Power BI Desktop + Copilot

Power BI Desktop + Copilot

From raw data to executive dashboards — the complete workflow, accelerated by Microsoft Copilot.

Most teams collect data. Few turn it into decisions. This course closes that gap. In three days you take raw, messy data and turn it into clean, executive-ready dashboards — the kind leaders actually use to run the business. You build one complete Power BI solution end to end: raw files, a clean model, and a live dashboard published in Power BI Service, refreshed daily and readable on a phone. Not a prototype, not an exercise. A real one. Every module is hands-on — around 70% of the three days is workshop — and the workflow is Copilot-first: Copilot drafts the DAX and the M, and you review, fix and refine what it writes. You leave with skills you can apply Monday morning.

What you'll be able to do

  • Build an automated data pipeline — from raw monthly files to a clean, analysis-ready model.
  • Create and validate DAX KPIs with Copilot — growth, YTD, achievement %, commission.
  • Publish, schedule, and present a live dashboard to leadership.
  • Clean and reshape data in Power Query — column names, dates, nulls, duplicates, split and unpivot — and combine a folder of monthly files automatically.
  • Model your tables as a star schema — single versus bi-directional relationships, a marked date table, and the clear names, descriptions and synonyms that make Copilot and business users get the right answer from it.
  • Write a Copilot prompt that carries context, data, intent and constraints — then run it through a review checklist before you trust the code.

Who this course is for

Data and BI analysts

You already produce the numbers. You want one model that answers the next question too, instead of another rebuild.

Finance and accounting teams

Month-end reporting eats days. You want a pipeline that swallows the new files and a dashboard that updates itself.

Sales teams and sales managers

Targets, achievement % and commission get argued over every month. You want one page where everyone reads the same number.

Operations and inventory teams

Stock moves faster than your reports. You want turnover and days of inventory on a live page, not in last week's spreadsheet.

HR and admin teams

Headcount, leave and training data sit in separate exports. You want them modelled once and published where leadership can read them.

Curriculum

  1. 01Day 1 — Prepare your data

    8 hours

    Turn raw, messy data into clean datasets ready for analysis.

    • Power BI architecture
      • Desktop, Service, and Gateway — what each does and when to use it.
      • Import vs. DirectQuery: tradeoffs in real projects.
    • Prompt engineering for Copilot
      • Anatomy of an effective prompt: context, data, intent, constraints.
      • Iterate and refine when Copilot's first draft isn't right.
      • Verify before you trust: a review checklist for AI-generated code.
      • Data privacy: what Copilot can see, and how to stay safe.
    • Power Query data cleaning
      • Standardize column names, formats, and whitespace.
      • Fix dates, nulls, and miscategorized values.
      • Split, merge, pivot, and unpivot to reshape data.
      • Detect duplicates, errors, and outliers.
      • Combine multiple monthly files automatically.
      • Copilot-first Power Query: describe the transform in plain language in chat, get the M back as text, then paste it into the Advanced Editor and refine it there. Copilot does not run inside the Power Query Editor in Power BI Desktop, so reading and correcting M is part of the work — which is exactly why Day 1 spends time on a review checklist.
    • Business data preparation
      • Sales: revenue, items, pricing, discounts.
      • Inventory: Beginning → In → Out → Ending.
      • Targets: monthly targets aligned to business keys.
      • Date dimension: the foundation for time-based analysis.

    Hands-on workshop

    Clean a real sales dataset, transform inventory data, prepare an Actual vs. Target table, and automate a monthly file pipeline — prompting Copilot to draft the M code, then reviewing and refining it yourself.

  2. 02Day 2 — Model & analyze

    8 hours

    Build the data model and write the formulas that produce real KPIs.

    • Star Schema modeling
      • Why Star Schema is the enterprise standard.
      • Single vs. bi-directional relationships.
      • Build models that stay fast as data grows.
    • DAX for analytics
      • Calculated columns vs. measures — when to use each.
      • CALCULATE, the engine behind every advanced metric.
      • SUMX, DIVIDE, FILTER, and other essentials.
      • Time intelligence: YTD, MTD, MoM, YoY, rolling periods.
      • Copilot-first DAX: Copilot drafts around 80% of your measures from plain-language prompts in DAX query view — you validate each one against a total you worked out by hand, then debug and optimize it.
      • Where Copilot gets DAX wrong: variables, calculation groups, and newer functions. Ask it for a query rather than a bare measure, and check every result before it reaches the model.
    • Business KPIs
      • Sales: revenue, quantity, average price, growth.
      • Inventory: turnover and Days Inventory Outstanding (DIO).
      • Performance: Achievement % and variance.
      • Ranking, top-N, and trend analysis.
    • Special: Sales Commission system
      • Tiered commission structures (3% / 5% / 7%).
      • Commission by product, category, or salesperson.
      • Performance dashboard with rankings.

