I am currently building a Workato recipe to automate our ‘Quality Evaluation’ assignments. We have a high-volume queue and we want to ensure that each supervisor is assigned exactly ten interactions per agent per week for review. I am using the Analytics API to find the interactions, but I am struggling with the ‘Random Sampling’ logic. How can I ensure that the interactions selected for evaluation are truly random and not just the first ten that occurred on Monday morning?
Greetings. I am an admin for our German organization and we have very strict rules about ‘Fair’ evaluations. You should not perform the sampling in your Workato recipe. Genesys Cloud has a native ‘Evaluation Policy’ feature that handles the random sampling for you. You can set the percentage of interactions to be evaluated and the system will automatically distribute them across the week. It is much more reliable and it satisfies our legal requirements for unbiased grading!
I evaluate over two hundred interactions every week. Yui14 is right. The native policies are great, but if you absolutely must use Workato (maybe for some custom criteria), you should use a ‘Reservoir Sampling’ algorithm in your script. This allows you to select a random sample from a stream of data without knowing the total count beforehand. It ensures that an interaction on Friday afternoon has the same chance of being selected as one on Monday morning.
I use similar logic for our campaign list selection. To follow up on Bia97, make sure your Workato recipe is also checking the ‘Evaluation Status’ of the interaction. You do not want to assign an interaction for review if it has already been evaluated by another supervisor, or you will end up with duplicate work and very frustrated agents!