Lorem ipsum dolor sit amet consectetur adipiscing elit. Quisque faucibus ex sapien vitae pellentesque sem placerat. In id cursus mi pretium tellus duis convallis. Tempus leo eu aenean sed diam urna tempor. Pulvinar vivamus fringilla lacus nec metus bibendum egestas. Iaculis massa nisl malesuada lacinia integer nunc posuere. Ut hendrerit semper vel class aptent taciti sociosqu. Ad litora torquent per conubia nostra inceptos himenaeos.

Lorem ipsum dolor sit amet consectetur adipiscing elit. Quisque faucibus ex sapien vitae pellentesque sem placerat. In id cursus mi pretium tellus duis convallis. Tempus leo eu aenean sed diam urna tempor. Pulvinar vivamus fringilla lacus nec metus bibendum egestas. Iaculis massa nisl malesuada lacinia integer nunc posuere. Ut hendrerit semper vel class aptent taciti sociosqu. Ad litora torquent per conubia nostra inceptos himenaeos.

Lorem ipsum dolor sit amet consectetur adipiscing elit. Quisque faucibus ex sapien vitae pellentesque sem placerat. In id cursus mi pretium tellus duis convallis. Tempus leo eu aenean sed diam urna tempor. Pulvinar vivamus fringilla lacus nec metus bibendum egestas. Iaculis massa nisl malesuada lacinia integer nunc posuere. Ut hendrerit semper vel class aptent taciti sociosqu. Ad litora torquent per conubia nostra inceptos himenaeos.

Lorem ipsum dolor sit amet consectetur adipiscing elit. Quisque faucibus ex sapien vitae pellentesque sem placerat. In id cursus mi pretium tellus duis convallis. Tempus leo eu aenean sed diam urna tempor. Pulvinar vivamus fringilla lacus nec metus bibendum egestas. Iaculis massa nisl malesuada lacinia integer nunc posuere. Ut hendrerit semper vel class aptent taciti sociosqu. Ad litora torquent per conubia nostra inceptos himenaeos.

Testing

Fast forward to today, and we now have a game-changer. The powerbi-modeling-mcp server for Claude Desktop can automate this entire process while keeping you fully in control.

The Challenge

Run the Best Practice Analyzer on any real-world Power BI model and you'll likely see something like this:

• 47 performance violations
• 89 DAX expression issues
• 134 formatting inconsistencies
• 28 maintenance warnings

Even a small sized model can surface 200+ recommendations across categories like:

Performance: Bi-directional relationships on high-cardinality columns, unpartitioned large tables, auto-date tables

DAX Quality: Division operators instead of DIVIDE(), IFERROR usage, improperly qualified column references

Formatting: Missing data categories, exposed foreign keys, unmarked primary keys

Maintenance: Unused columns, orphaned tables, unreferenced data sources

Clicking through each one to understand, fix, test, and validate takes a lot of time.

The Solution: Intelligent Automation using Claude Desktop MCP with Human Oversight

With Claude Desktop connected to your Power BI model via the powerbi-modeling-mcp server, you get an AI agent that understands Power BI semantics, can read your model structure, and systematically apply fixes—all while explaining each change and checking for potential issues. 

What You'll Need

Tabular Editor 2 (the free version—no need for TE3)Claude Desktop with powerbi-modeling-mcp server configured 

Your Power BI model (.pbix or .pbit file) and a backup of your model (seriously—create one before starting)

Guides 

How to install and setup your MCP server see here: Get started with the remote Power BI MCP server - Power BI | Microsoft Learn and here: GitHub - microsoft/powerbi-modeling-mcp: The Power BI Modeling MCP Server, brings Power BI semantic ...

