When data is loaded for a user step, sometimes the user may have to search for data in LOV with large volumes. Eg adding a spare part to a work order.

Adding such a LOV to a user step may take time to load as the flow wouldn't move to the next step without completing that step. In order to improve the user experience, it is better to have the following;

1. Load batch-wise
   The user step will get populated once the first 100 records are fetched, and the list has the option to load all or the next 100...

2. Load in the background.
   Ability to define some LOVs as 'Load in the Background'. These user steps will be executed as soon as the user starts the workflow in the background(asynchronous mode). So when the user reaches a certain user step data set is already loaded. An indicator may be needed in order to indicate if the data set is fully loaded or not.
   (somewhat similar to the offline dataset, but the relevant connector will start to load the data set in the background as soon as the flow is started without interrupting the user)