Kuken Industries searches 70 years of drawings in minutes

Kuken Industries, a major Japanese manufacturer of air conditioning and ventilation equipment, holds a large share of Japan's cooling tower market. Engineers needing replacement parts for older cooling towers previously relied on experienced colleagues to navigate an archive of drawings spanning 70 years. Searching the paper archive could take half a day, and even after digitization, engineers had to know exactly what to look up. Kuken partnered with Japanese AI startup Recursive, an official OpenAI partner, to build a search system and field app. After testing several AI models, they chose OpenAI's models for their accuracy and cost-effectiveness, and used the OpenAI API to integrate search directly into the app. Engineers can search by drawing or part number, or upload an image or PDF of a drawing or parts list; the app reads part numbers from screenshots and returns possible matches, including similar designs. Kuken reports that the tool is expected to cut drawing search time by about 90% over a year, free experienced engineers to mentor newer colleagues, reduce printing costs, and enable newer engineers to handle more jobs independently. Kuken is also developing an AI tool to check partner-prepared drawings against specifications; in prototype testing, about 90% of flagged issues were genuine discrepancies. The company aims to use AI to pass engineering expertise on to the next generation.

(Source：OpenAI)