[IT Donga x Korea Technology Venture Foundation] We bring you the stories of startups dreaming of a leap forward. The Korea Technology Venture Foundation is a global technology commercialization organization that supports incubating, accelerating, technology marketing, and global networking, creating the value and future of companies. Artificial intelligence has set the world abuzz this year. It analyzes and searches vast amounts of data in an instant, answers questions, and finds the most suitable way to solve problems. It has even shown off creative abilities, writing text, drawing pictures, and coding programs like a human. AI does not only create new things; it also shows the ability to improve existing ones. Applying AI to existing industries raises utility and efficiency together and creates new value. This is why the major industries leading the world, such as semiconductors and secondary batteries, are racing to adopt AI. In Korea, the startup 'AIBIZ' is researching this technology. AIBIZ CEO Ha Seung-jae has a unique background. He majored in livestock industry at university and worked as a smart factory manager at Harim Group. At that time, he reportedly predicted that smart technology and data would change the landscape not only of factories but of all industries. After retiring, he entered the Soongsil University Graduate School of Business as a late-in-life student, and then took up a doctoral program researching AI in order to make his prediction a reality. While researching and analyzing the data held by machinery, equipment and devices, he devised a way to dramatically raise their operating efficiency with AI. Combining his long experience in smart factory process management with the new knowledge he acquired at Soongsil University, he founded AIBIZ. It was also then that he gathered the staff who would research and develop AI together. CTO Choi Hyeon-jin is a Ph.D. in AI engineering and an AI design specialist with 10 years of experience. CIO Lee Jong-seop lends strength to AIBIZ with his experience as a Soongsil University professor and 25 years of IT consulting. CEO Ha Seung-jae first analyzed AI technologies and papers in Korea and abroad, and on that basis filed and registered patents. He has registered three patents, including a technology for precisely sensing process equipment in smart factories and a technology for analyzing the causes of defects in semiconductor device manufacturing processes, and has filed eighteen more. He then visited smart factory managers, listened to their voices, and reflected them in the technology. As a result, AIBIZ developed a platform that manages semiconductor manufacturing processes with AI and a solution that handles semiconductor device R&D with AI. It also introduced an AI solution that optimizes vaccine culture processes, a machine vision function for sorting defective products, and a big-data-based demand forecasting AI solution. All of them are industrial AI technologies that can be applied to existing industries to raise utility and efficiency together and create new value. The operating principle of AIBIZ's industrial AI technology is simple. It richly collects and precisely analyzes time-series data, the data generated by industrial equipment that changes moment by moment over time. Take the process of a smart factory that makes semiconductors as an example. When operating semiconductor production equipment, various data such as temperature changes, the presence of vibration, process speed, output, and defect rate are recorded over time. By collecting and analyzing this data, it is possible to identify what change occurred in which process and whether it caused defects. After finding and eliminating the cause of the change in that process, the equipment is operated again while time-series data is collected. Repeating this removes the potential for defects and raises both output and yield. The types of equipment in a smart factory differ depending on the characteristics and quantity of the products to be made and the scale of the process. Equipment for making semiconductors differs from equipment for making secondary batteries. AIBIZ's strength is quickly and accurately collecting and analyzing the time-series data of the numerous pieces of equipment that differ from one smart factory to another. Thanks to this, it is effective not only in general production factories but also in places equipped with advanced facilities, such as pharmaceuticals, semiconductors, and secondary batteries. Also notable is that it changed the way of thinking and improved the efficiency of data labeling (processing). Usually, when teaching data to AI, abnormal data is labeled. AIBIZ does the opposite and teaches the AI using normal data. It watches whether things operate normally in ordinary times, and when a defect occurs, it acquires that data, labels it, and teaches it. The advantage of this structure is that it can be put into industrial sites right away without going through data labeling work that takes years. It also has the advantage of immediately identifying equipment defects. A semiconductor process consists of more than 200 pieces of equipment. Detecting which part of which equipment has a defect is very difficult. Because AIBIZ detects defective data based on normal data, it detects and responds to this immediately. Furthermore, AIBIZ has prepared technology to grasp