Voice Recorder: Integrating AI Assistants into HR Processes

Abstract

This article examines the subjectivity and low efficiency of initial candidate screening processes in IT recruiting. A solution is proposed based on a voice-based intelligent assistant using the GPT-4 model for adaptive interviews, multimodal response analysis (text and voice), and the formation of an objective assessment. The system reduces manual processes for HR managers. The central element of the study is the development of a mathematical model for assessing the system's effectiveness and a detailed description of its architecture. The key element of the study is the developed architecture of a voice-based intelligent recruiter based on GPT-4, including a speech interaction module, an analyzer of the semantic content of responses, a unit for assessing the emotional coloring of speech, and an insight generation system. The article proposes a methodology for integrating the system into IT recruiting processes, describing the stages of adaptive interviewing and the formation of a final candidate assessment. Particular attention is paid to solving problems of processing multimodal data and ensuring the correct operation of the voice interface. Experimental studies demonstrate increased objectivity in initial selection and reduced operational costs while maintaining hiring quality. The results of the study are of practical interest to HR specialists and developers of intelligent decision support systems for personnel decisions. doi 10.54708/19926502_2025_29411019