Training · Power BI for Management
From data to decision, in two levels.
Hands-on Power BI training for people who manage with numbers. Start with the fundamentals or move on to DAX, Microsoft Fabric and analytical agents — always validating what the AI produces.
- In person or onlineAlways live and synchronous
- Up to 10 participantsPer cohort, in the room and on Teams combined
- Taught in PortugueseThe Copilot exercises are done in English
- ~70% practiceCases with synthetic data and fictitious names
- Environment includedAn account with a Power BI Pro licence and Microsoft Fabric capacity
Where to start
Three questions, one starting point.
Level 2 assumes you already import data with Power Query, build a star schema and create reports with filters and interactions. If you do not yet do all three, start with Level 1.
These are the Level 2 prerequisites. If you are between the two, start with Level 1.
The programme
Seven modules per level, sixteen hours each.
Each module ends with an exercise you solve on your own computer.
Level 1 · Fundamentals — Modern, interactive and AI-assisted reports
Turn scattered data into reports that support decisions: prepare the data, organise it in a model, choose the right visual for each message and share the result. Copilot and the Power BI analytical visuals are part of the work from the first report — and the course teaches you to check what the AI produces before you take it to a meeting.
- Who it is for
- Professionals from any area, with or without business intelligence experience: managers, controllers, accountants, sales and HR specialists, consultants.
- Prerequisites
- Everyday use of Excel. No Power BI knowledge needed.
- You leave with
- Three complete, published reports, one per case: Web Analytics, Business Analytics and Telecom.
The modules
- 011h30
Power BI, Microsoft Fabric and business intelligence concepts
See the topics
- Power BI Desktop, Power BI Service and Power BI Mobile: what each one is for.
- Power BI within Microsoft Fabric: OneLake, workspaces, capacities and licences, explained for people who build and consume reports.
- Excel and Power BI: what each does well, and the risks of using spreadsheets as a reporting system.
- The architecture of a solution: extract, transform and load (ETL), semantic model, report, sharing.
- Where AI comes in: Copilot, analytical visuals and agents — and why the quality of the answers depends on the quality of the model.
- 023h00
Importing and preparing data with Power Query
Import and transformation exercises
See the topics
- Data sources: Excel, CSV, folders, PDF, XML, web pages, SharePoint and OneDrive.
- Column and row transformations: data types, splitting and merging columns, promoting headers, fill, transpose and unpivot.
- Combining queries: append and merge, and the join kinds.
- Columns from examples and conditional columns.
- Data profiling and quality: distribution, errors and empty values.
- Good practice: clear names, readable steps, reference and duplicate queries.
- 031h30
Data model and first measures
Web Analytics case
See the topics
- Fact and dimension tables; star and snowflake schemas.
- Relationships: cardinality and filter direction.
- The calendar table, and why to turn off auto date/time.
- First DAX measures: SUM, COUNTROWS, DIVIDE; explicit and implicit measures.
- Names, descriptions and measure folders: a model that people and Copilot can read.
- 044h00
Storytelling and report design
Web Analytics case; do's and don'ts exercise
See the topics
- Choosing the visual for the message: comparison, change over time, composition, distribution, relationship and deviation from a target.
- Essential visuals: card, matrix, bar, line and combo charts, maps.
- Button and input slicers, date picker; visual, page and report-level filters; interactions between visuals.
- Filter, group and summarise: the essential analysis requirements.
- Design principles: placement, rule of thirds, proximity and contrast; page wireframes before building.
- Themes: base theme, light and dark themes, custom JSON themes in the company colours.
- Mobile page layouts.
- Do's and don'ts exercise: which version of each visual communicates better, and why.
- 052h00
Interactive reports
Business Analytics case (a chain of sports stores)
See the topics
- Bookmarks, buttons and page navigation.
- Page and bookmark navigators.
- Page tooltips and help tooltips.
- Drillthrough pages.
- Hierarchies and drill-down.
- 063h00
AI applied to analysis and reporting
Telecom case (customer churn)
See the topics
- Copilot in the report: summarise a page or a report, ask questions about the data, ask for a visual to be explained.
