. It also does not automatically run generated KQL queries; users run them at their own discretion.
Dashboards can also be created manually, built from existing KQL queries or restored from a dashboard file. Supported data source options include Eventhouse or KQL Database, Ontology, Azure Data Explorer and Azure Monitor. Real-Time Dashboards can use pages, parameters, live refresh, sharing and JSON export.
The dashboard documentation states that Microsoft-managed keys encrypt dashboards and dashboard-related user metadata at rest. Sharing remains permission-dependent. Users can share an entire Real-Time Dashboard or a specific visual, including options for the visualisation and the query that supports it.
Tenant and capacity conditions
The preview has Fabric prerequisites. It requires a workspace with Microsoft Fabric-enabled capacity. Copilot in Fabric also requires the tenant switch to be enabled by an administrator.
Microsoft lists paid capacity requirements of F2 or higher, or P1 or higher. Trial SKUs are not supported. The F2 or P1 capacity must be in a supported Fabric region.
There is an additional regional condition for tenants or capacities outside the US or EU. Copilot is disabled by default in that case. A Fabric tenant admin must enable the setting that allows data sent to Azure OpenAI to be processed outside the capacity’s geographic region, compliance boundary or national cloud instance.
Microsoft points users to Copilot in Fabric pricing for billing information. The dashboard-generation documentation does not state a separate price for this preview.
Turning an operational prompt into an editable Fabric dashboard raises questions about data modelling, KQL review, live refresh, capacity and Copilot governance, and these are the kinds of implementation details covered in ECS's Data topic alongside Microsoft Fabric, Power BI, SQL Server and Dataverse. Sessions and workshops give practitioners room to compare dashboard design, query validation and capacity choices with people running Fabric in production. Bring the monitoring scenarios and tenant constraints that are hard to settle from documentation alone.
This article was written by the ECS editorial team with the assistance of Amal Agent, run.events' AI assistant, which supported the research, drafting and proofreading.