News
Nov 6, 2026Company
A feature article by Masahiro Taima on the future of building review has been published in BELCA NEWS
Automatically translated from the Japanese original.
Magazine Coverage
An article by our CEO, Masahiro Taima, has been published in the October 2026 issue (No. 197) of BELCA NEWS, the journal of the Building and Equipment Long-life Cycle Association (BELCA), as part of its special feature "Construction and Generative AI." The piece, titled "AI in Building Confirmation Review: Where We Stand and Where We're Headed — An AI Review Support System Seen from Both Practical and Technical Perspectives," explores where AI stands today in building confirmation review and where it is headed.
About BELCA and BELCA NEWS
The Building and Equipment Long-life Cycle Association (BELCA) is a public interest incorporated association dedicated to promoting long-life buildings through proper maintenance, conservation and renovation. BELCA NEWS is the association's journal; each issue features the latest insights on maintaining and extending the life of Japan's building stock.
Visit the BELCA NEWS information page
Overview of the Article
The article examines the review support system, powered by an industry-specific AI agent, that Bureau Veritas Japan and mign have been using in live building confirmation and inspection work since 2026. It approaches the topic from two angles: the practical perspective of the building inspectors who use the system day to day, and the technical perspective of mign, which developed it.
On the practical side, the article shares what the system has delivered in real-world use — for example, verifying use restrictions under Article 48 of the Building Standards Act, and flagging a calculation error in supporting documents while also suggesting its likely cause. It is equally candid about the areas where today's AI still struggles, particularly in interpreting visual information: misreading figures on drawings, or failing to distinguish solid lines from dashed ones. On the technical side, it explains the difference between a standalone generative AI model and an industry-specific AI agent, and walks through how the system works — from reading the application documents to identifying the applicable regulations, making determinations and presenting the supporting rationale, all feeding into the inspector's final judgment. The article closes with a look ahead at how much decision-making can reasonably be entrusted to AI, and how the role of the reviewer is likely to evolve as AI becomes part of the process.
Authors
Masahiro Taima (CEO, mign Inc.)
mign will continue to work hand in hand with building confirmation and inspection bodies and other practitioners in the field to advance the use of AI across the architecture sector.