Launch Deal · 50% off 1st month ·

3.3 — Use Copilot to Create Pages, Suggest Content, Write Narratives and Summarize the Model

PL-300 Domain 3 · Skill 3.3

What You Need to Know

Copilot in Power BI uses the semantic model to build report pages and answer questions in natural language. It's on the exam in four places: creating a page, suggesting content, the narrative visual, and summarizing the semantic model.

Prerequisites

  • The report must be in a workspace on a paid Fabric capacity (F2 or higher) or Power BI Premium (P1 or higher). Trial capacities and Pro or PPU workspaces alone don't support Copilot.
  • The Fabric admin must enable the Copilot tenant settings. Depending on the region, the setting allowing data to be processed outside your geographic region may also be needed.
  • In Desktop, you select a workspace with a supported capacity for Copilot to use.
  • You need edit access to the report to create pages; readers can use Copilot to ask questions.

Create a new report page

Open the Copilot pane and ask, for example, "Create a page showing sales performance by region and product category for 2026". Copilot picks fields, visuals and a layout. Review what it built, then edit it like any page.

Suggest content for a new page

Ask "Suggest content for a new report page". Copilot proposes several page ideas based on the model (titles and the fields they would use). You pick one and Copilot creates it.

Narrative visual with Copilot

The Narrative visual writes a text summary of the page or the whole report, and updates when filters change.

  • Add the narrative visual, choose Copilot, then pick what to summarize and optionally a prompt such as "Give a bulleted summary of the top three insights".
  • You can reference specific visuals, or the whole report.
  • The text is regenerated when the report is filtered or refreshed.

Summarize the underlying semantic model

Ask Copilot "What's in this data?" or "Summarize the semantic model" to get a description of the tables, measures and the questions the model can answer. This helps report creators who are new to a model.

Make the model Copilot-friendly

Copilot's output is only as good as the model:

  • Clear table, column and measure names, descriptions and synonyms (lesson 2.1)
  • Star schema with correct relationships, and hidden technical columns
  • Prep data for AI: an AI data schema (choose which fields Copilot uses), verified answers and AI instructions that add business context

Key Concept: Copilot needs a supported paid capacity and tenant settings, and it works from the semantic model's names, descriptions and structure. It can build a page, suggest pages, write a narrative visual, and describe the model.

Exam Traps

  • Expecting Copilot in a workspace on Pro or PPU without a Fabric/Premium capacity, or on a trial capacity
  • Using the narrative visual when the requirement is a new full report page (ask Copilot to create a page)
  • Blaming Copilot for poor visuals when columns have cryptic names and no descriptions
  • Forgetting that the narrative visual updates with filters (it isn't static text)

Check your understanding

Practice Scenario

🔒 This practice scenario is part of Skillmarked Pro

The lessons stay free. Pro unlocks the other 22 PL-300 practice scenarios and the 50-question timed mock exam, plus everything for every other exam.

See plans

Launch Deal: first month €9.95, then €19.90/month. Enter code LAUNCH50 at checkout. Ends 1 December.

Cancel anytime. Or €199 a year, which saves 17%.

Not sure yet? Try 10 free mock exam questions.

Already Pro? Log in.