A dashboard built for a finance team looks fine until someone in a board meeting asks a question the underlying model was never designed to answer.

The visible part of Power BI, the charts and dashboards, is the smallest part of the work. The data model underneath determines whether the numbers a board sees are trustworthy. That includes how tables relate, how measures are defined, and how time intelligence handles Hijri and Gregorian calendars side by side. We build the model first and treat visualization as the final step.

Connecting to ERP and financial systems

Power BI connects to most ERP platforms, such as SAP, Oracle, Dynamics and Odoo, either directly or through a data warehouse layer. For Riyadh businesses the practical question is usually how to bring in Arabic-labeled master data cleanly, and how to handle multi-entity consolidation where entities report in different currencies or on different chart of accounts structures. That is a design decision made once and relied upon by every report built afterward.

Measures, not just visuals

A well-built model defines business logic once, in DAX measures, so that gross margin means the same thing on the CFO's dashboard and in the sales report. Where measures are recalculated inconsistently from report to report, numbers quietly disagree. The resulting loss of trust in the reporting does more damage than any missing visual would.

Governance and workspace structure

As Power BI usage grows past a handful of reports, workspace structure, access permissions and a naming convention stop being administrative details. They start determining whether people can find and trust the right report. We set this up deliberately from the start so it does not accumulate into sprawl.

A hypothetical example

Consider a group with many disconnected Power BI reports built by different people over time. Several answer the same question with different numbers because each was built on its own extract. Consolidating onto a single governed data model, with one definition of revenue and one of margin, lets the duplicate reports be retired.

Bilingual reporting

Power BI supports Arabic labels, right-to-left layout and dual-language reports from the same model. This matters directly for boards and staff who work primarily in Arabic. It is a deliberate build decision, not a default, and it rests on the same discipline as Arabic financial statement configuration.

Row-level security for sensitive figures

Not every user should see every entity's margin or every region's headcount. Power BI supports row-level security that filters what each user sees from the same report. That matters for a group where a regional manager should see their own performance without seeing another entity's, or where salary-linked figures must be restricted to HR and finance leadership. It is decided during modeling and connects to the same access discipline as data governance more broadly.

Data refresh and source connections

How Power BI gets its data depends on the source. Some sources support a live connection that queries the source directly, while others need scheduled refresh. We state this plainly at the outset, because users who expect live data from a report that only refreshes on a schedule lose trust in the same way they do with any undisclosed timing gap.