Chatbot to fetch PowerBI Report data

hoosha_11 · ‎03-07-2025

Hi,

We have an application that uses a SQL Server hosted in Azure, with Power BI Embedded for reporting. The setup includes publishing a master dataset to Power BI Service and linking multiple reports to it.

The goal is to create a simple chatbot where users can interact naturally.

For example, a user asks: What is the Average order number for July 2023? The chatbot should then find the related report, apply the filters and return the result.

I was thinking of calling the Azure OpenAI API with Python and connecting to the Power BI REST API or XMLA endpoint to retrieve all tables, columns, and measures. Then, I would define multiple prompts to guide OpenAI in selecting the appropriate measure, performing calculations, and delivering results based on the question asked.

Does it make sense? Would there be a better way? Any ideas or suggestions are appreciated.

v-veshwara-msft · ‎05-20-2025

Hi @hoosha_11 ,

We’re following up once more regarding your query. If it has been resolved, please mark the helpful reply as the Accepted Solution to assist others facing similar challenges.

If you still need assistance, please let us know.
Thank you.

v-veshwara-msft · ‎05-05-2025

Hi @hoosha_11 ,

Following up to see if your query has been resolved. If any of the responses helped, please consider marking the relevant reply as the 'Accepted Solution' to assist others with similar questions.

If you're still facing issues, feel free to reach out.

Thank you.

v-veshwara-msft · ‎04-21-2025

Hi @hoosha_11 ,

We haven’t heard from you since our last response and wanted to check if the provided solutions met your needs. If they did, please consider accepting the solution to help others benefit. Otherwise, feel free to reach out for further assistance.

Thank you.

v-veshwara-msft · ‎04-14-2025

Hi @hoosha_11 ,
Thanks for using Fabric Community,
Thanks @andrewsommer for your valuable response,

Agree with your approach of using the Azure OpenAI API and Power BI REST API or XMLA endpoint. It’s a solid strategy to create a chatbot for natural language queries.

Additionally, I wanted to mention the Semantic Model support with AI Skill introduced in the February Fabric update. It allows for easier integration of structured business data and automatic translation of user queries into DAX, simplifying the process and reducing the need for multiple prompts. This could align well with your approach and improve the chatbot experience.
Fabric February 2025 Feature Summary | Microsoft Fabric Blog | Microsoft Fabric

Hope this helps. Please reach out for further assistance.

Please consider marking the helpful reply as "Accepted Solution" to assist others with similar issues.

Thank you.

andrewsommer · ‎03-07-2025

Microsoft just announced in the February Fabric update the Semantic Model support with AI Skill:

“The AI Skill now supports Semantic models, a new data source that enables AI Creators to integrate structured business data for natural language querying. Semantic Models provide a curated data layer that defines business logic, relationships, and aggregations, making it easier for users to analyze and explore data without needing deep technical expertise.

AI Creators can add Semantic models and selectively scope which tables are available for querying. Once configured, both AI Creators and consumers can ask natural language questions, which the AI Skill translates into DAX (Data Analysis Expressions) queries using LLMs. Users can trace the entire query process in the chat canvas, viewing how their input is interpreted, which transformations occur, and the exact DAX queries executed.”