Copilot, data agents and the new Power BI apps answer from your semantic model. Microsoft is direct about models that aren't prepared: "Without proper preparation, Copilot can generate less helpful and inaccurate results" (Microsoft Learn). The audit tells you what to fix in your model before your users start asking it questions.

What we check

The seven checks of our 7-point checklist, on your real model:

  • Names a person can read, and no duplicate field names across tables.
  • Descriptions on the tables, columns and measures people use most.
  • Technical fields (IDs, keys, sort-order helpers) hidden or left out of the AI's view.
  • Unused tables, columns and measures, traced through every measure, relationship and visual.
  • A clean star schema, without ambiguous or many-to-many paths the AI would have to guess between.
  • Formats and sort orders: percentages that show as 0.34, month names that run April, August, December.
  • Prep data for AI: what to put in the AI data schema, the AI instructions and the first verified answers.

Plus every finding of our Model Health Check, such as visuals that point to fields that no longer exist, risky relationships and slow DAX patterns, each with the objects it concerns.

What you get

  • A written report with every finding ranked by impact, starting with what to fix first.
  • The exact fix for each finding: the DAX to change, the Power Query step or the setting.
  • Your Model Health Check score and findings, so you can run the free check again after the fixes.

Audit + fix, with proof that no number changed

If you want, we then apply the fixes in a copy of your model that you approve, and set up Prep data for AI: the AI data schema, the AI instructions and a few verified answers. We run the same DAX queries on the model before and after and show the results side by side, so you can see that no number changed.

Scope and timeline

  • One semantic model per audit.
  • You send a Power BI template (.pbit: the model without any data rows), or we work in a session on your machine. Your data never leaves your machine.
  • Nothing in your model changes without your approval.
  • The report in English or Arabic. Any Arabic text for your model is supplied or approved by you.
  • The written report within 5 business days of receiving the model.

Price

  • Audit: AED 4,500 (launch price; list price AED 6,500).
  • Audit + fix: quoted after the audit, once we know how much there is to fix.
  • Gulf calendar add-on: AED 2,000. Hijri, Ramadan and Eid columns in your calendar table, and AI instructions for Gulf periods.

Prices are per semantic model.

Start with the free check

Export a .pbit from Power BI Desktop (File > Export > Power BI template) and drop it on the Model Health Check. It runs in your browser and nothing is uploaded. If the findings look familiar, the audit is the next step.

Questions

Do you need access to our data?

No. A Power BI template (.pbit) holds the model and the report layout but no data rows. If you prefer, we work in a session on your machine instead.

Do we need Copilot switched on for the audit?

No. The audit checks the semantic model itself. Copilot needs a paid Fabric capacity (F2 or higher) or Power BI Premium (P1 or higher); our licensing guide explains the options.

Can the report be in Arabic?

Yes, in English or Arabic. Arabic names, descriptions and AI instructions for your model are written with you or approved by you.

How long does it take?

The written report arrives within 5 business days of receiving the model.

What if our business runs on the Gulf calendar?

Add the Gulf calendar add-on (AED 2,000): Hijri, Ramadan and Eid columns in your calendar table, and AI instructions for Gulf periods.