Yeonam Institute of Technology — Legacy Groupware AWS Migration & AI-Powered Administrative Automation

About the Customer

Yeonam Institute of Technology, an engineering-specialized university founded by LG Yeonam Foundation, aims to cultivate field-ready professionals suited for the era of the Fourth Industrial Revolution, offering a practice-oriented curriculum optimized for science and engineering students. To maximize administrative efficiency and realize its vision of nurturing "New Collar" talent, the university migrated its legacy groupware IT infrastructure to AWS.

Customer Challenge (Problem Statement / Definition)

The university's existing on-premises groupware had reached its technological limits due to aging systems and lack of scalability, making it difficult to integrate the AI-driven modern technologies the university sought to adopt. This rigid infrastructure environment prevented the automation of repetitive tasks and led to serious administrative inefficiencies, including the following:

Inefficient document verification: Approving expenditure approval requests required manually cross-checking whether the amounts on draft documents matched those on attached receipts and quotations (PDFs and images).
Complex regulations and budget category errors: Faculty and staff drafting documents frequently lacked familiarity with frequently changing university regulations and budget guidelines — such as the "Innovation Project Budget Plan" — resulting in drafts submitted under incorrect budget line items. This led to repeated rejections and re-submissions, slowing administrative processing.
Simple typos and omissions: Human errors such as missing attachments or typos had to be caught manually, leading to high levels of fatigue.

The university set out to delegate routine approval reviews to AI, aiming for a "0% rejection rate," and targeted a transition to an "AI-powered intelligent campus" where faculty and staff could focus on high-value work.

Proposed Solution

NxtCloud executed the project on the AWS cloud environment following the Assess – Migrate – Modernize framework.

Phase 1: Assess

Cloud transition roadmap design: Conducted a thorough analysis of the aging on-premises infrastructure and designed an optimal AWS architecture capable of supporting future AI service expansion — going beyond a simple lift-and-shift.
Data modernization strategy: Analyzed the existing legacy database structure and validated database migration scenarios aligned with a cloud-native environment, enabling the AI Agent to access groupware data securely and efficiently.

Phase 2: Migrate

All-in Migration execution: Leveraged AWS Application Migration Service (MGN) to rapidly rehost over 10 groupware servers onto Amazon EC2, minimizing business disruption.
Database stability: Migrated the operationally burdensome on-premises PostgreSQL database to the fully managed Amazon RDS for PostgreSQL, improving data availability and laying the groundwork for future upgrade to Amazon Aurora.

Phase 3: Modernize

AI approval review agent: Built on Amazon Bedrock, implemented a system that automatically cross-checks amount consistency between draft documents and supporting materials (receipt and quotation images/PDFs) and validates complex budget regulations in real time.
Hyper-automation of administrative processes: Configured the system so that AI delivers a preliminary analysis the moment a draft is submitted, dramatically reducing rejection rates caused by simple errors or unfamiliarity with regulations, and creating an environment where faculty and staff can focus on high-value work.

Results and Benefits (Success Metrics)

[Elimination of Data Silos and AI-Driven Administrative Advancement through AWS Migration] The core achievement of this project lies in migrating the legacy system to AWS to eliminate data silos and establishing an integrated environment where AI can freely access administrative data. The cloud agility gained through migration has been demonstrated by the following tangible metrics:

83% reduction in approval review time: High-performance AWS instances now handle OCR and data cross-checking tasks that were difficult to process on the legacy servers, minimizing manual intervention by administrators and reducing review time from [30 minutes] to [5 minutes].
40% decrease in simple rejection rates: By leveraging flexible cloud resources, AI delivers real-time guidance without performance degradation even during peak draft submission periods, improving accuracy at the drafting stage.
Accelerated complex review workflows: For tasks requiring advanced judgment — such as restructuring admissions processes and executing industry-academia MOU agreements — AI analyzes relevant regulations and past cases to provide draft "Risk Review" and "feasibility assessment" reports. This has reduced the average time spent by staff on research and drafting from [72 hours / 3 days] to [30 minutes], enabling faster decision-making.

Source: https://news.unn.net/news/articleView.html?idxno=590805