
How Kuken Industries Uses AI to Search 70 Years of Drawings

Kuken Industries makes air conditioning and ventilation equipment for large buildings, shopping centers, and factories, and it holds one of the largest shares of Japan’s cooling tower market. When a replacement part is needed, engineers search a drawing archive spanning 70 years of installations. On paper, a search could take half a day; after digitization, engineers still needed to know which name or number to look up and where to find the file. Newer engineers often had to ask experienced colleagues for guidance.
To let engineers at every experience level search on their own, Kuken worked with Recursive, a Japanese AI startup and an official OpenAI partner, to build the search system and its field app. They tested several AI models on a range of drawings, measuring how accurately each extracted details and at what cost. OpenAI’s models performed well on both, and the OpenAI API gave the team flexibility to build search directly into the app. Field engineers shaped the workflow: they can enter a drawing or part number, upload an image or PDF, or paste a screenshot of a parts list to avoid entering each number by hand. The app reads the part numbers and returns possible matches, including drawings for similar designs, and engineers verify against tower specifications and past replacement records.
Kuken reports that the tool is expected to cut the hours spent searching for drawings by about 90% over a year. Experienced engineers use some of the time saved to mentor newer colleagues, and engineers can review drawings in the app, reducing printing costs. Newer engineers now handle a wider range of jobs on their own, while veterans have more time to teach and take on their own engineering work. Takatsugu Naraki, Director and Factory Manager at Kuken Industries, said, “People now have time to do work they’ve wanted to try and take on new challenges. That’s been a real source of motivation.”
Kuken is also developing a tool to check drawings prepared by partners. Engineers currently compare each drawing with specifications by eye, looking for mismatched details or quantities. In prototype testing, about 90% of the issues the AI flagged were genuine discrepancies, and Kuken expects the tool to substantially reduce drawing-check time if it moves into production. The broader goal is to pass engineering experience to the next generation by making the decisions recorded in old drawings easier to find and understand.

