Without a way to locate similar past drawings, teams designed new parts from scratch each time. Part numbers multiplied as a result. Locating past drawings heavily required veteran instinct and know-how, and with those veterans nearing retirement, preventing the loss of that expertise became urgent.
A "check with AI and CADDi first" mindset has taken hold across the team, expanding the scope of work employees can handle on their own and speeding up new-hire training and collaboration across departments. Drawing data is now searchable at scale, so the team can group parts suited for process rationalization instantly and request consolidation and review right away.

Recognizing when instinct and experience aren't enough
YKK's competitive edge comes from its commitment to fully integrated in-house production, handling everything from melting raw materials to manufacturing its own machinery. But Matsui, Managing Executive Officer and General Manager of the Machinery Manufacturing Department, saw that this self-reliant approach had become a drag on change. The team regularly created new drawings without referring to the existing inventory, and part numbers ballooned as a result. That approach worked in the era of mass production, but with low-cost, high-performance equipment from countries like China entering the market and demand shifting toward smaller-lot, higher-variety production, Matsui recognized YKK couldn't maintain its competitiveness without breaking from that mindset.
This wave of part number proliferation and over-reliance on individual expertise created real tension on the manufacturing floor. Ogawa, General Manager of the Manufacturing Control Group, describes manufacturing's basic idea as building exactly what the drawing shows, but in practice, so much tacit knowledge lived on the floor that the team relied on veteran instinct to get the job done. Certain drawings could only be produced by certain people or certain machines, and a closed culture kept technical information from leaving the department, so the team's pace began slipping against external risks.
Nishida, General Manager of Business Operations, recalls that procurement work had become so dependent on individuals that switching the person in charge would cause problems. Miyazaki, General Manager of the Production Engineering Office, saw the same dynamic at the design stage: even when a similar shape already existed, people would draw up a new one and repeat similar quote requests. Across departments, a negative cycle had taken hold in which past work simply wasn't being reused.
An external shock sparks organizational change
The catalyst for change was the disruption of the COVID-19 pandemic. With travel restricted, Matsui took a hard look at operations and was struck by how analog things still were: drawings, for instance, could only be exchanged by mail.
A further push came after the pandemic eased, from an experience with an outside partner company.
YKK had handed a supplier a drawing that listed 15 process steps, only to learn the supplier was completing the same work in a single step. That realization made Matsui feel the urgent need to shift the mindset across the organization.
{{quote-1}}
Around the same time, an organizational restructuring tightened management oversight. Nishida recalls the pressure of that period: the old approach wasn't going to cut it anymore, and the team needed a new way to visualize and standardize individual know-how, one that could eliminate the over-reliance on individuals in procurement.
A speed that even veterans find impressive
Among the many products available, the department chose CADDi, the manufacturing AI data platform. The deciding factors were CADDi's startup-like speed, its technical fit for the problem, and the passion for manufacturing CADDi's staff brought to the table.
Adoption wasn't without resistance on the ground. Shima, of the Production Engineering Office, admits the previous vendor's system took more than 30 seconds to search, and there was some skepticism on the floor that this would just be more of the same. What dissolved that skepticism was the speed of CADDi Drawer, CADDi's cloud application for manufacturing data utilization, and the reduction in time spent on daily tasks.
{{quote-2}}
Even veteran staff have told Shima's team it's easy to use, and mid-career employees are discovering things too, like realizing another department already found a cheaper procurement option, simply by using the tool themselves.
Takamatsu, of the Manufacturing Reform Execution Team, drove adoption by getting directly involved on the shop floor and building trust with the team there. Being able to access information easily was essential to the goal of shortening manufacturing lead time, and Takamatsu was surprised the team could link a drawing to its past defect history within just a few days. CADDi Drawer has sped up information extraction and informed decisions on process consolidation, and now even the quality team uses it regularly.
These results aligned with a broader management vision, articulated by Miyazaki, of making CADDi Drawer the go-to search entry point to accelerate parts standardization, and momentum quickly built for taking CADDi adoption to an even higher level.
From tacit to explicit knowledge: CADDi as a bridge across generations
CADDi also proved to be a solution to a demographic imbalance the Production Engineering Office had been facing: an age gap split between employees in their sixties and those in their twenties, with almost no one in between. Shima sees real value in accumulating, as a digital asset, the know-how that had been at risk of disappearing as veterans retired or transferred elsewhere.
Shima's team brought all of its defect root-cause analysis and supplier guidance records into CADDi. New employees are now told to check with AI and CADDi Drawer first, and only ask veteran staff if they still can't find the answer, which has grown the range of work people can handle on their own.
{{quote-3}}
Miyazaki also praised the dramatic increase in the speed of process rationalization for parts, while pointing to what comes next: this is currently an effort within the Machinery Manufacturing Department, but true efficiency gains will require bringing the design department into the fold and driving change further upstream.
Ogawa sees this shift as a major step toward turning the tacit knowledge on the shop floor into explicit, quantified knowledge through AI. Shima also praises CADDi's customer success team, noting that it responds quickly, often coming back the same day with proposals related to information extraction whenever the team raises a request, energizing everyone involved.
CADDi's path to "Unrivaled Cost Competitiveness"
The project is now moving toward a fundamental transformation of the manufacturing process. Matsui wants to build an organization where people spend their time on more creative work rather than being buried in rework. Using AI to ensure the same quality fasteners can be delivered from any factory anywhere in the world is the vision the team is aiming for. Digitalization and AI adoption are essential if Japanese manufacturing is going to compete, and Matsui wants to expand this challenge beyond the department to the rest of the manufacturing industry.
{{quote-4}}
Nishida envisions sales planning, manufacturing, and procurement automatically connected, able to run with a minimal team.
YKK's long tradition of fully integrated production, combined with AI technology, stands as a signpost for how Japanese manufacturing can break free of its self-reliant mindset and gain new competitiveness through digital transformation.
Ready to see CADDi Drawer in action?
Get a personalized demo.
"We'd been spending 15 steps on something that could genuinely be done in one."

"CADDi Drawer is fast enough that engineers genuinely enjoy using it. Even veteran staff have told us it's easy to use, and mid-career employees are discovering things too, like realizing another department found a cheaper procurement option, simply by using the tool themselves."

"It used to be impossible to manually check tens of thousands of drawings, but now we can extract the data we need in just a day or two."

"We want to build a manufacturing platform with unrivaled cost competitiveness."


.png)




.png)