The Step-by-Step Workflow

1. Run the Best Practice Analyzer

Open your model in Tabular Editor 2:

• Launch it from Power BI Desktop's External Tools ribbon, OR
• Open your .pbix or .pbit file directly in Tabular Editor 2

Run the analyzer:

• Press F10 or navigate to Tools → Best Practice Analyzer
• Review the violations organized by category
•  CenterLargeAccessibilityDelete 

2. Connect Claude to Your Model

Open Claude Desktop and use this prompt structure:

Task: Optimize Power BI Semantic Model Using Best Practice AnalyzerStep 1: Connect to ModelConnect to the Power BI semantic model file: [path/to/your_file.pbix]Step 2: Analyze and Apply ImprovementsReview the following Best Practice Analyzer violations and apply corrections sequentially.Step 3: Safety FirstBEFORE applying any correction, evaluate:- Will this break existing calculations, relationships, or functionality?- Could this cause the model to fail or become unstable?If potential issues are identified, skip that violation and request my approval before proceeding.Step 4: Document EverythingLog each change: object name, property modified, old value, new value.

Claude will connect to your model and be ready to process violations. 

Guide

For a comprehesive Prompt guide you can download this prompt file from the attachement and adopt it for your use case.

3. Work in Controlled Batches

This is crucial: don't paste all 200 violations at once. Instead, work in logical groups:

Batch 1: All formatting violations (safest to start)Batch 2: Simple DAX improvements (DIVIDE, qualified references)Batch 3: Performance optimizations (hiding columns, marking keys)Batch 4: Maintenance items (removing unused objects)

Copy one batch of violations from the BPA window and paste them into Claude. The agent will:

• Analyze each violation
• Evaluate the safety of the fix
• Apply the correction using the MCP server
• Document what changed and why
• Flag anything that needs your review

4. Validate After Each Batch

After Claude processes a batch:

1. Save in Tabular Editor (Ctrl+S)
2. Open the model in Power BI Desktop
3. Test your reports and key calculations
4. Verify data refresh works (if applicable)

If something breaks, restore from your backup and proceed more cautiously with that specific type of change.

Note, that sometimes a refresh of your model is required. 

Understanding the Rules

Want to know what a specific BPA rule actually does?

1. In Tabular Editor, go to Tools → Manage BPA Rules
2. Click "Rules for local user"
3. Select any rule and click "Edit Rule"
4. Review the description and reference documentation

Each rule includes an explanation of the best practice and why it matters—often with links to detailed Microsoft documentation.

Some Tips for Success

Start with formatting: These changes are typically safe and give you confidence in the process

Test incrementally: Never apply 50+ changes without validating between batches

Read the rules: If you don't understand why a rule exists, look it up before applying the fix

Version your backups: Create dated copies as you progress (Model_Jan30, Model_Jan31, etc.)

Review Claude's log: The agent documents every change—use this as your change record

You're the expert: Claude automates the mechanics, but you make the final call on complex scenarios 

Pro Tip:

Create a project in Claude AI so you can use the BPA on each of your models during development just with a simple prompt.

Real-World Impact

Before: Manually reviewing and fixing 200+ violations across a production model = 4-6 hours of tedious point-and-click work

After: Claude processes batches in minutes while you review, validate, and make strategic decisions = 45-60 minutes of focused oversight. 

Hope that helps. Drop your questions in the comments. 

Resources

• Original Michael Kovalsky blog post
• Best Practice Analyzer documentation
• Optimize Power BI Models Fast with Tabular Editor’s Best Practices Analyzer ()
• GitHub - microsoft/powerbi-modeling-mcp: The Power BI Modeling MCP Server, brings Power BI semantic ...
• Get started with the remote Power BI MCP server - Power BI | Microsoft Learn

 BPA Claude Prompt Guide.docx 

QA Issue.

Fast forward to today, and we now have a game-changer. The powerbi-modeling-mcp server for Claude Desktop can automate this entire process while keeping you fully in control.

The Challenge

Run the Best Practice Analyzer on any real-world Power BI model and you'll likely see something like this:

• 47 performance violations
• 89 DAX expression issues
• 134 formatting inconsistencies
• 28 maintenance warnings

Even a small sized model can surface 200+ recommendations across categories like:

Performance: Bi-directional relationships on high-cardinality columns, unpartitioned large tables, auto-date tables

DAX Quality: Division operators instead of DIVIDE(), IFERROR usage, improperly qualified column references

Formatting: Missing data categories, exposed foreign keys, unmarked primary keys

Maintenance: Unused columns, orphaned tables, unreferenced data sources

Clicking through each one to understand, fix, test, and validate takes a lot of time.

