How to automate report updates in Copilot in Power BI

João Barros 15 de August de 2026 4 min read

This tutorial shows how to configure Copilot in Power BI to automatically generate update notes and tasks after a data refresh — useful for teams that need to communicate report changes without writing them manually. I explain why and show a practical flow, including triggers and simple prompts.

Prerequisites

Account with licensing that supports Copilot in Power BI.
Report published to Power BI Service with dataset configured for refresh.
Permissions to edit the report and create a Power Automate flow (if needed).
Basic knowledge of Power BI Service, Power Automate and Copilot prompts.

Step 1: Define the automation objective

First, clarify what you want Copilot to generate automatically: a summary of the main changes, a list of KPIs that changed, or assigned tasks. That defines the prompt you will use. For example, choosing “generate a 3-point summary about sales variations and highlight KPIs above 10%” is a clear objective.

Step 2: Prepare the report to feed Copilot

Ensure the report has clear markers: measure titles, column descriptions and a Bookmark or dedicated Page with the visuals that Copilot should analyze. Copilot uses metadata and report context, so descriptive names help.

Step 3: Create a standard prompt in the report

Open the report in Power BI Service, go to the Copilot pane (or the prompts field) and create a base prompt that the flow can reuse. A clear prompt reduces ambiguous responses.

Prompt example:
"Update summary: analyze the 'Resumo Vendas' page and return 3 main points, KPIs changed >10% and suggested action. Output in bullet points."

Step 4: Configure the automation trigger (Power Automate)

Use Power Automate to trigger Copilot generation after a dataset refresh. Create a flow with the trigger "When a data-driven alert is triggered" or "Schedule - Recurrence" combined with a refresh check. Then add an action to call Copilot via connector (or the action that integrates Copilot/Power BI).

Example (pseudo-steps in Power Automate):
1. Trigger: Power BI - Dataset refresh completed
2. Action: HTTP / Copilot connector - send prompt (see step 3)
3. Action: Create file or message (Teams/Email) with the response

Step 5: Format Copilot's output

Decide how you want to receive the response: create a note in the report (use a text box on a page), send to Teams/Email, or save to a file on SharePoint. Format the prompt to request structured output (JSON, bullet points) to facilitate processing in the flow.

Prompt adapted for JSON:
"Analyze the 'Resumo Vendas' page and return JSON with: {summary:[], kpis_changed:[{name:,old:,new:,pct_change:,note:}], suggestions:[] }"

Step 6: Insert the response into Power BI or communicate to the team

In Power Automate, take Copilot's response (JSON or text) and insert it where useful: update a text indicator in the report via API, post to a team in Teams or send an email. If inserting into the report, use the Power BI REST API to update a text file or a source that the report reads.

Simplified example (pseudo-HTTP):
POST https://api.powerbi.com/v1.0/myorg/groups/{groupId}/datasets/{datasetId}/tables/UpdateNotes/rows
Body: {"rows":[{"note": ""}]}

Verify the result

Perform a manual dataset refresh and observe the flow. Confirm that Copilot returned the expected content (check Power Automate logs). Verify where the output was placed: report page, Teams message or SharePoint file. Test scenarios with small changes to confirm that KPIs are correctly identified.

Conclusion

Automating updates with Copilot in Power BI reduces manual work communicating changes after a refresh and speeds up detection of variations. Next steps: refine the prompt, adjust filters and add validations in the flow to handle unexpected responses. Tip: start with plain text output before requesting JSON to understand Copilot’s behavior.