Developing soft communication skills with virtual reality and artificial intelligenceBohinc, Leon (Avtor) Perko, Igor (Avtor) soft skillsbusiness communicationvirtual realityartificial intelligenceinstant feedbackexperiential learningThis paper presents a methodological framework for examining how AI-enhanced virtual reality environments can be used to study soft communication skills within a social research framework. Specifically, it focuses on how experiments conducted in VR environments can integrate AI-generated behavioural metrics with conventional social science research methods such as surveys and transcript analysis. The framework is structured around a small-scale social experiment employing a pre-test/post-test design. The design includes two VR-based communication simulations using the VirtualSpeech platform. To examine how different data sources can be integrated within the experimental framework, AI-generated analytical feedback across multiple speech-related performance categories is combined with structured questionnaires capturing user experience and self-assessed competence before and after the intervention, transcript-based analysis of participants’ spoken responses, and AI-assisted analytical tools. The proposed experimental design combines three complementary layers of data collection: automated behavioural indicators generated by the VR platform, transcript-based qualitative analysis of communication performance, and participant self-report questionnaires capturing perceived competence and experience. The paper aims to demonstrate how heterogeneous data sources can be systematically integrated into a single experimental framework for studying communication behaviour in immersive environments. It provides methodological insights for researchers designing AI-enhanced VR experiments and supports future interdisciplinary studies of communication.20262026-07-27 13:13:01Znanstveno delo99033UDK: 005.336.2:004.8OceCobissID: 283282435COBISS_ID: 286117379DOI: 10.18690/um.epf.7.2026.52sl