- Four stages: Excel is king, Power BI enters the scene, things start getting done properly, and talking to your data.
- Each stage solves something and brings something new: Excel is flexible, but everyone has their own version of the truth. Power BI makes sharing better, but adds manual work and licenses.
- AI in stage four rests on everything you build before it: data in one place, governed and documented.
Contents 7 sections
- ·The four stages in one table
- ·Stage 01, Excel is king
- ·Stage 02, Power BI enters the scene
- ·Stage 03, things start getting done properly
- ·Stage 04, talk to your data
- ·How do you tell which stage you're in?
- ·FAQ
The four stages in one table
Stages of the data journey
| Stage | What works | What hurts |
|---|---|---|
| 01 Excel is king | flexibility, no extra licenses | everyone has their own version of the truth, sharing and updates |
| 02 Power BI enters the scene | better visuals, easier sharing | manual work, overcrowded reports, licenses |
| 03 Things start getting done properly | automated backend, sources in one place | rising costs, governance becomes a serious matter |
| 04 Talk to your data | AI helps you ask questions and find answers | most companies aren't here yet |
Stage 01, Excel is king
Data lives in spreadsheets. We have pivots, we have summaries, sometimes even a nice chart. For a lot of things this works and will keep working. This is where ideas and the first questions are born: couldn't we see this better?
“Can't we look at this in a better way?”
What works
- Still better than nothing.
- Incredibly flexible.
- No extra license costs.
What hurts
- Everyone has their own version of the truth.
- Data visualization in Excel has its limits.
- Sharing, updating, adding more sources. Each one hurts in its own way.
Stage 02, Power BI enters the scene
We start using more advanced tools. Data gets a home, maybe just a SharePoint folder, but at least it's something. For some processes we already know which KPIs to track, and reports help us with that.
In our view, this is the ideal moment to do it properly.
What works
- Visuals get better.
- Sharing is much easier.
- The first outlines of something called data governance.
What hurts
- Still a lot of manual work.
- Overcrowded reports with no clear focus.
- License costs add up.
- Updating data still takes manual work, even though the way to automation is opening up.
Stage 03, things start getting done properly
Data volume and company growth pushed us to move on. The first data warehouses, the first fully automated reports. Less is more: the right KPIs and a clear focus. Governance stops being a buzzword and becomes a real need.
Data finally runs on its own.
What works
- Insights actually shape business decisions.
- A decent backend finally removes the manual work.
- More and more sources joined in one place.
What hurts
- It took a lot of work, and costs are slowly rising: storage, compute, maintenance, development, licenses.
- Governance is now a serious matter: access, data ownership, accountability.
- It's still easy to churn out piles of reports that do nothing for anyone.
How to get data from your own server into Power BI is covered in the guide to data on your own server. What the Microsoft Fabric data platform brings is covered in the Microsoft Fabric guide.
Stage 04, talk to your data
This is where the journey leads. Once your data is in one place, well governed and documented, AI can step in and start working for you.
AI finds the insights for you.
We picture a future where every part of the organization has instant access to the data it needs. AI helps people ask the right questions and find answers before problems turn into real problems. Most companies aren't here yet, but everything you build in the earlier stages is what will one day make this possible.
In Power BI, this role is played by Copilot, which answers on top of the semantic model. What it can do and how you prepare the model for it is covered in the guide to Copilot in Power BI.
How do you tell which stage you're in?
Go through the questions in order and stop at the first no.
- Is your most important number produced some other way than by hand in Excel? If not, you're in stage 01.
- Does the data in your reports refresh on its own, with no manual export or edits? If not, you're in stage 02.
- Is the data in one place, and does every number have a description and an owner? If not, you're in stage 03.
- Do people ask about the data in plain language and get answers they trust? If yes, you're in stage 04.
Not sure where you stand? Get in touch, we'll be glad to look at it together.
FAQ
Does every company have to reach stage 04?
No. But each stage builds on the one before it. AI in stage four needs data in one place, governed and documented, and that comes about in stage three.
What does AI need from our data?
Copilot in Power BI answers on top of the semantic model. Microsoft warns that without preparing the model, Copilot can misunderstand the data and return generic or inaccurate answers (Copilot in Power BI overview). You prepare the model for it with the Prep data for AI feature: by scoping the fields, and with instructions and verified answers.
Which stage are most companies in?
Most companies we meet are somewhere in the middle, in stage 02 or 03.
Where do you stand?
Data Review
€3,350excl. VAT
We find out which stage you're in and what it costs you, build one report for real, and you leave with a Data Plan for 3, 6 and 12 months.
