LUCY CI
Lucy (Lucy), a specialist in financial data infrastructure, said Thursday that it has been selected as the final candidate for an OpenAI collaboration track in the ‘Yeollim’ global corporate collaboration program, jointly hosted by the Ministry of SMEs and Startups and the Korea Institute of Startup and Entrepreneurship Development.
The selection is significant because it recognizes Lucy’s distinctive financial data preprocessing technology as being on a global standard. Lucy has led efforts to precisely structure the vast disclosure data from the U.S. Securities and Exchange Commission (SEC) and Korea’s Electronic Disclosure System (DART) to fit large language model (LLM) and retrieval-augmented generation (RAG) environments.
As adoption of AI in the finance sector accelerates, Lucy provides an advanced data infrastructure that goes beyond simple text transformation and allows machines to understand data perfectly. Major domestic banks and securities firms have already delivered outstanding results using Lucy’s technology in analyzing pension funds and exchange-traded funds (ETFs). In particular, by flawlessly preprocessing complex offering memorandums, Lucy has helped produce in-depth investment insights such as a company’s theme classification, performance by business segment, and the connectedness among disclosures.
Through its collaboration with OpenAI, Lucy plans to verify from multiple angles how its data preprocessing infrastructure affects real-world LLM and RAG performance. It also plans to adopt the latest methodologies that use LLMs as evaluators, and to quantitatively demonstrate the superiority of its technology in terms of accuracy, evidentiality, reproducibility, and data quality—key priorities for financial institutions.
Performance in global markets is also coming into view. Lucy has secured a U.S.-based asset manager as a customer and is poised to record export sales of about $1.08 million by 2026. In Korea as well, it is strengthening service completeness by supplying customized investment information data to major securities firms.
Park Ji-hoe, the head of AI development at Lucy, emphasized that for LLMs to operate successfully in the financial sector, the quality of training data is absolutely critical, as much as—and perhaps more than—the model’s performance. He said his company has focused its enterprise-wide capabilities on building an infrastructure that enables AI to accurately decode complex financial disclosures, and he added that, using this program as a stepping stone, Lucy aims to provide the most trustworthy AI infrastructure to financial institutions at home and abroad.
Going forward, Lucy plans to significantly expand its data coverage centered on key analysis targets for financial practitioners, including SEC filings, ETF and ETP documents, and regulatory documents. Through that, it intends to strongly support asset managers, banks, and others in reliably adopting AI into core tasks such as investment-information chatbots, research automation, and regulatory analysis.