📋 Prerequisites
- Run inside a Microsoft Fabric workspace notebook
- The semantic model must be published to the same (or an accessible) Fabric workspace
- You need Build (or higher) permissions on the semantic model
semantic-link-labswill be installed automatically in the next cell
A scored Scorecard (0–100) with a rating (AI Ready / Mostly Ready / Needs Improvement / Not Ready) and a prioritised action list (Critical → Important → Recommended).
How to use
Upload to a Microsoft Fabric workspace notebook
Set dataset and workspace parameters in cell 5
Run all cells
Limitation:
It can not able to access Prep Data AI setup
It is checking the best practices documented at https://learn.microsoft.com/en-us/fabric/data-science/semantic-model-best-practices#prep-for-ai-make-semantic-model-ai-ready
11 Checks + 1 Bonus
1 Star Schema — M:M relationships, bidirectional cross-filter, isolated tables 15
2 Business-friendly naming — detects DIM_, FACT_, _AMT, all-caps, abbreviations 10
3 Object descriptions — coverage % across tables, columns, measures 15
4 Synonyms — inspects TOM for synonyms on tables/columns/measures 5
5 Implicit measures — numeric columns with Summarize By ≠ None 10
6 Duplicate/overlapping measures — semantic groups + near-identical names 5
7 Ambiguous date fields — multiple date columns without guidance 5
8 Hidden objects risk — hidden columns that break Verified Answers 5
9 Model complexity/bloat — visible helper measures, column/measure counts 5
10 Prep for AI — annotation scan for AI Schema, Instructions, Verified Answers + full manual checklist 15
11 Best Practice Analyzer — runs the full 60+ rule BPA 10
- Bonus: Measure Dependencies — shows what each measure depends on for AI Schema config —