Updated
Updated · OpenAI · Nov 9
Kuken Industries Cuts 70-Year Drawing Searches by 90% With AI Tool
Updated
Updated · OpenAI · Nov 9

Kuken Industries Cuts 70-Year Drawing Searches by 90% With AI Tool

1 articles · Updated · OpenAI · Nov 9

Summary

  • Kuken Industries has deployed an AI search system that lets engineers find replacement-part drawings from a 70-year archive in minutes instead of spending up to half a day searching.
  • The tool was built with Japanese startup Recursive using OpenAI models after Kuken found digitized files still required engineers to know exact names, numbers and file locations.
  • Engineers can search by drawing or part number, or upload a screenshot, image or PDF; the app extracts part numbers and returns likely matches, including similar designs.
  • Kuken says the system should cut annual drawing-search time by about 90%, help newer engineers handle more jobs independently and reduce printing by enabling in-app review.
  • The company is also prototyping AI drawing checks for partner submissions, with tests showing about 90% of flagged issues were genuine discrepancies.

Insights

Could relying on AI to instantly find and verify decades-old engineering drawings erode the foundational problem-solving skills of younger engineers?
How does Kuken's AI accurately interpret faded, handwritten annotations on 70-year-old paper blueprints without making critical errors?
Will AI-driven retrieval of legacy blueprints become the new standard for modernizing aging infrastructure like data centers and power plants?