Free Microsoft Power Platform PL-900 Actual Exam Questions - Question 12 Discussion
DRAG DROP A manufacturing company is evaluating Al Builder. You need to select Al Builder models to address specified requirements. Which model types should you use? To answer, drag the appropriate model types to the correct requirements. Each model type may be used once, more than once, or not at all. You may need to drag the split bar between panes or scroll to view content. 
I thought about the requirements that focus on grouping or categorizing items—those really scream classification models since they assign labels based on input features. For the ones where the goal is to estimate or forecast values, prediction models make more sense because they handle continuous outcomes. If any requirement involves detecting specific things in images or forms, object detection or form processing models fit better, but I didn’t see anything like that here. So basically, match classification models with categorical decisions and prediction models with numeric forecasts.
For this, I looked at the nature of the tasks—if it’s about labeling or grouping, classification models seem right. For things that involve numbers or forecasting, prediction models fit better. Also, if there’s something with text extraction or understanding text, that points to an entity extraction model. So basically, match the model type to the data output you expect: categories with classification, continuous values with prediction, and text data with entity extraction. That covers all the bases without overcomplicating it.
For the requirement about sorting or categorizing data, I’d go with classification models since they’re built to assign labels. For anything that involves estimating future values or continuous data, prediction models seem like the better fit. If there’s something that needs extracting info from forms or documents, then a form processing model would be the way to go. So basically, match the type of output you’re after—categories, numbers, or extracted text—to classification, prediction, or form processing respectively.
I’ve seen similar setups with AI Builder for classification and prediction tasks. Usually, classification models fit well when you need to sort data into categories, while prediction models work better for forecasting numbers or outcomes. The drag-and-drop part can get tricky because you gotta map each requirement carefully. If one is about sorting types of inputs, that’s probably a classification model; if it’s about estimating a value, prediction’s the go-to. Hope this helps clear things up a bit!