Home/Newsletter/Seventy Years of Drawings Became a Search Box
Edition #31

Seventy Years of Drawings Became a Search Box

Dan Toma·November 10, 2026·4 min read
Key Takeaway

Kuken Industries turned 70 years of engineering drawings into a search tool that finds the right file in 4.5 minutes instead of nearly an hour. The model was a commodity choice. The archive was the asset, and most companies are sitting on an equivalent archive they have never made searchable.


FAQ

How did Kuken Industries use AI to search engineering drawings?

Kuken worked with Japanese startup Recursive to build a search tool on the OpenAI API covering 70 years of cooling tower drawings. Engineers can search by part number or upload a screenshot of a parts list. Finding a drawing dropped from nearly an hour to 4.5 minutes.

What company data is most valuable for AI search tools?

Archives tied to repeated, high-value questions are the best starting point. Examples include past proposals and win notes, campaign results, support tickets and engineering records. The data matters more than the model, because models can be swapped while proprietary history cannot.

Do you need to clean all your data before building an AI search tool?

No. Start with one narrow question and prepare only the data that question needs. Kuken made existing records searchable by identifiers engineers already used, such as part numbers, rather than reorganising its entire archive first.

Subscribe to The Weekly Vibe

Every Tuesday. 5-7 original takes on what matters in AI, Marketing, and Business Growth. No spam, no fluff, unsubscribe anytime.