Kuken uses OpenAI API to search 70 years of drawings in minutes
Original title: Kuken Industries searches 70 years of drawings in minutes
Engineers now locate a replacement-part drawing in 4.5 minutes, down from nearly an hour, via a Recursive-built app.
Kuken Industries, a Japanese maker of cooling towers and ventilation equipment, digitized a 70-year drawing archive but still needed experienced staff to locate the right file. With Recursive, an OpenAI partner, it built a field app that accepts a part number, image, or PDF and returns matching drawings. Kuken projects about a 90% cut in annual search hours. A separate prototype that flags mismatches in partner drawings is not yet in production.
Unconfirmed: Search-time savings and the 90% discrepancy rate are Kuken’s projections and prototype results, not independently verified.
- Kuken holds one of the largest shares of Japan’s cooling tower market.
- Paper searches once took half a day before the archive was digitized.
- OpenAI models were chosen after tests of extraction accuracy and cost.
- Kuken expects about a 90% yearly cut in drawing-search hours.
- Prototype drawing checks flagged genuine discrepancies in about 90% of cases, Kuken says.
- Vendors
- OpenAI
- Products
- OpenAI API
- Organizations
- Kuken IndustriesRecursive
- Countries
- Japan
From the original article
6 min read at openai.comKuken Industries uses the OpenAI API to search 70 years of cooling tower drawings. Engineers can find the drawing for a repair in 4. 5 minutes, down from nearly an hour.
openai.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.