Our recommendation model stores the necessary interactions, such as "Invite to apply" or "Discard", that users perform with the recommendation results. This data is used to further develop the recommendation model.
Read more details in our article "Active sourcing with automated candidate recommendations".