It is now widely accepted that decisions made by AI systems must be explainable to their users. However, in practice, it often remains unclear how this explainability should be concretely implemented. This is especially important for nontechnical users, such as claims assessors at insurance companies, …