A No-Retreat Transformation: Wei Sheng Drying Pushes AI Competitiveness with KPIs

(Central News Agency, Taipei, July 4) Hidden on an industrial road in Baiji Mountain, New Taipei City, the roar of machinery echoes from a factory. Inside, giant ovens, these "semiconductor drying equipment" weighing 2 tons each, are being prepared for shipment to leading manufacturers, becoming essential components in precision manufacturing processes.

Wei Sheng Drying Industry, which has specialized in "how to stably dry chips" for half a century, serves clients in the semiconductor, printed circuit board, and optical industries. Leveraging professional technologies such as precise temperature control, enhanced cleanliness, and heat recovery, the company assists Taiwan's "sacred mountain" (a term for the semiconductor industry) in its advanced back-end packaging processes, ensuring smooth shipments.

Although Wei Sheng's position in the semiconductor supply chain suggests a steady stream of orders, the company is not content with the status quo. This established enterprise is undergoing an unseen revolution – an AI upgrade and transformation.

The instigator of this revolution is Chen Wen-han, the second-generation leader and general manager of Wei Sheng Drying. He states that the rise of AI has made him realize that "technology that could sustain a company for 30 years might now last less than 3 years." The pace of industrial competition has accelerated significantly. However, conversely, AI also presents opportunities for small and medium-sized enterprises to achieve significant results with limited resources, provided they seize this moment of transformation.

Rejecting a Gentle Transition: AI Integrated into KPI Assessment Mechanism to Motivate Employee Learning

"The hardest part of transformation is always the people," Chen Wen-han stated directly in an interview with CNA's team. Initially, employees generally held a resistant attitude towards new technologies and were unwilling to change their existing work habits. In the past, simply guiding employees from 2D to 3D drawing encountered considerable backlash. Senior employees were reluctant to abandon their long-used existing libraries and thus resisted learning 3D drawing.

He recalled that employees at the time felt that with already tight delivery deadlines, learning new things was too troublesome and they were unwilling to learn. This mindset is often the primary challenge faced by many traditional industries when initiating digital transformation.

When Chen Wen-han sought to fully implement AI tools, the resistance from employees was predictable. Beyond mere inconvenience, employees harbored a deeper fear: "Will AI replace me?" Wang Ting-liang, a project manager at Wei Sheng Drying, also observed that during the promotion of Knowledge Management, employees were extremely resistant to converting their in-house experience into structured documents, either unwilling to share or resisting the operation of new systems.

Chen Wen-han realized that if only gentle persuasion was used, AI adoption would remain limited to a few employees. Therefore, he decided to push strongly from the top down, directly incorporating the use of AI tools and the generation of knowledge documents into KPIs. This approach combines performance bonuses with guidance, meaning that the higher an employee's work efficiency improves through AI tools, the more it will be reflected in their mid-year (or year-end) bonuses.

To disarm employees, Chen Wen-han personally led by example. He demonstrated how AI could precisely categorize VIP information, a task that previously took his secretary 8 hours, in just 5 minutes.

The stark difference in work efficiency for the same task naturally created pressure on colleagues unwilling to use AI. "We were half-forced by the boss to use it," Wang Ting-liang said with a laugh.

Previously 3 Days to Verify International Standards to Avoid Infringement; Now Only 3 Minutes

Today, Wei Sheng's transformation revolution is reflected in concrete results. Over 90% of Wei Sheng's orders are customized. In the past, R&D required significant time to compare international patents and regulations to avoid infringement risks. Now, through AI semantic analysis, manual comparisons that previously took 3 days can be completed in 3 minutes.

After introducing multimodal reasoning models, employees can now produce 3D construction drawings through deep collaboration with AI by simply sketching or using voice input, shortening R&D timelines by 40% to 50%. In thermal fluid analysis, the processing efficiency of AI Agents has significantly improved compared to manual interpretation. Furthermore, by using NotebookLM for knowledge mining, decades of maintenance records have been converted into structured data, significantly shortening new employee training time by approximately 70% and effectively addressing the experience gap caused by technician retirements.

For small and medium-sized traditional industries, capital investment is often a major challenge. However, Chen Wen-han adopts a different investment mindset. He does not view transformation expenses as mere costs but as investments in long-term competitive assets. He emphasizes that, similar to how the company set up an in-house gym during the pandemic to invest in employee health, investing in AI is for the long-term survival and development of the enterprise, ultimately contributing to overall growth.

Wei Sheng Sets Goal of "AI-Native Enterprise" This Year to Maintain Advantage Amidst Change

Wei Sheng has set a goal to become an "AI-Native Enterprise" this year. For Chen Wen-han, adopting AI is not just a technological upgrade but a competition for corporate survival. He points out that the pace of industrial competition has greatly accelerated, and AI provides small and medium-sized enterprises with opportunities to achieve significant results with limited resources.

Therefore, Wei Sheng advocates a "try first, then fix if wrong" trial-and-error culture to cope with the rapidly changing technological environment. Chen Wen-han emphasizes that in the AI era, traditional decision-making models relying on manual labor and experience are no longer sufficient. Only by continuously accumulating data and driving decisions with data can enterprises maintain their competitive advantage amidst rapid change. (Editor: Pan Yi-ching) 1150704

Stand with the facts. Your sponsorship is a force for safeguarding press freedom.

Download the CNA "One-Stop News" APP for real-time updates.

Text, images, and videos on this website may not be reproduced, publicly broadcast, publicly transmitted, or used without authorization.