Korea Deep Learning supplies 'DEEP Parser' to Gyeonggi Provincial Government's generative AI platform.

Korea Deep Learning (CEO Jihyun Kim) , a company specializing in visual intelligence AI solutions for the public and corporate sectors, announced on the 26th that it has supplied its document structure analysis solution, "DEEP Parser," to the Gyeonggi Provincial Government's "2025 Generative AI Platform Construction Project." This project is being implemented in collaboration with MT Data, a company specializing in AI, big data, and cloud computing.

This project, which is being implemented with a total scale of 13.1 billion won, is the first case among local governments nationwide to fully introduce generative AI technology across administrative tasks. The goal is to establish an AI-based work support system centered on practical tasks such as document creation, meeting management, and information search over the course of approximately 11 months.

Korea Deep Learning's DEEP Parser is a solution that converts unstructured documents like HWP and PDF into structured data for AI utilization. Targeting unstructured format documents, which account for over 90% of Gyeonggi Provincial Government documents, it digitizes and structures administrative documents and policy data, enabling automated document processing and information utilization.

DEEP Parser's unique ability to simultaneously analyze the meaning and structure of documents, based on Vision Language Model (VLM) technology, is key. Unlike conventional OCR technologies that focus on simple text extraction, it precisely recognizes complex document elements such as tables, charts, hierarchies, and merged cells, thereby building a knowledge database usable by AI.

By utilizing this technology, Gyeonggi Provincial Government will be able to implement functions such as ▲precise digitization and structuring of unstructured documents ▲information extraction and refinement of policy documents and reports ▲summarization of meeting minutes and automatic extraction of key contents ▲document-based question-and-answer AI service.

Kim Ji-hyun, CEO of Korea Deep Learning, said, “This project is significant as a practical application case for digital transformation in the public sector,” and added, “DEEP Parser is a core technology that goes beyond simple document recognition and builds a knowledge base that AI can understand and utilize. We will continue to contribute to the advancement of the generative AI ecosystem.”

• See more related articles