The Solution: Intelligent Automation using Claude Desktop MCP with Human Oversight

With Claude Desktop connected to your Power BI model via the powerbi-modeling-mcp server, you get an AI agent that understands Power BI semantics, can read your model structure, and systematically apply fixes—all while explaining each change and checking for potential issues. 

What You'll Need

Tabular Editor 2 (the free version—no need for TE3)Claude Desktop with powerbi-modeling-mcp server configured 

CenterLargeAccessibilityDelete

Your Power BI model (.pbix or .pbit file) and a backup of your model (seriously—create one before starting)

Guides 

How to install and setup your MCP server see here: Get started with the remote Power BI MCP server - Power BI | Microsoft Learn and here: GitHub - microsoft/powerbi-modeling-mcp: The Power BI Modeling MCP Server, brings Power BI semantic ...

The Step-by-Step Workflow

1. Run the Best Practice Analyzer

Open your model in Tabular Editor 2:

• Launch it from Power BI Desktop's External Tools ribbon, OR
• Open your .pbix or .pbit file directly in Tabular Editor 2

Run the analyzer:

• Press F10 or navigate to Tools → Best Practice Analyzer

CenterLargeAccessibilityDelete

• Review the violations organized by category

CenterLargeAccessibilityDelete

•  CenterLargeAccessibilityDelete

CenterLargeAccessibilityDelete

2. Connect Claude to Your Model

Open Claude Desktop and use this prompt structure:

Task: Optimize Power BI Semantic Model Using Best Practice AnalyzerStep 1: Connect to ModelConnect to the Power BI semantic model file: [path/to/your_file.pbix]Step 2: Analyze and Apply ImprovementsReview the following Best Practice Analyzer violations and apply corrections sequentially.Step 3: Safety FirstBEFORE applying any correction, evaluate:- Will this break existing calculations, relationships, or functionality?- Could this cause the model to fail or become unstable?If potential issues are identified, skip that violation and request my approval before proceeding.Step 4: Document EverythingLog each change: object name, property modified, old value, new value.

Claude will connect to your model and be ready to process violations. 

CenterLargeAccessibilityDelete

Guide

For a comprehesive Prompt guide you can download this prompt file from the attachement and adopt it for your use case.

3. Work in Controlled Batches

This is crucial: don't paste all 200 violations at once. Instead, work in logical groups:

Batch 1: All formatting violations (safest to start)Batch 2: Simple DAX improvements (DIVIDE, qualified references)Batch 3: Performance optimizations (hiding columns, marking keys)Batch 4: Maintenance items (removing unused objects)

Copy one batch of violations from the BPA window and paste them into Claude. The agent will:

• Analyze each violation
• Evaluate the safety of the fix
• Apply the correction using the MCP server
• Document what changed and why
• Flag anything that needs your review

4. Validate After Each Batch

After Claude processes a batch:

1. Save in Tabular Editor (Ctrl+S)
2. Open the model in Power BI Desktop
3. Test your reports and key calculations
4. Verify data refresh works (if applicable)

If something breaks, restore from your backup and proceed more cautiously with that specific type of change.

Note, that sometimes a refresh of your model is required.

Understanding the Rules

Want to know what a specific BPA rule actually does?

1. In Tabular Editor, go to Tools → Manage BPA Rules
2. Click "Rules for local user"
3. Select any rule and click "Edit Rule"
4. Review the description and reference documentation

Each rule includes an explanation of the best practice and why it matters—often with links to detailed Microsoft documentation.

Some Tips for Success

Start with formatting: These changes are typically safe and give you confidence in the process

Test incrementally: Never apply 50+ changes without validating between batches

Read the rules: If you don't understand why a rule exists, look it up before applying the fix

Version your backups: Create dated copies as you progress (Model_Jan30, Model_Jan31, etc.)

Review Claude's log: The agent documents every change—use this as your change record

You're the expert: Claude automates the mechanics, but you make the final call on complex scenarios

Pro Tip:

Create a project in Claude AI so you can use the BPA on each of your models during development just with a simple prompt.