- Create report pages with Copilot and ask for content suggestions; the narrative visual.
- Key influencers: which variables explain customer churn.
- Decomposition tree: break a result down by dimensions, with AI-suggested splits.
- Anomaly detection in time series.
- Forecasting in line charts: what the model assumes, how to read the confidence interval and when not to trust it.
- Reading AI critically: answers that vary for the same question, checking against visuals built by hand, what Copilot does not do (forecasts, causal explanations), language limits and capacity cost.
- 071h00
Publish, share and follow up
Publishing the three cases
See the topics
- Publishing to the Power BI Service; workspaces and roles.
- Sharing: links, org apps with audiences, export to PDF, PowerPoint and Excel.
- Email subscriptions, including summaries generated by Copilot.
- Alerts on the values of a visual.
- Publish to web: why not to use it with company data.
- Power BI Mobile.
- Course summary, final self-assessment and personal commitment.
The two days: Day 1: modules 1 to 3 and the first part of module 4. Day 2: the rest of module 4 and modules 5 to 7.
By the end, you can
- Tell apart the Power BI components (Desktop, Service, Mobile) and place Power BI within Microsoft Fabric.
- Import data from files, folders and the web and transform it with Power Query.
- Build a star schema with relationships and a calendar table, and write simple DAX measures.
- Choose the right visual for each management message and apply report design principles.
- Build interactive reports with bookmarks, navigation, tooltips and drillthrough pages.
- Use Copilot to summarise reports, answer questions about the data and create pages — and check every answer against the model.
- Use the analytical visuals (key influencers, decomposition tree, anomaly detection, forecasting) and read their results and limits.
- Publish and share reports and set up subscriptions and alerts.
Bibliography
- Few, S. (2013). Information Dashboard Design: Displaying Data for At-a-Glance Monitoring (2nd ed.). Analytics Press.
- Knaflic, C. N. (2015). Storytelling with Data: A Data Visualization Guide for Business Professionals. Wiley.
- Puls, K., & Escobar, M. (2021). Master Your Data with Power Query in Excel and Power BI. Holy Macro! Books.
- Financial Times. Visual Vocabulary. github.com
- Microsoft. Overview of Copilot for Power BI. Microsoft Learn. learn.microsoft.com
- Microsoft. Power Query documentation. Microsoft Learn. learn.microsoft.com
Level 2 · Advanced — DAX, Microsoft Fabric and analytical agents
For people who already build reports and need to answer harder management questions: budget control, financial ratios, customer segmentation, forecasting. DAX stays at the centre; around it, Microsoft Fabric, models prepared to be read by AI, agents that answer questions about the company data and forecasting models. AI writes DAX and answers questions — participants learn to read and validate what it produces.
- Who it is for
- Professionals who already build reports in Power BI and want to deepen their analysis: controllers, financial analysts, accountants, reporting leads, consultants.
- Prerequisites
- Level 1 or equivalent experience: importing data with Power Query, building a star schema, creating reports with filters and interactions.
- You leave with
- A complete project in your own workspace: the data of a chain of sports stores in a lakehouse, a Direct Lake model prepared for AI, a data agent, a forecast and a distributed app.
The modules
- 013h00
DAX: context, CALCULATE and iterators
Sports Stores and Shoe Shop cases
See the topics
- How DAX evaluates an expression: row context, filter context and context transition.
- Calculated columns, measures and calculated tables: when to use each.
- Aggregations and iterators: SUM and SUMX, AVERAGEX, MINX, MAXX, COUNTROWS.
- CALCULATE and the filter modifiers: FILTER, ALL, ALLEXCEPT, REMOVEFILTERS, KEEPFILTERS.
- Variables and DIVIDE: readability and performance.
- Navigating relationships with RELATED and RELATEDTABLE; ranking with RANKX and TOPN.
- DAX query view: test a measure as a query before saving it to the model.
- Copilot in DAX query view: write and explain queries. The course rule: only what you can read and validate goes into the model.
