Credit C-GPT: A Domain-Specialized Large Language Model for Conversational Understanding in Vietnamese Debt Collection

TL;DR

Credit C-GPT addresses the challenge of understanding multi-turn, emotionally charged Vietnamese debt-collection conversations by fine-tuning a 7B-domain specialized LLM to jointly predict emotion, sentiment, intent, call stage, and slot-value information. It adopts a unified, instruction-tuned model based on Qwen2.5-7B, trained with QLoRA on proprietary simulated data to operate in real-time and post-call analytics with on-premise deployment. The results show it outperforms a traditional BERT-based pipeline and remains competitive with GPT-5, while reducing deployment risks and privacy concerns in BFSI settings. This work demonstrates practical, domain-adaptive fine-tuning for enterprise conversational analytics and highlights the value of unified modeling for complex, context-rich spoken dialogues in Vietnamese debt collection.

Abstract

Debt collection is a critical function within the banking, financial services, and insurance (BFSI) sector, relying heavily on large-scale human-to-human conversational interactions conducted primarily in Vietnamese contact centers. These conversations involve informal spoken language, emotional variability, and complex domain-specific reasoning, which pose significant challenges for traditional natural language processing systems. This paper introduces Credit C-GPT, a domain-specialized large language model with seven billion parameters, fine-tuned for conversational understanding in Vietnamese debt collection scenarios. The proposed model integrates multiple conversational intelligence tasks, including dialogue understanding, sentiment recognition, intent detection, call stage classification, and structured slot-value extraction, within a single reasoning-based framework. We describe the data construction process, annotation strategy, and training methodology, and evaluate the model on proprietary human-annotated datasets. Experimental results show consistent improvements over traditional pipeline-based approaches, indicating that domain-specialized conversational language models provide a scalable and privacy-aware solution for real-time assistance and post-call analytics in enterprise contact centers.