Naver Cloud and JB Financial Group Team Up to Apply AI Across Corporate Lending
Naver Cloud signed an MOU with JB Financial Group-Gwangju Bank, Jeonbuk Bank, and JB Woori Capital-on Dec. 26 to drive AI-enabled innovation across lending and operations. Leaders from all four organizations attended the signing, including Lee Jae-gwan (JB Woori Capital), Kim Yu-won (Naver Cloud), Tak Hyeong-jae (Jeonbuk Bank), and Byeon Dong-ha (Gwangju Bank).
The collaboration centers on Naver Cloud's HyperCLOVA X and AICC, with the two sides planning to co-develop AI research and sector-specific models for finance. The focus: shorten cycle times, improve consistency, and create audit-ready documentation without adding operational drag.
What the partnership covers
- Discovery of AI use cases using HyperCLOVA X and AICC across the lending lifecycle.
- Co-development of specialized AI models for banking and capital operations, informed by best practices in AI for Operations.
- Long-term roadmap to extend AI to additional business lines beyond corporate lending.
How AI will be applied across the lending lifecycle
- Consultation: Extract and standardize data from intake notes and supporting documents to reduce manual prep and re-keying.
- Screening: Summarize applications, financial statements, and transaction histories to support credit decisions with consistent, comparable views.
- Post-decision: Auto-generate approval rationales to strengthen transparency, enable better audit trails, and support reviewer oversight.
In short, AI will assist bankers throughout corporate lending-from first touch to monitoring-while preserving human judgment where it counts. A Naver Cloud representative said the goal is to advance AI services that fit the needs of financial institutions and improve operational efficiency.
Why this matters for finance teams
- Faster turnaround with cleaner inputs: Standardized data reduces back-and-forth and enables apples-to-apples analysis.
- Higher documentation quality: Machine-generated rationales make decisions easier to audit and review.
- More consistent screening: Summaries built on the same templates help reduce variance across analysts and regions.
- Better monitoring: Structured data from day one improves ongoing portfolio surveillance and early warning signals.
Execution checklist for banks and lenders
- Data governance: Lock down PII, define retention rules, and ensure model inputs/outputs are fully traceable.
- Model risk management: Document objectives, limitations, validation plans, and performance thresholds in line with supervisory guidance such as SR 11-7.
- Fair-lending and bias controls: Test for disparate impact, maintain challenger models or rules, and keep human-in-the-loop for material decisions.
- Compliance and audit: Archive prompts, outputs, and decision rationales with time stamps and versioned model references.
- KPIs to track: Application-to-decision time, rework rate, exception rate, override rate, approval rationale completeness, and post-approval defect rate.
- Integration: Plan connectors into document systems, core banking, and risk engines; avoid isolated pilots with no path to production.
- Change management: Train front-line teams, set usage guardrails, and establish a feedback loop to improve prompts and policies; consider role-specific training like the AI Learning Path for Business Unit Managers.
For finance leaders, the win is simple: faster, more consistent credit decisions with stronger controls and clearer auditability. The partnership sets a practical path to apply large language models where they deliver measurable outcomes without compromising oversight.