@CHANDANPAN

1)What are the Prerequisites to use Process AI, additionally if there is any zip/jar/package need to be installed  ?

            If you want to use adaptive models as part of predictions then you need to have DDS and ADM node, if not dataflow node should be enough. No additional installation is required from Pega platform 8.6 and above.

 For case management prediction with case completion outcome, we create a prediction with an adaptive model by default whereas, if you use other two outcomes (Claims fraud & Custom) we create a prediction with Scorecard as the default model. These predictions will not require a DDS or ADM node, also these scorecard rules can be replaced with a predictive model using below path

1. Open the prediction
2. Navigate to the models tab and select replace model option
3. Follow the replace path

2)From the Case designer, under setting there is a configuration to add Prediction , I can create new Prediction from Prediction Studio but when I create a new prediction from Prediction I can only see Adaptive model are beign created, how can I   add Predictive model there?

            New -->Case management -->Case completion  This will create a prediction with an adaptive model by default, if you want to replace the model with a predictive model then you need to follow the MLOps process to replace it with a predictive model by following the below path

1. Open the prediction
2. Navigate to the models tab and select replace model option
3. Follow the replace path

3)From the process / stage / steps / flow  what are the various connecting bridge/shapes (if any) between AI Model and and case ?

            From case designer wherever we see support for condition builder we should see predictions listed provided predictions are added to the case from case settings tab.

4)How the context class of the Strategy / Model is connected with Work- class of Case , because inputs to these models will be triggered frm case ?

            When creating a prediction user gets to select the subject for the prediction, the class of the subject will be used for creating the prediction. If the subject class and case class are same then the prediction strategy context class will be same as that of the case.

If the subject is a data object then the class of the data object will be the class of the prediction, when such a prediction is added to a case then it looks for existing relation between the case type and the data object which typically is a page type property in a case class. If such a relation exists then we list all the predictions that are defined in the data object. For such predictions, the data object class will be the primary context for the prediction strategy.