Customer Story
Jeju Bank,
becoming a bank that works alongside AI
An XGEN-based generative-AI platform was built inside the bank, giving every employee a working AX environment where they can create AI agents themselves and put them to use in real work. From compliance and audit through consumer protection and sales support, AI agents are being embedded into bank-wide processes step by step.
- Internal generative-AI platform built
- Network-separated, on-premise operation
- Built by the business teams themselves
- Embedded across bank-wide processes
The Challenge
AI the business side can use, on top of the security it has to keep
Banking runs on strict security and network-separation requirements. In an environment where an external generative-AI service cannot simply be adopted as is, Jeju Bank had three problems to solve at once.
Adoption without loosening security
Apply generative AI to internal work while keeping the network-separated, on-premise environment intact
A tool the business side uses directly
Not only IT — business staff themselves had to be able to build and use AI
Embedded, not one-off
Settle into real bank-wide processes rather than stopping at a demo
The Solution
An AX platform employees build themselves
The XGEN-based generative-AI platform was built on-premise into the bank's own infrastructure, together with a working model in which business staff design, verify, and use AI agents themselves.
AI agents employees build themselves
- · Start easily by wiring nodes on the canvas, or from a template or tutorial
- · Create an agent through chat alone — describe the goal and be guided through the steps and tools needed to finish it
- · Upload internal documents, connect them to a knowledge collection, and use it as a Q&A agent right away
Wired straight into the work
- · Call an agent from the collaboration tool to answer a question, with the best-suited agent among several responding
- · Produce real work output, such as document drafts on standard templates
- · Extend the reach by linking internal systems through MCP and APIs
Technology applied
The XGEN-based Agentic AI platform was built into an on-premise, network-separated environment. Business staff create agents on the canvas, from templates, or through chat, and connect internal documents to knowledge collections (RAG) to use them for Q&A. MCP and APIs link internal systems and collaboration tools, while generation-stage controls such as PII masking and blocked-term filters and usage monitoring meet the security requirements of the financial sector.
What We Built
AI agents applied across six bank-wide areas
Bank-wide work was divided into six areas, a representative task was chosen in each, and AI agents were applied. Rollout continues across the bank around the lead agent in each area.
Compliance and audit automation
Continuous monitoring, KYC, selection of internal-control and audit items, and risk-based analysis support
Consumer-protection support
Drafting responses to voice-phishing cases and customer complaints, plus complaint statistics analysis
Sales and product efficiency
Automating reporting and analysis — branch opinion papers, daily briefings, new-product planning and research
Credit and market monitoring
Monitoring procurement bid openings and tenders, comparing and summarizing markets and products
Bank-wide shared agents
A shared support layer for work QA, document classification, and quality checks
Data-driven performance sharing
Agent-assisted data reports and performance analysis that make the state of the work visible
Past assisting individual tasks, into bank-wide process
From document drafting and data analysis through internal control and sales support, agents are being embedded into bank-wide processes step by step.
Outcomes
Work innovation, a culture of use, and room to expand — bank-wide
AI-driven work innovation, settled in
- · Standardized work and automated repetition through generative AI
- · Less time spent on drafting, research, and reporting
- · An environment where staff concentrate on higher-value work
A bank-wide culture of using AI
- · An AI platform anyone can use easily
- · A working culture built around AI agents
- · Rising AI literacy and a shift toward an AI-native organization
A base for extending AI services
- · Internal systems connected through MCP and APIs
- · Department-level agents developed on the shared AI platform
- · Data and AI assets accumulated and reused over time
Voice of Customer
In the customer's words
Head of the AI Innovation Team, Jeju Bank
Jeju Bank · Finance
Building a generative-AI platform was not simply a project to introduce a system — it was a process of changing how we work and how the organization thinks. It was not an easy journey, but XGEN's stable platform and the delivery team's fast, flexible response let us carry the project through successfully. What stood out was how quickly the varied requirements coming from the business side were reflected while the build stayed stable through to completion. Jeju Bank will now go a step further — past using AI, to become an AI Native Bank that works alongside AI.