Deepfaine, an industrial AI platform company based on spatial AI, is developing a smart glasses-based AI research support system to drive the digital transformation of preclinical trial research environments.
Deepfaine announced Monday that it is participating in the "2026 AI Voucher Support Program" promoted by the Ministry of Science and ICT (MSIT) and the National IT Industry Promotion Agency (NIPA). The AI Voucher Support Program is a government initiative that supports companies' AI technology development and its application in industrial settings. The lead institution for the project in which Deepfaine is participating is Eunpyeong St. Mary's Hospital at the Catholic University of Korea. Deepfaine will develop the system and conduct field verification by advancing its smart glasses-based work collaboration solution "DAO" to suit preclinical trial environments.
Through this project, Deepfaine is developing a preclinical trial-specific AI research support system that combines a retrieval-augmented generation (RAG)-based AI agent with vision AI monitoring functions. The goal is to convert a research environment that has relied on analog documents and researcher proficiency to an AI-based one, supporting the standardization of research procedures, improved information accessibility, and enhanced research productivity and safety.
The core of the system is a RAG-based AI agent. Deepfaine builds an AI knowledge base by refining and structuring research site documents such as laboratory standard operating procedures (SOP), equipment manuals, safety guidelines, and accident response procedures. When a researcher wearing smart glasses asks a question by voice, the AI agent searches related documents and procedures, generates evidence-based answers through a large language model (LLM), and provides guidance on screen and by voice. This allows researchers to immediately check necessary procedures and precautions while keeping both hands free during experiments. Without checking a separate PC or paper documents, they can receive on-site guidance on experiment procedures by step, equipment usage, safety rules, and accident response methods.
The company is also developing a vision AI-based laboratory animal monitoring function. This function analyzes video from laboratory animal breeding cages to identify laboratory animals at the individual level and track their behavior. It complements the limitations of existing methods that involve marking laboratory animals directly or inserting devices, such as tail marking, ear tags, and radio frequency identification (RFID) chip insertion, and supports monitoring the condition of laboratory animals in a non-invasive manner.
When a researcher points smart glasses at experimental equipment, the vision AI recognizes the equipment and automatically provides related information such as usage procedures, safety rules, and inspection items. In addition, the research site and external experts can share the same screen in real time through smart glasses and jointly check equipment status or experiment procedures.
Deepfaine plans to optimize the system for a domestic AI semiconductor-based environment and link it with cloud services to secure a foundation for applying AI solutions that can be stably used in actual research sites. Through this, it will improve performance and increase service scalability using research institution data, implementing on-site AI services specialized for the preclinical trial field.
"This project will become a case that changes how knowledge is searched and procedures are performed at research sites by combining smart glasses, vision AI, and an LLM/RAG-based AI agent," Deepfaine CEO Kim Hyun-bae said. "Through verification at a university hospital, we will validate Deepfaine's on-site AI agent technology competitiveness, and based on this, we plan to expand the scope of applying our industrial AI platform to various research sites such as preclinical trial institutions and pharmaceutical and bio research labs."
