TOMOMI RESEARCH CTO Dr. Seong-Hun Choe to Speak at TAMA Association AI and DX Seminar
kotai2003Manufacturing sites hold a great deal of "tacit knowledge," such as the experience of veteran staff and know-how for handling past projects. At the same time, knowledge that depends on specific individuals has become a challenge for passing on skills and improving operational efficiency.
Dr. Seong-Hun Choe, CTO and Doctor of Engineering at TOMOMI RESEARCH Inc., will speak at the AI and DX subcommittee hosted by the TAMA Association (Shutoken Sangyo Kasseika Kyokai, an industry promotion association for the Tokyo metropolitan area, general incorporated association), introducing an initiative that uses generative AI to turn on-site knowledge into a company asset.
We are pleased to report that the seminar is now fully booked. With interest in generative AI in manufacturing growing, we are truly honored to have been invited as a speaker presenting a "manufacturing case study."
Session overview
FY2026 1st AI and DX Subcommittee
"Case Studies of Applying Generative AI to Manufacturing
Turning Person-Dependent On-Site Know-How into Corporate Competitiveness"
| Item | Details |
|---|---|
| Organizer | TAMA Association (Shutoken Sangyo Kasseika Kyokai, general incorporated association) |
| Date and time | Friday, June 26, 2026, 15:00 to 18:00 |
| Venue | Conference Room 3, 3F, Tokyo Tama Mirai Messe |
| Our session | 16:10 to 17:00 |
| Title | Advancing Similar-Project Search with Generative AI × kintone |
On the day, case studies of applying generative AI to on-site operations in manufacturing will be presented, including turning tacit knowledge into shared knowledge, similar-project search, and standardizing price estimates and order decisions.
Similar-project search with generative AI × kintone
In manufacturing, knowledge such as "how we handled a similar project before" or "under what conditions we prepared the estimate" is often accumulated as the experience of veteran staff.
Even when past data is stored in a system, if the search method is complicated, the necessary information cannot be put to use easily.
In this talk, we will introduce a system that links the following three components.
- kintone: stores project information and business data
- Agentic AI: understands the user's question and intent
- SQL agent: converts the question into a database search and runs it
Staff on site can search past projects and business data simply by asking questions in everyday language, without having to learn specialized search methods or SQL.
For example, if you enter "Show me the sales ranking," the AI understands the intent of the question, searches for the necessary data, and returns the results in an easy-to-understand form. Using the same mechanism, it becomes possible to quickly find past projects with similar conditions and estimate information.
Expected benefits
This system is expected to deliver benefits such as the following.
- Turning veteran experience and past data into knowledge assets
- Reducing dependence on specific individuals
- Standardizing decisions and operations based on past projects
- Faster and more accurate estimates and similar-project searches
Moving from "you can't find out unless you ask that person" to "you can get the information you need by asking the system" is the first step toward turning tacit knowledge into corporate competitiveness.
*Results of implementation vary depending on the content of the data, the business environment, and other factors.
Speaker profile
Dr. Seong-Hun Choe
CTO and Doctor of Engineering, TOMOMI RESEARCH Inc.
After completing graduate school at Tohoku University, Dr. Choe worked on the development of MEMS devices at the Mechanical Engineering Research Laboratory of Hitachi, Ltd. Dr. Choe then led the launch of an AI-OCR inspection system for HDDs at Western Digital.
Dr. Choe currently drives the implementation of AI and DX in manufacturing, including the development of visual inspection systems that combine AI anomaly detection and image processing.
Summary
In this talk, we will introduce a system that links kintone, agentic AI, and an SQL agent to search past projects and business data in natural language.
The aim is to turn the tacit knowledge accumulated on the manufacturing floor into knowledge assets anyone can use, leading to less dependence on individuals, standardized operations, and faster decisions.
The content of the talk and the materials are scheduled to be published later on the TAMA Association's official website. We will also report back on our blog and social media.
TOMOMI RESEARCH Inc. works on visual inspection systems that use AI anomaly detection and image processing, as well as operational DX through generative AI and data utilization. If you are interested in AI-based quality inspection or improving the efficiency of on-site operations, please feel free to contact us.
