Power Platform Community Forum Thread Details

Hi fellow experts!

I’m currently working on an exciting project at our startup, where we’re developing a leadership app using Power Apps (specifically, a Canvas App). Our data source of choice is Dataverse, which provides a robust foundation for our app. However, we recently hit a roadblock while trying to incorporate AI-driven insights into our solution.

Our initial approach involved experimenting with AI Builder, Microsoft’s tool for creating machine learning models within the Power Platform. Unfortunately, we quickly discovered that AI Builder has some limitations:

AI Builder restricts us to using only one table for model training. This posed a significant challenge because our app relies on multiple critical tables.
Even when working with just one table, the performance of our AI models wasn’t up to par. We needed a more robust solution.

To get an understanding of what we want to achieve, Imagine this scenario: A user interacts with our app’s chat bot and asks, “What were the top 10 selling cars at Dealership XYZ last month, and who were the customers that purchased them?” Achieving this requires combining data from multiple tables (Car Brand, Dealership, and Customer).

The reason I want to use the AI Builder is to create new records in Dataverse table depending on the response. For example I want to take the list of top 10 selling cars and create an approval to increase the inventory for dealership by pressing a button. And this all needs to happen within the app.

Categories:

AI Builder