    Hands-on workshop

    Build a complete 3-page executive dashboard plus a working Sales Commission analytics system — extending a supplied commission starter model into tiered rates and salesperson rankings, generating the core DAX measures with Copilot, then validating and tuning them.

  3. 03Day 3 — Deploy & share

    8 hours

    Move from your laptop to the enterprise. Make reports run themselves.

    • Deployment workflow
      • Publish reports from Desktop to Service.
      • Workspaces: structure and governance.
      • Build and distribute Power BI Apps to leadership.
      • Track adoption with usage analytics.
    • Refresh & data gateway
      • Schedule automatic daily refresh.
      • Install the Gateway for on-premises data.
      • Troubleshoot failures and configure incremental refresh.
    • Mobile & presentation
      • Design layouts that work on phones and tablets.
      • Export to PowerPoint with executive polish.
      • Present live, interactive dashboards in meetings.
    • Copilot in Power BI Service
      • Generate draft report pages from a plain-language prompt.
      • Create executive summaries and smart narrative visuals in one click.
      • Let leaders ask their own questions of a report in plain language, through the Copilot chat pane. This is Copilot, not the older Q&A visual, and it needs the same paid capacity.
      • Prepare the model so Copilot answers correctly: descriptions, synonyms, hidden fields, clear names, one meaning per field. Copilot reads your model's metadata, so an untidy model gives confident wrong answers.

    Hands-on workshop

    Publish your dashboard, configure refresh, generate an executive summary with Copilot, and present it to a live audience.

Tools you'll actually use

  • Power BI Desktop on Windows — free to install, and where you build
  • Power Query and the M language
  • The star-schema data model and DAX
  • DAX query view with Copilot — where Copilot drafts your measures
  • Copilot in Power BI — access provided on our training accounts during the course
  • Power BI Service in a web browser — workspaces, apps, usage analytics
  • On-premises data gateway and scheduled refresh
  • Power BI Mobile layouts and export to PowerPoint

What you'll build

Automated monthly data pipeline

Day 1's workshop, ready to reuse: a real sales dataset cleaned, inventory reshaped into Beginning → In → Out → Ending, an Actual vs. Target table prepared, the date dimension that every time-based measure depends on, and a folder of monthly files that loads itself — with the M drafted by Copilot in chat and reviewed by you in the Advanced Editor.

Executive dashboard with a Sales Commission system

Day 2's workshop: a 3-page executive dashboard on a star-schema model, plus a working commission analytics system built out from a starter pattern into tiered rates and salesperson rankings, every measure validated against the raw data.

A published, self-refreshing report

Day 3's workshop: you publish your dashboard to Power BI Service in our training workspace, configure a daily scheduled refresh, add a mobile layout, and generate an executive summary with Copilot. The published copy lives on our training tenant and comes down after the course — what you take home is the .PBIX file with the mobile layout inside it, and the Service guidebook that walks you through the same steps on your own tenant.

The files you'll practise on

  • Sales

    Line-item sales — revenue, items, pricing and discounts, plus unit cost, which Day 2 needs to compute cost of goods sold for inventory turnover and Days Inventory Outstanding. It arrives raw, with inconsistent column names, stray whitespace, broken dates and miscategorized values, so the Day 1 cleanup is real work rather than a demonstration.

  • Inventory

    Stock movement in the Beginning → In → Out → Ending shape. You reshape it on Day 1 with split, pivot and unpivot, then use it on Day 2 for turnover and Days Inventory Outstanding (DIO).

  • Target

    Monthly targets aligned to business keys, so Day 2's Achievement % and variance measures have something honest to compare against, and Day 1's Actual vs. Target table has somewhere to come from.

  • Commission

    Salesperson, product and category sales behind the Special session on Day 2 — tiered commission at 3% / 5% / 7%, commission by product, category or salesperson, and the ranking dashboard.

What you take home

  • Working .PBIX filesComplete end-to-end project files — the pipeline, the model, the measures and the dashboard, exactly as you built them in class.
  • Reference datasetsSales, Inventory, Target and Commission, the same four you work with across all three days.
  • Reusable templatesPower Query scripts and DAX patterns you can lift straight into your own reports.
  • Copilot prompt libraryTested prompts for data prep, DAX, and dashboards — the ones that worked in class, not a generic list — plus the finished M and DAX for every exercise, so you can complete every lab with or without working Copilot access.
  • Dashboard templatesExecutive-grade starter layouts, so the next report begins from a page that already reads well.
  • Service guidebookA step-by-step deployment reference: publish, workspace, app, gateway, scheduled refresh.
  • CertificateFormal recognition of completion from ALL IN ONE STATION, stating 24 hours.
  • Q&A supportInstructor access after the course ends.

