The purpose of this paper is to analyze written and voice messages of five bilingual speakers of Croatian and Italian language exchanged via WhatsApp application. Having in mind all particularities of computer-mediated communication as a medium that influences the features of linguistic interaction carried out through it, qualitative sociolinguistic and pragmalinguistic methods and methods of discourse analysis were applied in the corpus analysis. The analysis was conducted with a special regard to the nature of the bilingualism of the examinees and their code switching. The results of the analysis show formal and functional differences in the use of code switching between the two language microcommunities to which the examined speakers belong. It is also shown that code switching normally follows a default system of rules and that in terms of gender there are no significant differences in its structuring among the speakers. The paper is a contribution to the study of Croatian-Italian bilingualism as well as a contribution to the study of influences of the new media on Croatian and Italian language.