[For more diverse company information on the startups mentioned in this article, please visit Unicorn Factory's big data platform 'Data Lab'.]

As generative AI (artificial intelligence) spreads across all industries, technologies that convert analog documents into digital data to improve work efficiency are gaining attention. In particular, even the financial sector, which had faced constraints on using generative AI due to security concerns, is now seeing regulatory easing movements. This has created new growth opportunities for 'Document AI' companies that support AI-based data construction and work automation based on OCR (optical character recognition) technology.
The Financial Services Commission announced last month a plan to improve network isolation measures to expand the use of generative AI in the financial sector. Initially, regulations will be relaxed only for security vulnerability checks and defense system construction using AI. However, if financial institutions can demonstrate sufficient security capabilities, there is a possibility that the scope of AI usage will expand to other business areas such as innovative financial services in the future.
The industry expects that as the financial sector's AI transformation (AX) gains momentum, demand for related infrastructure will also increase. In particular, for financial institutions to effectively utilize generative AI in actual operations, they must convert contracts, financial statements, credit-related documents, and various supporting materials into data formats that AI can understand. In this process, the importance of Document AI technology is expected to grow, as it goes beyond simply reading documents via OCR to classifying and analyzing them, extracting necessary information, and integrating with business systems.
In fact, the Document AI market has recently evolved from a stage focused on competing in character recognition accuracy to one that supports document understanding and work automation. In the financial sector, it is expected that demand will expand as these technologies can be applied to various tasks such as credit assessment, corporate finance, internal controls, and customer due diligence (KYC).

Domestic leading Document AI companies Korean DeepLearning and Upstage are also accelerating their efforts to target the financial sector. They are not only combining OCR technology with large language models (LLMs) and vision-language models (VLMs) to digitize documents but are also building platforms that automate actual business processes.
Korean DeepLearning is targeting the corporate document automation market with its proprietary platform, 'Deep Agent.' This platform recently achieved first place in the English category of 'OCR Bench v2,' a global benchmark where multimodal AI models compete, surpassing Google Gemini and OpenAI's GPT.
In particular, it highlights vertical solutions specialized for industries with many technical terms and unstructured documents, such as finance and manufacturing. 'Deep Agent for Finance' automates the classification, extraction, verification, and system input of supporting documents generated during loan assessment and credit operations. By meeting requirements for personal information anonymization and financial sector security standards, it has secured over 80 clients, including major corporations and institutional investors.
Upstage is expanding its market reach through 'Upstage Studio,' which combines LLM capabilities. This platform integrates document parsing (conversion), classification, information extraction, and question-and-answer functions. It particularly emphasizes its technical strength in converting complex unstructured documents containing tables, charts, and mixed multi-column layouts into structured data.
Upstage has already secured numerous cases of large-scale document processing implementation with major domestic financial institutions such as Samsung Life Insurance and Hanwha Life Insurance. In addition to domestic operations, it is also carrying out global document automation projects targeting U.S. insurance companies and institutional investors.
An industry insider stated, "Financial institutions that had hesitated to adopt due to security concerns are showing high interest in Document AI solutions that can be safely operated on internal networks (on-premises) following improvements in network isolation regulations." They added, "Moving beyond simple text extraction, integrated platforms that understand the context of complex unstructured documents and integrate with actual business systems to maximize efficiency will lead the market in the future."
[MoneyToday startup media platform Unicorn Factory]