We present a new approach to automate the spectroscopic redshift reliability assessment based on machine learning (ML) and characteristics of the redshift probability density function (PDF). We propose to rephrase the spectroscopic redshift estimation into a Bayesian framework, in order to incorporate all sources of information and uncertainties related to the redshift estimation process, and produce a redshift posterior PDF that will be the starting-point for ML algorithms to provide an automated assessment of a redshift reliability. As a use case, public data from the VIMOS VLT Deep Survey is exploited to present and test this new methodology. We first tried to reproduce the existing reliability flags using supervised classification to describe different types of redshift PDFs, but due to the subjective definition of these flags, soon opted for a new homogeneous partitioning of the data into distinct clusters via unsupervised classification. After assessing the accuracy of the new clusters via resubstitution and test predictions, unlabelled data from preliminary mock simulations for the Euclid space mission are projected into this mapping to predict their redshift reliability labels.