Occupational classifications are essential tools built by National Statistical Institutes in order to deal with the diversity of jobs. The French socio-professional classification (professions et catégorie socio-professionnelles, PCS, in French) enables statisticians to group individual professional status based on job contents, together with economic and institutional contexts. The dictionary used today was set in 2003 (PCS2003) and needed to be updated (PCS2020). In this context, we experiment supervised machine learning algorithms to code into the new one in order to try to make the production process easier.