Read-only audit

Atlas Supply Chain · Supply chain / logistics · sample

AI

This is a finished audit of a synthetic model.

Run the same static audit on your own PBIP in seconds, without an account.

Audit your model
63/ 100

Not ready for production

507085

This semantic model has critical structural flaws that produce incorrect results for key metrics. The use of two separate date tables makes cross-functional time intelligence impossible, and snapshot inventory data is aggregated incorrectly, leading to vastly inflated stock figures over time. Several percentage measures like supplier share and on-time rates are also calculated incorrectly. The immediate priority is to fix the '[Preferred Supplier Share]' measure, which currently provides misleading results when filtered. Following this, a strategic refactoring to a unified star schema with a single date table is essential for the model's reliability and usability.

43 findings · 13 critical · 7 verified fixes · 5 hygiene notes, not counted in the score

Tables

10

Measures

17

Relationships

13

Pages

–

no report layer

Critical

13

AI findings

15

7 fixes

Fix these first

Ordered by severity, confidence and how many visuals depend on the object.

By area

Sub-scores out of 100; click to filter.

Verified fixes

Rewrites checked against your model and reviewed for semantic equivalence.

The model

Shape, sources and the report layer.

Shape

Fact tables3
Dimensions5
Helper / parameter2
Relationships6 critical13
Bidirectional2
Date table markedyes

Data sources

SQL Server6
Power Query3
Excel workbook1

Report layer

Upload the PBIP with its .Report folder to get lineage, unused measures and page findings.

363 static checks · 33 firedAI review: 15 findings, 7 rewrites verified · $0.67Engine semantic-pargemini-2.5-pro+gemini-2.5-pro9/22/2026, 2:10:47 PM

Nothing from this audit is stored. How your data is handled