This article presents a non-exhaustive analysis of alternatives for the development of models to predict the ultimate cost of claims and explainability techniques that seek to facilitate the understanding of the different models proposed for a set of claims that occurred between 1985 and 2005.The models with the best results and the process that was carried out for the use of the different methods are shown, the tests carried out are commented and the best alternatives are identified for the prediction of the cost of claims based on the characteristics available in the base. as well as the respective analysis of explainability that seek to facilitate their understanding and, to a certain level, their review.