- 022h00
Time intelligence and reusable calculations
Sports Stores case; the inactive relationships case
See the topics
- A calendar table with CALENDAR and ADDCOLUMNS, marked as a date table.
- Year-to-date, same period last year, growth rate, three-month moving average, comparison with previous and following periods.
- Fiscal and custom calendars with calendar-based time intelligence.
- Inactive relationships and USERELATIONSHIP: order date and ship date, civil and fiscal calendar.
- Calculation groups: apply year-to-date and year-over-year to any measure without duplicating it.
- Visual calculations: calculations over the visual matrix, such as percent of total, moving average and custom totals.
- DAX user-defined functions: write a business rule once and reuse it across measures.
- 033h00
Management cases with DAX
Sales team, trial balances, budget control, ABC classification, simulation
See the topics
- Sales team: parent-child hierarchies with PATH and analysis by level.
- From trial balances to a report: semi-additive measures for balances, income statement and balance sheet, financial ratios.
- Budget control: compare a monthly budget with daily sales, allocating and accumulating at different granularities.
- ABC (Pareto) classification of products or customers.
- What-if simulation: parameters and the sales break-even point.
- 043h00
Microsoft Fabric for analysts: from file to Direct Lake model
Sports Stores continuous project; advanced Power Query exercises
See the topics
- Fabric architecture for Power BI users: OneLake, workspaces and capacities; lakehouse, warehouse and semantic model.
- Create a lakehouse and bring data in: shortcuts to SharePoint and OneDrive, file upload, Dataflows Gen2.
- Advanced Power Query in Dataflows Gen2: irregular data, regional settings, errors and missing values, parameters and M functions, data from a web API.
- Copilot in Dataflows Gen2: describe transformations in natural language and ask for a query summary.
- AI functions in Dataflows Gen2: classify customer comments, extract sentiment and summarise text.
- A Direct Lake semantic model over the lakehouse; when to choose Import, DirectQuery or Direct Lake.
- Scheduled refresh and monitoring.
- 052h30
Analytical agents: preparing the model and building data agents
Sports Stores continuous project
See the topics
- Why the quality of the answers depends on the model: names, descriptions, relationships and explicit measures.
- Prepare the model for AI: AI data schema, AI instructions, verified answers and approval for Copilot.
- Copilot across several reports and models at once, and the summary of a semantic model.
- Fabric data agent: choose the sources (lakehouse, warehouse, semantic model), write instructions and example queries, test and publish.
- Evaluate an agent: a battery of questions with known answers, an error log, corrected instructions and a new round.
- Using the agent in Microsoft Teams, Microsoft 365 Copilot and Copilot Studio (demonstration).
- Governance: row-level security, read-only users, capacity consumption, answers limited to read queries.
- Demonstration: agents that build semantic models through the Power BI MCP server, and what they require (version control, human review, tests).
- 061h30
Forecasting and predictive analysis
Continuous project; subscription and sales cases
See the topics
- What Power BI does out of the box: forecasting in line charts and anomaly detection. What Copilot does not do: forecasts.
- RFM (recency, frequency, monetary value) customer segmentation with DAX.
- AutoML in Fabric: train a sales forecasting or subscription churn model with the wizard and read the quality metrics.
- Generate predictions, write them to the lakehouse and use them in a Direct Lake report.
- Communicating a forecast: interval, expected error and the decisions it can support.
- 071h00
Distribute, automate and govern
Continuous project
See the topics
- Org apps with audiences; workspace roles.
- Row-level security: define roles and test them.
- Deployment pipelines: development, test and production.
- Alerts and actions from reports with Activator.
- From report to action: Power Automate and Power Apps inside the report; updating data from the report with translytical task flows (demonstration).
- Course summary, final self-assessment and personal commitment.
The two days: Day 1: the DAX language (modules 1 to 3). Day 2: Fabric, agents and forecasting (modules 4 to 7).
By the end, you can
- Write and read DAX measures that depend on filter context, with CALCULATE, iterators and variables.
- Use Copilot to write and explain DAX queries and validate the result before it goes into the model.
- Build time-based analyses (year-to-date, same period last year, moving averages, fiscal calendars) and reuse logic with calculation groups, visual calculations and user-defined functions.
