The Credit Counseling and Recovery Service (CCRS) has launched a project to build an integrated artificial intelligence (AI) platform. The agency plans to upgrade its counseling and operational support systems based on generative AI and retrieval-augmented generation (RAG) technologies and expand debtor support functions through data-driven customer analysis models.
According to the financial industry on the 15th, the CCRS is currently conducting preliminary technical negotiations ahead of selecting a preferred bidder for the "CCRS Integrated AI Platform Construction" project. The agency has allocated a maximum budget of 6.3 billion won for the project and plans to carry it out over approximately seven months.
The project aligns with the government's public sector AI transformation (AX) initiative. Presenting a vision of becoming a "digital-based comprehensive debtor support hub," the CCRS has set core tasks including the introduction of generative AI services, the establishment of data-based customer analysis models, the development of AI infrastructure, and the enhancement of AI organizational capabilities and governance.
The core of the project is internal operational innovation using generative AI and RAG technologies. The CCRS first plans to build a "committee-specialized generative AI" trained on debt restructuring regulations, counseling cases, and internal guidelines. Under this system, AI will locate relevant supporting documents and provide answers, eliminating the need for counseling staff to manually search through extensive rulebooks.
A real-time speech-to-text (STT)-based counseling assistant will be introduced at the call center. The function analyzes counseling content in real time and presents relevant regulations and counseling scripts. As current counseling work largely depends on individual staff capabilities, the CCRS aims to reduce variation in counseling quality and improve operational efficiency.
The agency is also pursuing the construction of AI-based customer analysis models. The CCRS plans to develop a model predicting the likelihood of next-installment payments, a model recommending optimal debt restructuring programs, and a model predicting delinquency among small-loan users. The aim is to move beyond simple statistical analysis and establish a data-driven decision-making system.
The platform will operate in an on-premises (closed-network) environment isolated from the external internet. Considering the institution's handling of personal and financial information, the agency plans to build masking functions for personal data such as resident registration numbers and account numbers, along with a document access control system. The CCRS will also establish AI guidelines to mitigate AI-related ethical and security risks, including the misuse of personal information.
"Our goal is to enhance internal operational efficiency and counseling quality through AI and to build a faster debtor support system," a CCRS official said. "We plan to go beyond reducing simple repetitive tasks and strengthen our data-based analytical capabilities to upgrade support functions for ordinary citizens and vulnerable borrowers."