What you'll need

  • Basic Excel — SUM, VLOOKUP and PivotTables. No Power BI experience required.
  • A 64-bit Windows 10 or Windows 11 laptop for each participant. Power BI Desktop is free to install but exists only for Windows — there is no Mac version. Tell us before the course and we can arrange a Windows machine.
  • For an in-house class you provide the training room, a projector or large screen, the participant laptops, and stable internet for all three days — Copilot and the whole of Day 3 run in the cloud. We provide the instructor, the four datasets, the exercise files, all course materials, the certificates, and Copilot access on our training accounts. You do not need to buy any Microsoft licence for the class itself.
  • To keep using Copilot at work afterwards, your organisation needs a paid Fabric capacity (F2 or higher) or Power BI Premium (P1 or higher) with the Copilot tenant setting on, and you need contributor access to at least one workspace on that capacity. A Power BI Pro or Premium Per User licence on its own is not enough, and trial capacities do not count. Because Laos sits outside the United States and the EU data boundary, your admin must also switch on the setting that lets data sent to Azure OpenAI be processed outside your region — without it Copilot stays off by default. Check with your IT before you book.
  • To publish into a shared workspace, share a report, distribute an app or run a scheduled refresh other people depend on, you need a Power BI Pro licence. Whoever builds and publishes always needs Pro, whatever the capacity. The leaders who only open the report need Pro too, unless the workspace sits on Fabric F64 or larger, or on Premium — there they can read it on the free licence.
  • Prompts are typed in English: Microsoft supports Copilot in Power BI officially in English only. Instruction is in Lao; course materials and the software interface are in English.

Where this can take you

  • Data or business intelligence analyst

  • Financial or management accountant

  • Sales analyst or sales manager

  • Operations, inventory or supply chain officer

  • HR or admin officer who owns the reporting

Questions about this course

Do I need Power BI experience before this course?

No. The course assumes basic Excel — SUM, VLOOKUP and PivotTables — and nothing more. Day 1 starts with what Power BI Desktop, the Service and the Gateway each do, then goes straight into cleaning data in Power Query. If you have taken our Excel Intermediate or Advanced course, the star schema and DAX on Day 2 will already feel familiar and you will move faster. If you have not, you still finish Day 3 with a published dashboard. Day 2 is the steepest day of the three. You will not be writing tiered commission logic from a blank page — we build it together from a starter pattern, and you leave with the file.

Is Power BI Desktop free, and what does it run on?

Power BI Desktop is a free download from Microsoft, and everything you build on your own laptop — Power Query, the model, DAX, the report — costs nothing. It runs on 64-bit Windows 10 or Windows 11 only. There is no Mac version, and Microsoft has not announced one. A Mac can run it inside a Windows virtual machine, but that setup is yours to maintain and we do not support it in class. Tell us before you book and we can arrange a Windows machine for the three days.

What can Copilot in Power BI actually do, and where does it work?

In Power BI Desktop, Copilot builds and edits report pages from a prompt, writes and explains DAX in DAX query view — including measures you then add to the model — summarises your semantic model and drafts measure descriptions. In Power BI Service it also drafts report pages, creates summary and narrative visuals, and answers questions about a report in plain language. One limit matters for Day 1: Copilot does not sit inside the Power Query Editor in Power BI Desktop. Microsoft's native Copilot for Power Query lives in Dataflow Gen2 in the Service. So in this course you have Copilot draft the M in chat, then paste it into the Advanced Editor and refine the steps yourself. That is also why Day 1 spends time on a review checklist: Copilot is a fast first draft, and Microsoft documents where it goes wrong on exactly the Day 2 workload — variables, calculation groups and newer DAX functions — so checking it is a skill rather than a formality. One more thing to expect. Parts of Copilot are still preview: the DAX query view Copilot has to be switched on under Options and Preview features in Power BI Desktop, the Desktop Copilot settings pane is still labelled preview, and Microsoft changes what Copilot can do month to month. We teach the version that is live on the day of your class and tell you what has moved since the last intake.

What licences does my organisation need to keep working this way after the course?

