Integrating Natural Language Processing and Psychological Assessment Scales in AI-Driven Career Counseling Platforms
Year: 2026
Type: article
Abstract: Conventional career counseling mainly depends on manual psychological assessments and face-to-face interactions, which limit its scalability in the digital age. This study presents and validates an integrated framework that merges Natural Language Processing (NLP) with standardized psychological assessment scales to improve accuracy and personalization in AI-powered career counseling platforms. The system combines three components: open-text analysis using a fine-tuned Bidirectional Encoder Representations from Transformers (BERT) model, trait extraction from the Big Five Inventory (BFI-44), and vocational type classification based on Holland's RIASEC Self-Directed Search (SDS). An experiment involving 240 university students, who completed both psychological tests and open-ended career narratives, was conducted. The integrated model achieved a career-type classification F1-score of 0.847, surpassing the NLP-only baseline (F1 = 0.731) and the scale-only baseline (F1 = 0.768). These results show that combining linguistic signals with psychometric scores significantly improves predictions of Holland vocational codes. The framework provides a replicable architecture for intelligent career counseling systems that connect psychological theories with computational language modeling. This study contributes both theoretically and practically to the field of AI-driven career counseling: theoretically, it provides empirical evidence that fusing transformer-based linguistic features with validated psychometric scales advances person-environment fit theory into a computationally testable framework, while practically, it offers a replicable, open-source architecture that career-guidance platforms and counselors can adopt to deliver more accurate, personalized, and evidence-based vocational guidance at scale.
Authors: Septien Dwi Savandha, Elvira Fitriyanti
Language: English
Subfield: Applied Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W7204683007