Internal Welfare Data Search & Chatbot for Employees

Overview

The regulations of the Seoul Welfare Foundation undergo frequent revisions due to constant changes in higher-level laws and municipal ordinances. To ensure that staff can accurately grasp and reflect these latest updates, the foundation launched a data search and chat service initiative in 2024. This project aims to enhance work convenience, reduce repetitive tasks, and improve overall productivity and accuracy.

As the demand for welfare systems to address low birth rates and an aging population grows, the timely application of the latest data and legislation is critical. However, because legal and administrative documents are often scattered in unstructured formats, finding necessary information requires significant time and effort. Therefore, the development of an advanced, intelligent search system and chatbot service became essential to streamline the foundation's operations.

Project description

For AI model training, the foundation collected 835 pages of internal regulations and 1,627 pages of relevant laws and ordinances (81 types). During the data pre-processing stage, non-standardized elements such as complex diagrams and images were filtered to optimize the structure for search indexing. Based on this refined dataset, a RAG-based Q&A service was developed.

The core service provides regulations, laws, and ordinances ranked by relevance and accuracy in response to queries about operational procedures or standards. This allows staff to verify the latest information instantly, with the system also supporting natural language answers for data contained within tables.

Additionally, a demo service was provided that allows users to test both search and chatbot functions simultaneously. This hybrid system matches user queries with pre-built chatbot knowledge first; if no direct match is found, it automatically switches to a document search to provide the most relevant information, thereby maximizing user convenience.

Key technological features

• RAG-based Q&A Service: Development of a search and Q&A chatbot service based on SearchGPT (a RAG solution), targeting internal regulation manuals, relevant laws, and local ordinances
• SmartChat Integration: Provision of a demo-version chat interface integrated with SmartChat to facilitate interactive user engagement

Expected outcomes

By reducing simple and repetitive tasks such as manual data searching and initial report drafting, this service allows employees to focus on core strategic tasks. The application of real-time updated information improves administrative accuracy and significantly shortens the learning curve for new employees.

Furthermore, the chatbot service ensures operational continuity by providing immediate responses even during the absence of specific personnel, thereby resolving potential work gaps. Ultimately, this intelligent system strengthens the foundation's capacity to deliver timely and accurate welfare services to the citizens of Seoul.

Launch year: 2024

Attribution

Case study courtesy of the National Information Society Agency, Republic of Korea, Apolitical's Content Partner. Sourced from Use Cases of Public AI Service, Vol.2, https://eng.nia.or.kr/site/nia_eng/ex/bbs/View.do?cbIdx=31975&bcIdx=29182&parentSeq=29232