Ricoh, Kawasaki Heavy tie up to pass on engineers' know-how via AI

The partnership rests on a problem the two companies describe as common across manufacturing: large volumes of design documents and drawings sit underused, while the experience of veteran engineers stays locked up with individuals. With labor shortages and retirements advancing, the stated aim is a mechanism for passing that knowledge around the organization rather than leaving it to whoever happens to hold it.

Ricoh's side of the deal is a set of LLM and LMM models it says are strong at reading internal documents, plus an on-premises generative AI base and installation know-how. Hi.DEEN, announced in December 2025, is the enterprise platform the company has already put forward for using internal documents, drawings, reports and unstructured data. Kawasaki Heavy's contribution is its design and engineering knowledge from developing rail cars, ships, aircraft and robots, along with technical documents containing drawings, tables and graphs, business data and a test environment.

The work is framed as verification in an environment close to actual operations at Kawasaki Heavy, which handles complex products such as aircraft and robots. Going forward the two firms say they will accumulate AI know-how for more advanced design work and spread the resulting use cases, with the stated goal of an environment where engineers can concentrate on more creative tasks.