Read this part before you book. Two separate licences are involved and neither comes with Power BI Desktop. Copilot first. A Power BI Pro or Premium Per User licence on its own is not enough. Microsoft requires your organisation to hold a paid Fabric capacity — F2 or higher — or Power BI Premium P1 or higher, with the Copilot tenant setting switched on, and you need contributor access to at least one workspace on that capacity. Trial capacities do not count, and Premium Per User workspaces are not supported. Because Laos sits outside the United States and the EU data boundary, your admin must also switch on the setting that lets data sent to Azure OpenAI be processed outside your region — without it Copilot is off by default no matter what has been bought. One thing keeps the cost down, and it is worth knowing before you rule this out: only the workspace you point Copilot at has to sit on the capacity. Your reports and semantic models can stay in ordinary Pro workspaces. A company can buy the smallest paid capacity, put a single workspace on it, and use Copilot in Power BI Desktop against models published anywhere. Publishing and sharing next. Whoever builds and publishes needs Power BI Pro, whatever the capacity — that covers publishing into a shared workspace, sharing a report, distributing an app, and owning a scheduled refresh other people depend on. The leaders who only open the report need Pro as well, unless the workspace sits on Fabric F64 or larger, or on Premium, where they can read it on the free licence. A free licence still covers building in Power BI Desktop and saving to your own My Workspace, including a refresh schedule there, and installing the on-premises data gateway costs nothing. During the three days you work on our training accounts, so Copilot is there for every exercise. Back at the office, if your company has no Fabric capacity, you will write the DAX and the M yourself instead of prompting for it — everything else you learned still works, and the course deliberately teaches you to read, fix and validate the code rather than only to ask for it. Send us your company's Microsoft licence position before you book and we will tell you plainly what you would need.

What language is the course taught in?

Instruction is in Lao. Course materials, the Power BI interface, the function names and the error messages are in English — that is how the software ships, and reading them is part of the work. Copilot prompts are written in English as well: Microsoft supports Copilot in Power BI officially in English only, and prompts in other languages may or may not return a sensible answer. We teach the prompt patterns in Lao and you type them in English.

What if someone misses a day?

Tell us as early as you can. The three days stack: Day 2 models the data you shaped on Day 1, and Day 3 publishes the model you built on Day 2, so a missed day is harder to make up here than on our Excel courses. We send the exercise files and the finished .PBIX for that day, and the instructor runs a catch-up session before the next class day.

How is this different from the Excel Advanced course?

Excel Advanced does much of the same thinking inside a workbook: Power Query, a star schema in Power Pivot, DAX measures, a dashboard in a file that lives on your laptop and gets emailed around. This course leaves Excel. You build in Power BI Desktop, then publish to Power BI Service — workspaces, an app for leadership, a daily scheduled refresh through a gateway, a mobile layout, and Copilot inside Power BI itself. Take Excel Advanced if your reporting stays in Excel files. Take this one if leadership needs a live dashboard that refreshes itself. You do not have to take Excel Advanced first, but if you have, Day 2 will move quickly for you. The extra 100,000 LAK buys Day 3 — publishing, workspaces, apps, the gateway and scheduled refresh — and the platform that keeps your report running after you close your laptop.

What do I take home, and what support do I get afterwards?

The working .PBIX files from all three days, the Sales, Inventory, Target and Commission datasets, reusable Power Query scripts and DAX patterns, the Copilot prompt library we test in class along with the finished M and DAX for every exercise, executive dashboard templates, a step-by-step Power BI Service deployment guidebook, and a certificate of completion from ALL IN ONE STATION stating 24 hours. Everything you take home is yours to use and adapt inside your organisation. It is not licensed for resale or for teaching outside it. You can also message the instructor after the course ends.

We want this for our team — what do we provide, what does it cost, and can you run it on our own data?

The price shown is per participant. For a whole team booked in-house we quote per class — tell us your headcount and your dates and we send the figure in writing before you commit. The fee covers the instructor, the four datasets, the exercise files, all course materials, Copilot access on our training accounts for the three days, and the certificates. It does not cover any Microsoft licence your organisation buys to keep working this way afterwards. You provide the training room, a projector or large screen, one Windows laptop per participant, and stable internet — Copilot and the whole of Day 3 run in the cloud. Class size is capped, and the cap comes from our training capacity rather than from the room: every Copilot exercise runs against one Fabric capacity, and past a certain number the responses slow down for everyone. Ask us for the current number when you book. On running it with your own data: often, yes — but tell us early. The three days run on four datasets that carry through every workshop, so swapping in your own tables means rebuilding the pipeline, the model and the commission exercise around them. We need a sample and a scoping call at least three weeks before the course, we will say plainly which exercises we can rebuild on your data and which we cannot, and where the rebuild is substantial we quote it separately. The option that costs nothing: work the class exercises on our datasets and bring a copy of your own file to build alongside them — no notice needed for that. Either way bring a copy rather than the original, and remove any columns you are not allowed to share. One privacy point before you decide. What you send to Copilot — your prompts and your model structure — is processed by Microsoft's Azure OpenAI service, which for a capacity in this region runs outside the country. Day 1 covers what Copilot can and cannot see so you can make that judgement yourself. If your data cannot leave your organisation at all, do the Copilot exercises on our sample datasets and build your own file by hand.

Interested in this course?

Message us and we'll honestly tell you whether it's the right fit. No charge, no obligation.

WhatsApp