Real-World Impact

Before: Manually reviewing and fixing 200+ violations across a production model = 4-6 hours of tedious point-and-click work

After: Claude processes batches in minutes while you review, validate, and make strategic decisions = 45-60 minutes of focused oversight. 

Hope that helps. Drop your questions in the comments. 

Resources

• Original Michael Kovalsky blog post
• Best Practice Analyzer documentation
• Optimize Power BI Models Fast with Tabular Editor’s Best Practices Analyzer ()
• GitHub - microsoft/powerbi-modeling-mcp: The Power BI Modeling MCP Server, brings Power BI semantic ...
• Get started with the remote Power BI MCP server - Power BI | Microsoft Learn

 BPA Claude Prompt Guide.docx 

CenterLargeAccessibilityDelete

Play Video

CenterLargeAccessibilityDelete

Fast forward to today, and we now have a game-changer. The powerbi-modeling-mcp server for Claude Desktop can automate this entire process while keeping you fully in control.

The Challenge

Run the Best Practice Analyzer on any real-world Power BI model and you'll likely see something like this:

• 47 performance violations
• 89 DAX expression issues
• 134 formatting inconsistencies
• 28 maintenance warnings

Even a small sized model can surface 200+ recommendations across categories like:

Performance: Bi-directional relationships on high-cardinality columns, unpartitioned large tables, auto-date tables

DAX Quality: Division operators instead of DIVIDE(), IFERROR usage, improperly qualified column references

Formatting: Missing data categories, exposed foreign keys, unmarked primary keys

Maintenance: Unused columns, orphaned tables, unreferenced data sources

Clicking through each one to understand, fix, test, and validate takes a lot of time.

The Solution: Intelligent Automation using Claude Desktop MCP with Human Oversight

With Claude Desktop connected to your Power BI model via the powerbi-modeling-mcp server, you get an AI agent that understands Power BI semantics, can read your model structure, and systematically apply fixes—all while explaining each change and checking for potential issues. 

What You'll Need

Tabular Editor 2 (the free version—no need for TE3)Claude Desktop with powerbi-modeling-mcp server configured 

CenterLargeAccessibilityDelete

Your Power BI model (.pbix or .pbit file) and a backup of your model (seriously—create one before starting)

Guides 

How to install and setup your MCP server see here: Get started with the remote Power BI MCP server - Power BI | Microsoft Learn and here: GitHub - microsoft/powerbi-modeling-mcp: The Power BI Modeling MCP Server, brings Power BI semantic ...

The Step-by-Step Workflow

1. Run the Best Practice Analyzer

Open your model in Tabular Editor 2:

• Launch it from Power BI Desktop's External Tools ribbon, OR
• Open your .pbix or .pbit file directly in Tabular Editor 2

Run the analyzer:

• Press F10 or navigate to Tools → Best Practice Analyzer

CenterLargeAccessibilityDelete

• Review the violations organized by category

CenterLargeAccessibilityDelete

•  CenterLargeAccessibilityDelete

CenterLargeAccessibilityDelete

2. Connect Claude to Your Model

Open Claude Desktop and use this prompt structure:

Task: Optimize Power BI Semantic Model Using Best Practice AnalyzerStep 1: Connect to ModelConnect to the Power BI semantic model file: [path/to/your_file.pbix]Step 2: Analyze and Apply ImprovementsReview the following Best Practice Analyzer violations and apply corrections sequentially.Step 3: Safety FirstBEFORE applying any correction, evaluate:- Will this break existing calculations, relationships, or functionality?- Could this cause the model to fail or become unstable?If potential issues are identified, skip that violation and request my approval before proceeding.Step 4: Document EverythingLog each change: object name, property modified, old value, new value.

Claude will connect to your model and be ready to process violations. 

CenterLargeAccessibilityDelete

Guide

For a comprehesive Prompt guide you can download this prompt file from the attachement and adopt it for your use case.

