[DigitalToday reporter Ji-young Lee] PFCT (PFCTechnologies) is developing and moving to commercialise an "AI Agent Builder" that plans and executes financial tasks on its own. It links and automates a range of tasks from lending strategy to loan decision-making and risk management.

PFCT said on Tuesday it has developed a system based on its in-house AI lending solution, AIRPACK, that allows financial institutions to build and operate AI agents tailored to their business objectives.

The AI Agent Builder works by having multiple agents call up the data and tools they need and divide tasks according to the work goal. Staff review and approve results at key stages and control the final decision-making.

For example, after developing a credit risk model and setting a lending strategy in AIRPACK Lab, AIRPACK Studio supports strategy operations and real-time loan decisions after staff approval. AIRPACK Insight then analyses loan performance and delinquency and bad-loan risks and provides a report with improvements.

It is designed to operate on internal networks and in on-premises environments to meet financial institutions' security requirements. It also applies the open standard protocol MCP (Model Context Protocol) to support integration with external AI applications and existing financial systems.

The first application will be at Koryo Savings Bank. Koryo Savings Bank plans to introduce the AI Agent Builder to automate tasks such as reflecting and verifying lending strategies in systems and analysing causes of faults.

Kim Hwan-seok (김환석), head of the CSS Planning Team at Koryo Savings Bank, said, "We plan to enhance the efficiency of lending strategy operations and strengthen our risk response capabilities by linking data analysis, monitoring and decision-support with the AIRPACK Agent Builder."

Lee Jae-kyun (이재균), executive director and head of PFCT's financial business solutions division, said, "We will later combine it with the Entry solution of N2Soft, a specialised subsidiary for integrated lending management systems, to expand adoption of AI agents across the financial sector, including savings banks and capital companies."