Forum Discussion
Microsoft EmployeeShare your thoughts on Prep data for AI (preview)
Hit Reply to tell us what you think about the new Prep data for AI feature so we can continue to improve.
Thanks,
-Power BI product team
To read more about the feature see our documentation.
8 Replies
v-pnaroju-msftCommunity Support
Hi emlisa,
Thank you for reaching out and gathering feedback on the "Prep data for AI (preview)" feature in Power BI.
At present, there are no significant issues reported by the community regarding the use of the "Prepare your data for AI (preview)" feature. We will keep you informed if we receive any updates or concerns.
Thank you.
Most enterprises have a singular semantic model punished in the web with multiple reports connected to it via Live Connection. In this scenario, when do you think the Prep for AI can be unlocked?
It doesn't work when the development of the semantic model is done via XMLA endpoints directly in the DEV workspace
- emlisa
Today, Prep for AI is tied to the semantic model rather than the report. So in the scenario you're describing, where multiple reports connect to a single semantic model via Live Connection, you would need to either enable Prep for AI on the underlying semantic model itself, or create a new semantic model using DirectQuery to that model and define Prep for AI there. For the scenario you’re describing, would you hope to be able to define Prep for AI directly on the connected reports?
An ideal outcome is to be able to define the Prep for AI on the existing semantic model. However neither can i download the model to desktop (xmla restriction) nor can i define the AI rules in the web.
I'm not sure why a direct query connection is the right answer here? Can you explain a bit more?
Hi can you discuss how "Prep Data for AI" and "Fabric Data Agents" are related? I cannot find documentation on this. Does prepping the data for AI have any impact of the creation of Fabric Data Agents? For example, both of them have an "AI Instructions" setting, so I am confused whether I should put instructions on the semantic model and also add them to the data agent?
laurenfaberHi! The instructions in each location are slightly different. When working in Power BI you are giving context/instructions on the semantic model. When working in the Data Agent you are giving instructions to the Data Agent. We are currently in the process of updating the data agent to automatically use the instructions set on the semantic model, but today they are separate experiences.
When i import my culture from another semantic model prep data for AI is not working due to Q&A being greyed out. Is this a known bug och intentinal?
We want our AI to be able to responde in swedish but cannot prep our data after our culture is created.BR
