Dear Silva, you are absolutely right. Evaluators are mainly interpreters of what has happened, and consequently not only a "tick the box" in a "multiple choice" questionnaire. Of course we have to fill some forms and tables, organize information in a pre-agreed way, but the most important work is when we are able to find a "non previewed" result or impact. Most of our work is "standard work" and it has to be "up to standard", but there is a piece of "original work" that is what adds brightness to the evaluation work. And that cannot be done by AI, as AI always is an "average" of previous experiences.