Most BC customers we meet already have Power BI. It is just not always the case that anyone opens the reports. Typically because they were built as a demo, show figures that do not match the trial balance, or because management still gets the same Excel sheet from the bookkeeping team every Monday.
Here is how we build Power BI on Business Central, so that the figures match and the reports get used.
How Power BI and Business Central fit together
There are three routes from BC to Power BI, and the choice decides how much work you have afterwards.
Benefits of the built-in integration
Microsoft provides ready-made Power BI apps for Business Central covering, among other things, finance, sales, purchasing, inventory, inventory value, projects and production. The connector is already installed in BC, the semantic models follow a star schema, and the reports can be embedded directly on the pages in BC. It is the right place to start — and often enough for finance.
Dataflow architecture
When the ready-made apps fall short, you fetch data through Business Central's APIs — the standard API v2.0 or custom API pages from your AL extensions. We use custom API pages when you have your own tables or fields that need to be included. Avoid building reports directly on old OData web services from pages: they are slow, and they break when someone changes the page. With large data volumes, the data should be placed in an intermediate layer, for example a Fabric lakehouse, rather than having Power BI pull all entries at every refresh.
Organisations that implement Power BI effectively with Business Central typically see a 40% reduction in report creation time and a 35% increase in data-driven decision making across departments.
Setting up the Power BI environment
The set-up does not take long, but a few choices here are hard to change later.
Licences
Power BI Desktop is free to build in. To share reports, both the person who publishes and the people who read need a Power BI Pro licence — or the report must sit in a Premium or Fabric capacity. Microsoft's Power BI apps for BC are free, but also require a paid Power BI licence.
Connect Business Central to Power BI
Run the guided set-up Connect to Power BI in BC, install the apps you need from Microsoft Marketplace, and enter the environment and company as parameters on the semantic model. Use one workspace per app, so you can control access per area. If you have several companies, decide from the start whether they go in one model or in separate ones.
Security and access control
Power BI does not know your BC permissions. A salesperson who is not allowed to see payroll accounts in BC can still see them in a finance report if nobody has thought about it. Use row-level security, for example per salesperson or department dimension, and control access to workspaces through security groups in Entra ID. Let the data source sign in with a dedicated account or app registration, not with an employee's own user.
Designing dashboards and reports that get used
The technical side is rarely the problem. The problem is that the report does not answer the question the recipient has.
Know the recipient
The managing director wants to see revenue, contribution margin and liquidity, and whether things are heading the right way. The warehouse manager wants to see which items are below minimum today. Those are two different reports. Ask the recipient which three figures they look for first — and build those.
Visualisation principles
What we stick to:
- Figures must reconcile with BC. If the controller cannot find the balance again in the trial balance, the report dies
- Use the structure of the account schedule for the income statement and balance sheet instead of reinventing it
- Few visuals per page. One bar chart and a table beat ten pie charts
- Slicers on period, company and department — and drill-through down to the entry
- Use BC's own names for fields and dimensions, so users recognise them
Reports per role
One report for everyone ends up being for no one. Examples of what we build:
- Finance: liquidity forecast, ageing analysis of receivables and payables, budget against actuals per department
- Sales: revenue and contribution margin per customer, item and salesperson — compared with last year
- Warehouse and operations: inventory value, turnover rate, items below minimum, delivery performance
- Projects: hours spent against budget and billable work that has not yet been invoiced
Advanced analysis techniques
Once the basic reports are running, three things typically give the most.
DAX measures
DAX is Power BI's formula language. This is where you make the calculations BC does not give you directly: year-to-date against last year, rolling 12 months, margin after discounts. Keep the measures in the semantic model, not in the individual reports, so that everyone uses the same definition of "revenue".
Forecasts
Power BI can make simple forecasts on time series directly in a line chart, and for more advanced models you can use Azure Machine Learning. This is useful for sales and liquidity forecasts — but only if there is enough clean history. Two years of data with a new chart of accounts in the middle does not give a credible forecast.
AI features
Key Influencers can show what drives, for example, a low margin, and Copilot in Power BI can answer questions about data in plain language. Both work best on a clean, well-documented model. If the model is messy, so are the answers.
Companies that implement advanced analytics with their Business Central data typically see a 15-20% improvement in forecast accuracy and a 25% reduction in inventory costs through optimised stock levels.
Self-service analytics
The goal is that the controller can answer the next question without ordering a new report.
A semantic layer people understand
It requires a semantic model with Danish field names, ready-made measures and relationships that are already in place. Then users can drag fields into their own report — or analyse the model directly from Excel, which most people in finance prefer. Microsoft's semantic models for BC are a good starting point. We extend them with your own fields and dimensions rather than starting from scratch.
Want your BC data in Power BI?
We build the semantic model, the API pages and the reports on your own BC data — and reconcile the figures with the trial balance before they go to management.