TaiwanManufacturing2025 – ongoingGW Instek
AI agent quality collaboration system for the DIP manual-insertion line
GW Instek reported a 50% better DIP rework rate after adding an AI agent to its manual-insertion line
Problem
Quality checks on the manual-insertion (DIP) assembly line at the Tucheng plant happened after the fact. Defects led to rework and repair labor, and new operators needed time to learn the line.
Approach
GW Instek built an AI agent quality collaboration system with the Institute for Information Industry under a Ministry of Economic Affairs program. It uses small language models, retrieval-augmented generation and edge computing. Modules include a smart workstation, a smart inspection station, a real-time monitoring center and a smart line advisor. The company positions the AI as an assistant to operators.
Results
Takeaway
Moving quality checks from after the build to the workstation gave the company a measurable drop in rework. The figure is the company's own and the source does not give a baseline.
Categories
- Industry
- Manufacturing
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Generative AIAI agents
- Adoption
- Partner co-development
- Evidence
- Company's own claim
Sources
Source summaries
Award coverage describing the DIP line AI agent system GW Instek built with the Institute for Information Industry. It lists the system's modules and says the DIP defect rework rate improved by 50%. The company says it also cuts scrap cost and shortens operator learning time, without figures.
Full details are in the original sources.