Construction-focused University of Tokyo AI startup EdgeLab releases 'Construction-focused Knowledge Transfer White Paper 2026' -- Why traditional manuals fail and the latest approach to extracting expert judgment logic
EdgeLab, a UTokyo AI startup, released a white paper addressing the construction industry's knowledge transfer crisis, offering a novel AI-driven approach to extract experts' tacit knowledge.
📋 Article Processing Timeline
- 📰 Published: April 23, 2026 at 20:00
- 🔍 Collected: April 23, 2026 at 11:31
- 🤖 AI Analyzed: April 23, 2026 at 13:21 (1h 49m after Collected)
EdgeLab Co., Ltd. (Headquarters: Bunkyo-ku, Tokyo; Representative Director and CEO: Ryoyoshi Yagi; hereinafter "EdgeLab"), a University of Tokyo-originated startup developing knowledge transfer systems specialized for the construction industry, has released a white paper titled 'Construction-focused Knowledge Transfer White Paper 2026'. The paper explains the risk of knowledge loss associated with the retirement of skilled technicians and the latest strategies for formalizing tacit knowledge using generative AI.
■ The 'Challenges of Skill Inheritance' and 3 Backgrounds Facing the Construction Industry
In the current construction industry, workers aged 55 and over account for about 36%, while those under 29 make up only about 12%. The loss of know-how accompanying the mass retirement of skilled technicians has become an urgent issue. Furthermore, due to the following three backgrounds, technology inheritance relying on traditional OJT (on-the-job training) is being forced to be reevaluated.
Difficulty securing time for training
With the application of the upper limit on overtime work (the 2024 problem), it has become difficult to secure sufficient direct instruction time on-site.
Declining retention rate of young workers
The turnover rate within three years for new high school graduates exceeds 40%, creating a situation where time-consuming training investments are unlikely to lead to skill retention in the organization.
Risks of person-dependent instruction
The traditional 'learning by observation' training system relies on individual discretion, making it difficult to completely eliminate risks in standardizing quality and managing safety.
The fundamental cause of these issues lies in the fact that important know-how remains inside the heads of individuals as the 'tacit knowledge of experts'. However, with traditional methods such as creating manuals and conducting interviews, it has been difficult to extract and formalize the 'judgment logic' that the technicians themselves apply unconsciously.
■ Highlights of This White Paper
This document analyzes the structural issues of traditional knowledge extraction methods and systematically explains a 'practical approach using AI' as a solution. It summarizes the following content for persons in charge of promoting organizational knowledge transfer projects.
Three requirements for successful formalization
Unraveling the issues of traditional methods, it presents points to balance 'non-invasiveness', which minimizes the workload of busy skilled technicians, and 'deep diving' to reach the depths of know-how.
EdgeLab's unique 'AI Interview Strategy'
It discloses a method using interactive AI that understands specialized construction terms and on-site contexts to effortlessly extract judgment logic from experts and accumulate it as organizational data.
On-site utilization model by 'Co-creation AI Agent'
Rather than letting the extracted know-how end in mere documentation, it explains the realization image of an 'on-site AI assistant' that allows young technicians to reference necessary information in real-time during actual work.
In addition, the opening pages feature a special interview with Tadao Adachi, an EdgeLab advisor and leading VE (Value Engineering) expert who spearheaded facility design, construction management, company-wide productivity improvement, and operational DX at Kajima Corporation for 36 years. He shares front-line insights on the future of data-driven, next-generation skill inheritance.
The 'Construction-focused Knowledge Transfer White Paper 2026' is a guiding document for management and project promotion personnel facing issues with technology inheritance, on-site DX promotion, and organizational knowledge management to consider specific measures.
You can apply via the dedicated page below.
https://edgelab.co.jp/WhitePaper
About EdgeLab Co., Ltd.
