Kuken Industries searches 70 years of drawings in minutes

Kuken Industries, a Japanese manufacturer of air conditioning and ventilation equipment with a leading share of Japan's cooling tower market, maintains an archive of engineering drawings spanning 70 years of installations. Previously, finding the right drawing for a replacement part could take half a day, and newer engineers often depended on experienced colleagues to locate and interpret records. Kuken partnered with Japanese AI startup and OpenAI partner Recursive to build a search system and field app using OpenAI's models and API, which performed well on accuracy and cost when tested on a range of drawings. Engineers can search by part number or upload an image or PDF of a drawing or parts list, and the app returns possible matches, including similar designs. Kuken expects the tool to cut drawing search time by about 90% over a year, reduce printing costs, and free experienced engineers to mentor colleagues. Newer engineers can now handle a wider range of jobs independently. Kuken is also developing an AI tool to check partner drawings against specifications; in prototype testing, about 90% of flagged issues were genuine discrepancies. The company aims to use the technology to pass engineering expertise on to the next generation.

(Source：OpenAI)