smart factory equipment data more accurately and analyze it intuitively. Equipment ages with long use, and the time-series data changes little by little accordingly. Then, when part of the equipment or a component is replaced, the pattern of the time-series data changes once again. AIBIZ's solution detects all such fine changes and reflects them in the data. In addition, it added sensors to turn the changing patterns of time-series data into graphs, and also built a deep learning model that teaches these to the AI. CEO Ha Seung-jae introduces AIBIZ's industrial AI as 'a technology that helps us see what we could not see until now.' He introduces it as a technology that clearly presents data that people could not perceive because it is too vast, not properly labeled, complex, or changing frequently. He stresses that on this basis it raises the completeness of smart factory processes and reduces defects, and raises yield and output together. AIBIZ has already applied industrial AI to major Korean semiconductor companies, solving problems and securing yield. However, this technology can be applied only when all of a smart factory's equipment and sensors are connected. Technology to integrate and manage the differing data of each piece of equipment and sensor is also needed. AIBIZ is developing not only industrial AI but also technology that integrates and manages the data of smart factory equipment and sensors. Then companies will be able to apply industrial AI to smart factories with much less burden and see effects right away. The technology soon turned into results. AIBIZ received the top rating in the Pre-Startup Package, an excellent rating in the Initial Startup Package, and a good rating in the Leap Package in succession. In 2021, it won the Grand Prize at SUJI IR-DAY with its 'defect AI solution using sensor data of semiconductor equipment.' It won an Encouragement Award at the 2022 Intellectual Property Startup Competition, followed by the Grand Prize at the 17th Korea Excellent Patent Awards 2023 in the electronics/IT category, and the AI Grand Prize at the 2023 Korea Excellent Company Awards in the AI semiconductor process category. It made its name known overseas by being selected as an AI semiconductor company in the KOTRA Korea Pavilion. In September, it supplied technology to a Japanese semiconductor company and a major Korean battery company, recording sales as well. AIBIZ recently decided to lend its strength to the 'AI+X Cluster' led by Yangjae AI Hub, Guro G-Valley, Soongsil University and Dongjak-gu. CEO Ha Seung-jae himself earned his Ph.D. in AI at Soongsil University. Receiving mentoring from Soongsil University faculty and a stage-by-stage growth support program, he laid the foundation for AIBIZ's growth. Having completed the Seoul National University AI CEO Academy, he also collaborated several times with companies affiliated with Yangjae AI Hub, improving the maturity of the technology. To repay this, he decided to join the AI+X Cluster and spread success stories and technology. CEO Ha Seung-jae also says he is grateful to the Korea Technology Venture Foundation. He says that thanks to their support for technology development and commercialization, AIBIZ was able to develop an AI platform for semiconductor manufacturing process management and enter overseas markets. He added that the mentoring and education programs to strengthen AIBIZ's technology and business capabilities were also helpful, and said he would repay the Korea Technology Venture Foundation by growing into a leading company in the AI+X field. Building on these results, AIBIZ is tackling several challenges. First, it will spread industrial AI and apply it to various industries. This is not easy. The concept and utility of industrial AI are still not well known in the industry. So many companies are reluctant to introduce this technology into their processes. Many also doubt that a Korean startup has created a technology that even the world's large corporations have not yet actively used. CEO Ha Seung-jae's strategy is 'references.' He builds up success stories by collaborating with advanced technology companies in Korea and abroad. Using these cases, he removes the prejudice attached to industrial AI and raises the likelihood of its adoption in industry. He is also not neglecting 'collaboration with other AI companies.' He plans to join hands with companies that believe in AI's growth potential and have developed solid technology, and to target the industrial AI market, which is close to barren land. AIBIZ recently established a U.S. branch with the help of the KOTRA San Jose IT Center. It is a forward base for supplying AI solutions to U.S. semiconductor foundries, secondary battery and solar companies. CEO Ha Seung-jae said he has already introduced the technology to several major local companies and that they will soon verify the concept together. CEO Ha Seung-jae said, "If AI is used to raise both the efficiency and the yield of semiconductor production processes, value on the scale of trillions of won will be created. It will show great power in Korea's main industrial sectors such as secondary batteries and solar panels. AIBIZ will be a company that applies AI to Korea's main industries to raise operating efficiency and helps them gain the competitiveness to compete in the world." IT Donga reporter Cha Ju-kyung (racingcar@itdonga.com)