Hesitant Pythagorean fuzzy sets represent a successful generalization of Pythagorean fuzzy sets that embeds the benefits of hesitation. As such they enable the practitioner to incorporate vague data and information very effectively. We take advantage of this feature in the procedure of risk evaluation in failure modes and effects analysis (FMEA). This technique has demonstrated to be a powerful and valuable framework for the identification of possible failures and the reduction of their occurrence. The advantages of hesitant Pythagorean fuzzy sets motivate us to extend the existing methods of risk assessment in FMEA to the case where data are in that format. Therefore, we propose an ELimination and Choice Translating REality-I (ELECTRE-I) approach under the hesitant Pythagorean fuzzy (HPF) environment to measure and rank risks. To assess the failures, the HPF ELECTRE-I technique creates hesitant Pythagorean fuzzy concordance/discordance matrices. Then, we use an outranking decision graph to obtain the most crucial failure mode. We illustrate the structure of this procedure through a flow chart. Furthermore, the model is validated through the study of real-life applications, such as the prevention of infant abduction and risk evaluation of health care failure modes in blood transfusion. Finally, to showcase the superiority and effectiveness of the suggested method, we provide a comprehensive comparison of the HPF ELECTRE-I approach with existing techniques.