- Solve management cases with DAX: hierarchies, trial balances and financial ratios, budget control at different granularities, ABC classification and simulation.
- Create a lakehouse, load and transform data with Dataflows Gen2 and build a Direct Lake semantic model on top of it.
- Apply generative AI functions to text (classify, extract sentiment, summarise) in a data transformation.
- Prepare a semantic model for AI and create, test and evaluate a Fabric data agent.
- Train a forecasting or classification model with Fabric AutoML, generate predictions and use them in a report.
- Distribute content with org apps, manage the lifecycle with deployment pipelines and automate actions from reports.
Bibliography
- Ferrari, A., & Russo, M. (2019). The Definitive Guide to DAX (2nd ed.). Microsoft Press.
- Puls, K., & Escobar, M. (2021). Master Your Data with Power Query in Excel and Power BI. Holy Macro! Books.
- Nogueira, N. (2026). Gestão com IA, transformar empresas inteligentes (in Portuguese).
- Russo, M. (2026, 22 September). Where we stand on AI and DAX in 2026. SQLBI. sqlbi.com
- Microsoft. Prepare your data for AI. Microsoft Learn. learn.microsoft.com
- Microsoft. Fabric data agents. Microsoft Learn. learn.microsoft.com
- Microsoft. Low-code AutoML in Fabric. Microsoft Learn. learn.microsoft.com
Side by side
Level 1 builds. Level 2 goes deeper.
| Level 1 · Fundamentals | Level 2 · Advanced | |
|---|---|---|
| Focus | From scattered data to a published report | Harder management questions |
| DAX | First measures: SUM, COUNTROWS, DIVIDE | Filter context, CALCULATE, time intelligence, calculation groups |
| AI | Copilot in the report and analytical visuals | AI-ready model, data agents, AutoML |
| Platform | Power BI Desktop and Power BI Service | Microsoft Fabric: lakehouse, Dataflows Gen2, Direct Lake |
| Cases | Web analytics, sports stores, telecoms | Trial balances, budget, ABC, sales team; one continuous project |
| You leave with | Three published reports | A complete project, from lakehouse to app |
| Duration | 16 hours · 2 days | 16 hours · 2 days |
Next cohorts
Choose a level, or take the full path.
Level 1 · Fundamentals
€590 + VAT
per person · 2 days of 8 hours
27 and 30 October 2026
Register for Level 1Full path · Levels 1 + 2
€1,150 + VAT
per person · 4 days of 8 hours
Both levels, 32 hours, 7% less than buying them separately.
27 October to 6 November 2026
Register for the full path →VAT (23% in Portugal) is calculated at checkout, from the billing country and VAT number.
Full refund up to 7 days before the start. After that, no refund — but you can send someone else in your place.
Private cohorts for companies: talk to us →
Frequently asked questions
Before you register.
Do I need a Power BI Pro licence?
No. Swell gives each participant an account in its training tenant, with a Power BI Pro licence and access to a workspace on Microsoft Fabric capacity with Copilot enabled. Level 2 also includes lakehouse, Dataflows Gen2, data agents, AI functions and AutoML.
What equipment do I need?
A Windows computer with an up-to-date Power BI Desktop. Power BI Desktop does not exist for macOS.
Can I take Level 2 without Level 1?
Yes, with equivalent experience: importing data with Power Query, building a star schema and creating reports with filters and interactions. The «Where to start» check above helps you decide.
Do we work with real data?
With management cases built on synthetic data and fictitious names. About 70% of the time is practice, and each module ends with an exercise on your own computer.
What language is the training in?
Portuguese. The Copilot exercises (and, in Level 2, the data agent exercises) are done in English, the only language Microsoft officially supports; the difference between asking in Portuguese and in English is covered in the exercise itself.
How is it assessed?
An initial and a final self-assessment, a short quiz at the end of each module, the trainer’s review of the exercises and a closing personal commitment: what you will build at work.
Want to bring your team to Power BI?
Either level can be delivered as a private cohort for your company.
Talk to us →
