How to create an automatic executive summary in Copilot in Power BI

João Barros 24 de September de 2026 6 min read

This tutorial explains how to create an automatic executive summary in Copilot in Power BI that synthesizes key report points (KPIs, variances and probable causes). An automatic summary is useful for sharing quick insights with stakeholders and reducing the time spent interpreting charts. Instead of analyzing multiple visuals, decision-makers receive 5–7 sentences that highlight the essentials: performance, drivers and actionable recommendations.

Prerequisites

Account with access to Power BI and licensing with Copilot enabled. Without the appropriate license Copilot will not be available in the service.
Power BI report with a well-structured data model (tables, relationships and basic measures). Ideally a calendar table with at least 2 years of data for YoY analyses.
Basic knowledge of DAX and report pages with visuals showing KPIs. Knowing how to create measures and cards makes it easier to extract exact values (e.g.: Revenue, Cost).

Step 1: Prepare key measures (KPIs)

Copilot works best if essential measures are explicit. Create clear measures for the KPIs you want to include in the summary (e.g.: Revenue, Cost, Margin, YoY Growth). Measures should be readable and have descriptive names — this way Copilot can reference the metrics exactly. Concrete examples help: if annual Revenue is ~€1.2M, or if the expected margin is 30%, that appears in the generated texts.

Receita = SUM(Vendas[Valor])
Custo = SUM(Vendas[Custo])
Margem = DIVIDE([Receita]-[Custo],[Receita])
Crescimento_YoY = DIVIDE([Receita] - CALCULATE([Receita], SAMEPERIODLASTYEAR(Calendar[Date])), CALCULATE([Receita], SAMEPERIODLASTYEAR(Calendar[Date])))

Test the measures with cards in the report to see real values (for example, Revenue = €1,250,000, Cost = €850,000, Margin ≈ 32.0%, Crescimento_YoY = 5.8%). Exact values allow Copilot to write percentages and variances as numbers, not just generic text.

Step 2: Create a report page dedicated to the summary

Add a new page called 'Resumo Executivo' and arrange simple visuals: cards for KPIs, a column chart for trend and a table with the top 5 products/segments. Simplicity helps Copilot identify the relevant points — many complex visuals can confuse text generation.

For example, use cards to show Revenue, Margin and Crescimento_YoY; a column chart with months (last 12) and a table with Top 5 products by Revenue. This gives Copilot temporal context and concrete drivers (e.g.: the top 3 products account for 62% of the quarterly variance).

Step 3: Configure descriptors for Copilot

Use text boxes with short descriptions to guide Copilot on what to include (e.g.: analysis period, report context). These phrases help generate text more aligned with what you want. Include information such as the time horizon (quarter, year) and the target audience (executive team, sales).

Contexto: Análise trimestral de vendas por região. Inclui variações YoY e top 5 produtos.
Objetivo do resumo: Destacar KPIs críticos, alterações relevantes e potenciais causas.

Practical example: if the report is for the sales team, ask for a tone oriented towards actions (e.g.: increase stock of the 3 products that grew more than 20%).

Step 4: Invoke Copilot and ask for the executive summary

Open the Copilot pane inside the report (Copilot icon). Make a clear prompt indicating what you want: include KPIs, compare with the previous period, point out the 3 main drivers and suggest actions. Be specific to avoid generic responses — the more context and constraints, the better the text quality.

Prompt exemplo: 'Gera um resumo executivo curto (5–7 frases) da página "Resumo Executivo". Incluir Receita, Margem, Crescimento YoY, os 3 produtos que impulsionaram a variação e duas recomendações de ação.'

If you want exact numbers, ask explicitly: 'Include values and percentages with 1 decimal place'. This forces Copilot to present €1,250,000 instead of 'around 1.2M'.

Step 5: Refine and customize the generated text

Review Copilot's text and request refinements: change the tone (more formal or concise), add exact numbers or explain causes based on the applied filters. Use follow-up prompts to adjust details and confirm hypotheses (e.g.: seasonality, campaigns or stock-outs).

Prompt de refinamento: 'Torna o texto mais conciso e inclui percentagens com 1 casa decimal. Explica se a tendência é sazonal ou relacionada com categorias específicas.'

Also ask Copilot to justify recommendations with metrics: 'Recommends increasing stock for Product A — justify with 24% growth and 18% share of revenue'.

Step 6: Insert the summary into the report

Copy the final text into a text box on the 'Resumo Executivo' page or use Power BI's export/share functionality. If you want automatic updates, repeat the process after data refreshes or create a flow with Power Automate (advanced) that triggers a prompt and saves the result to a file or scheduled email.

Automation example: schedule Copilot to generate a monthly summary after the dataset refresh and email it to 5 stakeholders — this reduces report preparation time by ~50% in a small team.

Verify the result

Confirm that the summary includes the KPIs with correct values, compares the requested periods and correctly identifies the main drivers. Validate two or three variances by requesting detailed explanations from Copilot and comparing with the report visuals. If you find discrepancies, adjust the measures or descriptors and rerun.

Conclusion

With these steps you have a practical method to generate automatic executive summaries in Copilot in Power BI, useful for management meetings and quick sharing. Next steps: automate with Power Automate, create prompt templates for different audiences and keep measures consistently named for better Copilot results. Practical tip: store 3–5 standard prompts to speed up production of weekly and monthly summaries.