The selection process composed of 3,105 candidates in which are collected 64 data (socio-demographic, languages, preferences, tests) and the CVs in different formats and languages. As a result of the process, 50 candidates were selected. The model predicted the pre-selected candidates with high accuracy.
Following the order proposed by the AI model, half of the pre-selected candidates (250) were positioned among the 500 proposed candidates (1:2 in comparison to 1:6 of the random process).
Small selection process with 110 candidates in which the CVs and cover letters were used, with no applicants structured information. As a result of the process, 6 candidates were selected. The model predicted the selected candidates with high accuracy. The great innovation of the process is the joint use of both CVs and covering letters, as well as the learning ability was based on the selection process itself.