Medium

Gulab Chand Tejwani

Gulab Chand Tejwani

·

3d ago

What Power BI Copilot Actually Reads When It Summarises a Report

Your report can be correct while its AI-generated summary is incomplete. The reason may be the visuals Copilot selected, the way a matrix…

What Power BI Copilot Actually Reads When It Summarises a Report

Gulab Chand Tejwani

·

Nov 2

Direct Lake Calculated Columns Are Query-Time. Same DAX, Different Execution Model.

Direct Lake on OneLake can now use DAX calculated columns without switching the table out of Direct Lake. That sounds like feature parity…

Direct Lake Calculated Columns Are Query-Time. Same DAX, Different Execution Model.

Gulab Chand Tejwani

·

Oct 26

Calculation Groups Can Eliminate Measure Explosion in Power BI.

Calculation groups solve one of the oldest maintenance problems in Power BI: repeating the same calculation pattern across dozens of…

Calculation Groups Can Eliminate Measure Explosion in Power BI.

In

Towards AI

by

Gulab Chand Tejwani

·

Oct 19

Copilot’s Fabric Token Rates Dropped. On October 1, Consumption Changes Again.

As of 18 September 2026, Microsoft’s public Copilot-in-Fabric documentation lists 100 CU seconds per 1,000 input tokens, 10 CU seconds per…

Copilot’s Fabric Token Rates Dropped. On October 1, Consumption Changes Again.

In

Towards AI

by

Gulab Chand Tejwani

·

Oct 16

Power BI Desktop’s October 2026 Deadline: Which Versions Are You Running?

March 2026 and earlier releases will lose built-in OneDrive and SharePoint saving and sharing. The update is only part of the job.

Power BI Desktop’s October 2026 Deadline: Which Versions Are You Running?

Gulab Chand Tejwani

·

Oct 12

Your Power BI Semantic Model Is More Than a Reporting Layer. Treat It Like an API.

Define what consumers can depend on, and test what happens when it changes.

Your Power BI Semantic Model Is More Than a Reporting Layer. Treat It Like an API.

In

Towards AI

by

Gulab Chand Tejwani

·

Sep 19

If You Tested Power BI Copilot by Asking the Same Question Twice, You Tested the Cache

Microsoft’s Copilot documentation contains a paragraph that quietly invalidates how most teams evaluate whether Copilot is reliable. There…

If You Tested Power BI Copilot by Asking the Same Question Twice, You Tested the Cache

In

Towards AI

by

Gulab Chand Tejwani

·

Sep 16

The Semantic Layer Is Standardising. Microsoft Isn’t in the Room.

Sixty organisations have joined an effort to make semantic models portable between tools. Snowflake started it. Databricks, dbt Labs…

The Semantic Layer Is Standardising. Microsoft Isn’t in the Room.

In

Towards AI

by

Gulab Chand Tejwani

·

Sep 8

One Query Engine Now Answers Your Questions Everywhere.

The August 2026 Fabric update brought advanced DAX generation to Fabric data agents on the Preview runtime. Most coverage has framed it as…

One Query Engine Now Answers Your Questions Everywhere.

In

Towards AI

by

Gulab Chand Tejwani

·

Sep 2

Your Semantic Model Can Now Power a Fabric App. That Is Bigger Than Another Dashboard.

Microsoft introduced Fabric Apps in preview in 2026, with Rayfin providing the SDK and CLI used to build and deploy them. One command…

Your Semantic Model Can Now Power a Fabric App. That Is Bigger Than Another Dashboard.

Gulab Chand Tejwani

Gulab Chand Tejwani

I build analytics solutions for real businesses and teach analysts how to think beyond reports using Power BI, SQL, DAX, and AI.

Following

Help

Status

About

Careers

Press

Blog

Store

Privacy

Rules

Terms

Text to speech