Tre percorsi, una ricerca: la codifica qualitativa tra manualità, CAQDAS e AI

The increasing use of digital tools in qualitative research has become a central topic in contemporary methodological debates, raising questions about how technologies effectively support researchers’ interpretive work. This article revisits a qualitative study on the reception of migrants in Southern Europe conducted in the provinces of Salerno (Italy) and Cádiz (Spain) and explores the analytical process through three modes: (a) traditional manual coding, (b) computer-assisted qualitative data analysis with NVivo 15, and (c) automated coding using the new NVivo functions Auto Codeand Lumivero AI Assistant. The study compares these analytical paths to identify convergences and divergences in interpretive depth, epistemological coherence, and processingtime.Findings show that automation accelerates pre-coding and lexical mapping but produces highly descriptive coding that requires critical review by the researcher. Beyond procedural comparison, the study contributes to the methodological debate by outlining areflexive model for integrating human judgment and machine assistance in qualitative analysis. It argues that while CAQDAS and AI can effectively support repetitive and exploratory tasks, the production of dense, context-sensitive knowledge ultimately depends on the researcher’s reflexive practice.