CASE 01·Precision machinery · Hinge bearings
Shin Zu Shing × Intellicon Solutions
AI transformation at a global bearing maker
from knowledge Q&A
all the way into the core of the process
Before the rollout
Open a MacBook, flip open a pair of AirPods, and the hinge that still closes precisely after hundreds of thousands of cycles comes mainly from Shin Zu Shing Co., Ltd., a company with 60 years of history. Shin Zu Shing is the world’s largest supplier of laptop hinge bearings, holds close to 3,000 patents worldwide, and vertically integrates precision processes including metal injection molding (MIM), stamping, turning and milling. It has led the precision machinery industry for years.
Deep as that technical base runs, the company faces the same problem as every other manufacturer. How do you preserve and pass on decades of process know-how? How do you search the sprawling data across R&D, manufacturing, quality assurance and management fast enough to use it? Shin Zu Shing’s answer was to bring in the enterprise AI Agent platform from Intellicon Solutions.
Goals
- Turn decades of process experience into a knowledge base AI can read, and avoid losing that technical expertise
- Make data from R&D, manufacturing, quality assurance and management searchable and cross-referenceable on the spot
- Judge Agents by whether they work in the field rather than by how many there are, and deploy them against the pain points that come up most often
- Make the Agent every engineer’s digital right hand, freeing people for the judgment and innovation worth more of their time
- Lay an intelligent R&D and manufacturing foundation for new fields: liquid-cooling parts for AI servers, aerospace, and precision automotive components
What Intellicon Solutions did
In three months Intellicon Solutions mapped AI use cases plant-wide and across departments, planning dozens of Agents that cover the R&D center, quality assurance, several manufacturing business groups and corporate administration. Each one answers a real pain point that comes up often:
- 01Systems-level mapping: work through every department’s pain points, data sources and decision logic at plant scale, then design Agents to match
- 02Into the core of the process: not stopping at knowledge Q&A or document tidying, but reaching specialist processes such as torque calculation and sintering quality analysis
- 03Experience as an asset: know-how scattered across engineers’ notebooks, Excel files and veteran technicians’ heads, turned into a digital asset anyone can query on the spot
- 04Human-AI collaboration: the Agent works as the engineer’s digital right hand, taking on repetitive work that needs heavy cross-checking
Use cases: AI across the whole precision manufacturing value chain
- R&D center — hinge torque calculation
- engineers can assess the torque behavior of a hinge mechanism and match components while still designing, without leafing through past test data and formulas one by one. Product development cycles get much shorter
- R&D center — chart analysis
- takes over work that used to depend on the statistical package JMP, supporting more than ten kinds of technical chart, among them decay-rate trends, torque-angle curves, box plots and process capability analysis. Engineers supply the test data and get the visualization back immediately
- Shop floor — work-in-progress tracking
- query production progress and operation status, so floor supervisors no longer switch between several systems to compare
- Shop floor — sintering quality analysis
- for the critical metal injection molding step, helps quality control staff read the relationship between sintering parameters and quality data quickly
- Shop floor — cutting-tool requisition analysis
- matches machining characteristics against past purchase records and recommends the right tool specification and quantity, moving purchasing from judgment by experience to decisions driven by data
- Quality assurance division — drawing specification extraction
- reads 2D engineering drawings and identifies dimensions, tolerances and geometric features. Compiling the FAI (First Article Inspection) dimension list went from hours to minutes
- Corporate administration — smart help desk, customs document reading, supplier search
- instant queries on internal rules and procedures, automatic extraction of the key fields from import customs documents, and fast comparison and filtering of supplier data
Results
- Annualized ROI of close to NT$850,000 from one Agent
- the drawing analysis Agent already in service handles 15 users a week and cuts each task from 4 hours per person to 20 minutes, about 55 hours saved every week
- Across departments in three months
- dozens of AI Agents planned for individual departments inside three months, covering the whole value chain from R&D and manufacturing to quality assurance and management
- AI in the core of the process
- torque calculation, Shaft-Clip (shaft and clip) matching analysis and sintering quality analysis are all in production. Among AI applications in Taiwanese manufacturing, this is front-rank work
- Passing on experience
- decades of process know-how became a knowledge base AI can query and cross-reference instantly, so every engineer stands on the shoulders of those who came before
What the management team says
“For 60 years Shin Zu Shing has run on an obsession with precision. We make hinges, we do MIM, and the tolerance on every part is measured in microns. Bringing in AI follows the same logic. We are not after flashy features. We want AI that genuinely understands the language of our processes and can give engineers analysis and advice they can trust at every step of design, production and quality control. We chose EgentHub because it lets us turn decades of process experience into a knowledge base AI can read, so that every engineer stands on the shoulders of those who came before.”
“With AI in manufacturing, what matters is not how advanced a model you use but how solid your data is and how thoroughly you have worked through your processes. What we value most is whether it works in practice, not how many we have. The goal is for the AI Agent to become every engineer’s ‘digital right hand’ — not replacing people, but handing the AI the work that is repetitive and takes a great deal of cross-checking, so that our people spend their time on judgment and innovation that is worth more.”
What the partner says
“We have worked with hundreds of companies, and what struck us most about Shin Zu Shing was how ready they were for AI. Plenty of companies want to use AI but cannot describe their own processes clearly. Shin Zu Shing is different: every department, every step of the work, the pain points and the data sources are all mapped out. Rarer still, their AI applications do not stop at knowledge Q&A or document tidying. They go into the core of the process, into torque calculation, Shaft-Clip matching and sintering quality analysis — this is fairly advanced practice among AI applications in Taiwanese manufacturing, and it proves that traditional manufacturers can not only use AI but use it at the deepest and most precise level.”
What’s next
Shin Zu Shing will next evaluate a hybrid cloud and on-premises deployment, keeping core process data and technical material inside the company. For precision manufacturing, protecting intellectual property leaves no other option. Once the data is protected, MCP (Model Context Protocol) will connect the Agents to existing systems, making them digital working partners that operate seamlessly alongside the company’s core systems.
From precision springs in 1965 to the largest laptop hinge supplier in the world, and on to a deep role in the core supply chains of the world’s leading technology brands, Shin Zu Shing has spent 60 years breaking through industry ceilings with technical innovation. It now treats AI Agents as the basis of its competitiveness for the next 60 years.
More customer impact stories
Want every employee to have their own AI teammate?
Places are limited. Our team will be in touch after you register.