How are teams using AI agents to automate data science workflows?
I am exploring how AI agents can support modern data science workflows by automating repetitive tasks and improving collaboration between data teams. Some potential use cases: Automating data exploration and preprocessing steps Generating insights from datasets using natural language queries Assisting with feature engineering and model evaluation Triggering ML pipelines based on business events Monitoring model performance and identifying data quality issues I would like to understand how data science teams are combining Microsoft Fabric, notebooks, ML workflows, and AI agents in real-world projects. What approaches, tools, or architecture patterns are you using for AI-assisted data science workflows?
Fabric Data Agent Limitations of 25 items
There's no way for Fabric Data Agent to give more than 25 items of a list. And this is quite confusing. I have a dataset with the States and Presidents of USA. It is full with 50 states and 47 presidents. As you can see here, when asking for the list, it only contains the 25 first. When looking at the DAX query, it gives the 50 in result. // List all U.S. states with their abbreviations, sorted by state name. EVALUATE SELECTCOLUMNS( // Use SUMMARIZE to ensure distinct state rows if duplicates exist SUMMARIZE( 'america_states', 'america_states'[nom], 'america_states'[abréviation] ), "State Name", 'america_states'[nom], "Abbreviation", 'america_states'[abréviation] ) ORDER BY [State Name] ASC No TOPN here, or any filter/limitation. When asked, Agent says there are 50 lines in the table (good answer), but says that the "tool" gives only 25. Obviously, the "tool" is not DAX because it is correct. It's another mysterious tool. I understand that, to maintain a "conversation context", the "tool" is limited on answers with 25 items. My colleagues and I are looking for use-cases when Fabric Data Agent could be relevant. But this limitation of 25 (which is very small, he can't even give the 50 states of USA) is a bad argument for us to use this Fabric Artefact. When trying to build cool agents for the company, and when testing them, users may ask some lists of items. They could have a warning for the partial results.. but 25... gasps !!! Of course, when you ask Copilot for the Web, you get a list with more than 25 items "Here’s the full list of the 50 U.S. states and their capitals in English: Alabama – Montgomery Alaska – Juneau Arizona – Phoenix Arkansas – Little Rock California – Sacramento North Carolina – Raleigh South Carolina – Columbia Colorado – Denver Connecticut – Hartford North Dakota – Bismarck South Dakota – Pierre Delaware – Dover Florida – Tallahassee Georgia – Atlanta Hawaii – Honolulu Idaho – Boise Illinois – Springfield Indiana – Indianapolis Iowa – Des Moines Kansas – Topeka Kentucky – Frankfort Louisiana – Baton Rouge Maine – Augusta Maryland – Annapolis Massachusetts – Boston Michigan – Lansing Minnesota – Saint Paul Mississippi – Jackson Missouri – Jefferson City Montana – Helena Nebraska – Lincoln Nevada – Carson City New Hampshire – Concord New Jersey – Trenton New York – Albany New Mexico – Santa Fe Ohio – Columbus Oklahoma – Oklahoma City Oregon – Salem Pennsylvania – Harrisburg Rhode Island – Providence Tennessee – Nashville Texas – Austin Utah – Salt Lake City Vermont – Montpelier Virginia – Richmond West Virginia – Charleston Washington – Olympia Wisconsin – Madison Wyoming – Cheyenne" Does anyone agree that the limitation is too smal ? Thanks
Fabric Data Agent in Copilot Studio not outputting message
I am having a problem that I have connected my Copilot Studio agent to a Fabric Data Agent. I can ask a question within the chat window, and it will output a response in the MCP popout window, but not in the chat window, and I can't get it to respond in Teams or M365 after I published it either. It is answering because if I ask it to repeat, it will, so just when it calls that MCP, it doesn't send that message.