3. Work in Controlled Batches

This is crucial: don't paste all 200 violations at once. Instead, work in logical groups:

Batch 1: All formatting violations (safest to start)Batch 2: Simple DAX improvements (DIVIDE, qualified references)Batch 3: Performance optimizations (hiding columns, marking keys)Batch 4: Maintenance items (removing unused objects)

Copy one batch of violations from the BPA window and paste them into Claude. The agent will:

• Analyze each violation
• Evaluate the safety of the fix
• Apply the correction using the MCP server
• Document what changed and why
• Flag anything that needs your review

4. Validate After Each Batch

After Claude processes a batch:

1. Save in Tabular Editor (Ctrl+S)
2. Open the model in Power BI Desktop
3. Test your reports and key calculations
4. Verify data refresh works (if applicable)

If something breaks, restore from your backup and proceed more cautiously with that specific type of change.

Note, that sometimes a refresh of your model is required.

Understanding the Rules

Want to know what a specific BPA rule actually does?

1. In Tabular Editor, go to Tools → Manage BPA Rules
2. Click "Rules for local user"
3. Select any rule and click "Edit Rule"
4. Review the description and reference documentation

Each rule includes an explanation of the best practice and why it matters—often with links to detailed Microsoft documentation.

Some Tips for Success

Start with formatting: These changes are typically safe and give you confidence in the process

Test incrementally: Never apply 50+ changes without validating between batches

Read the rules: If you don't understand why a rule exists, look it up before applying the fix

Version your backups: Create dated copies as you progress (Model_Jan30, Model_Jan31, etc.)

Review Claude's log: The agent documents every change—use this as your change record

You're the expert: Claude automates the mechanics, but you make the final call on complex scenarios

Pro Tip:

Create a project in Claude AI so you can use the BPA on each of your models during development just with a simple prompt.

Real-World Impact

Before: Manually reviewing and fixing 200+ violations across a production model = 4-6 hours of tedious point-and-click work

After: Claude processes batches in minutes while you review, validate, and make strategic decisions = 45-60 minutes of focused oversight. 

Hope that helps. Drop your questions in the comments. 

Resources

• Original Michael Kovalsky blog post
• Best Practice Analyzer documentation
• Optimize Power BI Models Fast with Tabular Editor’s Best Practices Analyzer ()
• GitHub - microsoft/powerbi-modeling-mcp: The Power BI Modeling MCP Server, brings Power BI semantic ...
• Get started with the remote Power BI MCP server - Power BI | Microsoft Learn

 BPA Claude Prompt Guide.docx 

CenterLargeAccessibilityDelete

Play Video

Here are a few tips to help you create great discussions:

• Make the discussion title or question as descriptive as possible. A good discussion title is a short preview of your post and is what gets people to click and read. A well written title is also going to help search engines better index your post which will bring more people into the discussion. For example, instead of ‘Won’t Connect’, try ‘Help, I’m having problems getting my Acme modem into bridge mode.’
• Proof read. Spelling mistakes, typos, and bad grammar will distract readers from the point you’re trying to make. Vanilla automatically saves drafts as you type. If you’re writing a long post, save it as a draft and come back to it after a few minutes or as long as it takes your brain to forget what you had written.
• Use minimal formatting. Overly formatted posts can also distract from the message and it encourages others to do likewise and you end up with a hard to read thread.
• Put your post in the right category. The right category can be the one that has a relevant category name or it can be a category where this kind of post is often made.
• Use tags. Tags are helpful for others to find keyword related posts. It also helps the site admins get a sense for what topics are popular.
• If you want responses, ask for them. In marketing this is called a ‘call to action’. If you want others to comment, you can encourage them by asking them to do so. If your post is just an FYI then don’t.
• Go easy on the insider jargon. Inside jokes and inside references can be fun and make the community unique but too much of it can turn off new members.
• Add an image. Images add visual interest and make your post look great when shared to social networks. You can embed an image using the button bar or you can upload one from your desktop or phone.
• Mention others. Credit other members if you are building off their previous comments or if you want to draw them into the discussion. Put the @ before a username to mention someone.
• Take ownership. Most important of all, take ownership of the discussions that you have created. Respond to comments promptly and thoughtfully. Thank others for commenting on your discussion and help with moderation if things get heated.

Thanks for posting! Have fun!