EdgeLab develops and implements AI agents that formalize the tacit knowledge of experts in construction management and design in the construction industry. In addition to a service that extracts know-how in three steps ('recall', 'deep dive', and 'formatting') and works alongside construction companies to provide tools that young workers can use freely, they also develop products.
Company Profile
Company Name: EdgeLab Co., Ltd.
Location: 2-35-17-306 Hongo, Bunkyo-ku, Tokyo (Relocated Sep 2025)
Representative Director: Ryoyoshi Yagi
Established: November 2024
Official Website: https://edgelab.co.jp/
Inquiries regarding this matter
E-mail: info@edgelab.co.jp
■ The 'Challenges of Skill Inheritance' and 3 Backgrounds Facing the Construction Industry
In the current construction industry, workers aged 55 and over account for about 36%, while those under 29 make up only about 12%. The loss of know-how accompanying the mass retirement of skilled technicians has become an urgent issue. Furthermore, due to the following three backgrounds, technology inheritance relying on traditional OJT (on-the-job training) is being forced to be reevaluated.
Difficulty securing time for training
With the application of the upper limit on overtime work (the 2024 problem), it has become difficult to secure sufficient direct instruction time on-site.
Declining retention rate of young workers
The turnover rate within three years for new high school graduates exceeds 40%, creating a situation where time-consuming training investments are unlikely to lead to skill retention in the organization.
Risks of person-dependent instruction
The traditional 'learning by observation' training system relies on individual discretion, making it difficult to completely eliminate risks in standardizing quality and managing safety.
The fundamental cause of these issues lies in the fact that important know-how remains inside the heads of individuals as the 'tacit knowledge of experts'. However, with traditional methods such as creating manuals and conducting interviews, it has been difficult to extract and formalize the 'judgment logic' that the technicians themselves apply unconsciously.
■ Highlights of This White Paper
This document analyzes the structural issues of traditional knowledge extraction methods and systematically explains a 'practical approach using AI' as a solution. It summarizes the following content for persons in charge of promoting organizational knowledge transfer projects.
Three requirements for successful formalization
Unraveling the issues of traditional methods, it presents points to balance 'non-invasiveness', which minimizes the workload of busy skilled technicians, and 'deep diving' to reach the depths of know-how.
EdgeLab's unique 'AI Interview Strategy'
It discloses a method using interactive AI that understands specialized construction terms and on-site contexts to effortlessly extract judgment logic from experts and accumulate it as organizational data.
On-site utilization model by 'Co-creation AI Agent'
Rather than letting the extracted know-how end in mere documentation, it explains the realization image of an 'on-site AI assistant' that allows young technicians to reference necessary information in real-time during actual work.
In addition, the opening pages feature a special interview with Tadao Adachi, an EdgeLab advisor and leading VE (Value Engineering) expert who spearheaded facility design, construction management, company-wide productivity improvement, and operational DX at Kajima Corporation for 36 years. He shares front-line insights on the future of data-driven, next-generation skill inheritance.
The 'Construction-focused Knowledge Transfer White Paper 2026' is a guiding document for management and project promotion personnel facing issues with technology inheritance, on-site DX promotion, and organizational knowledge management to consider specific measures.
You can apply via the dedicated page below.
https://edgelab.co.jp/WhitePaper
About EdgeLab Co., Ltd.
EdgeLab develops and implements AI agents that formalize the tacit knowledge of experts in construction management and design in the construction industry. In addition to a service that extracts know-how in three steps ('recall', 'deep dive', and 'formatting') and works alongside construction companies to provide tools that young workers can use freely, they also develop products.
Company Profile
Company Name: EdgeLab Co., Ltd.
Location: 2-35-17-306 Hongo, Bunkyo-ku, Tokyo (Relocated Sep 2025)
Representative Director: Ryoyoshi Yagi
Established: November 2024
Official Website: https://edgelab.co.jp/
Inquiries regarding this matter
E-mail: info@edgelab.co.jp