Quality Control & Defect Tracking System That Caught Problems Before They Reached the Customer
The company reversed a dangerous quality trend, rebuilt trust with its most important customers, and passed its ISO 9001 audit with flying colors. By replacing paper-based inspection with a digital quality system powered by AI-driven root cause analysis, the team reduced their defect rate from 4.2% to well below industry benchmarks and gained the ability to catch systemic problems — like a thermal expansion issue in a CNC lathe spindle — before they propagated across production runs. What had been a reactive scramble of corrective action requests became a proactive, data-driven quality operation with real-time visibility from the shop floor to the quality manager's dashboard.
A metal fabrication shop producing precision parts for automotive and industrial clients was losing credibility with its customers due to a defect rate that had crept up to 4.2%. Customer returns were increasing, corrective action requests were piling up unanswered, and the company's ISO 9001 certification — a hard requirement for most of their customer base — was at risk during their upcoming surveillance audit. They needed a system that would catch defects at the source, track them to root cause, and prove to auditors that their quality system was under control.
Client: A metal fabrication shop
Employee Size: 40 employees
Industry: Manufacturing
Services: - Custom Software Development - AI Automation & Workflow Design - Quality Management System - Mobile Inspection Applications
The company had held ISO 9001 certification for six years and served a customer base that included three Tier 1 automotive suppliers and a dozen industrial OEMs. Quality had always been managed through paper-based inspection forms filled out at the end of each production run. Inspectors recorded measurements, pass/fail results, and visual inspection notes on printed checklists that were filed in binders and rarely referenced again.
The paper system had three critical failures. First, defects were only caught at final inspection — after the entire batch had been produced. A dimensional error on a CNC operation discovered at the end of a 500-piece run meant 500 scrap parts and a full re-run. There was no in-process inspection checkpoint to catch drift before it became a batch-wide problem.
Second, when defects were found, there was no structured system to track corrective actions. The quality manager maintained a spreadsheet of open corrective action requests (CARs), but follow-through was inconsistent. At any given time, 30-40% of open CARs were overdue, and several had been open for more than six months without resolution. Customers who submitted CARs received inconsistent updates, damaging trust.
Third, root cause analysis was practically nonexistent. When asked why a defect occurred, the shop floor answer was almost always "operator error" — a response that satisfied no one and prevented no recurrence. Without data on defect patterns, the company could not identify systemic issues in tooling, fixturing, material, or process setup.
The defect rate had climbed from 2.1% two years prior to 4.2%, and customer returns had increased proportionally. Two automotive customers had placed the company on "quality watch" — a precursor to being removed from the approved supplier list. The ISO registrar had noted three minor non-conformances in the previous audit cycle related to corrective action timeliness, and the quality manager was concerned that the upcoming surveillance audit could escalate to a major finding.
JSG designed and deployed a digital quality management system that replaced paper inspection forms with tablet-based workflows, introduced in-process inspection checkpoints, and created a closed-loop corrective action system with AI-assisted root cause analysis.
Key Components
Digital Inspection Workflows on Tablets Paper inspection forms were replaced with Power Apps running on ruggedized tablets at each inspection station. Inspectors followed guided workflows that ensured every required measurement and check was completed in sequence. Data was captured digitally at the point of inspection, eliminating transcription errors and making results immediately available for analysis.
In-Process Inspection Checkpoints The system introduced mandatory in-process inspection gates at critical operations — first piece approval, mid-run sampling, and final inspection. Defects caught at earlier gates triggered automatic hold notifications on downstream operations, preventing bad parts from continuing through the process and compounding scrap costs.
Defect Tracking & Non-Conformance Management Every defect was logged with structured data — part number, operation, defect type, severity, quantity affected, and inspector notes with photo attachments. Non-conformance reports (NCRs) were generated automatically and routed to the quality manager for disposition (scrap, rework, use-as-is, or return to supplier).
AI-Powered Defect Pattern Analysis Azure OpenAI analyzed defect data across part numbers, operations, machines, operators, shifts, and time periods to identify patterns invisible to manual review. The model surfaced that 38% of dimensional defects on turned parts occurred on a specific CNC lathe during the first hour of the Monday morning shift — a finding that led to the discovery of a thermal expansion issue in the spindle that needed a longer warm-up cycle.
Corrective Action Workflow Engine N8N orchestrated the full corrective action lifecycle — from CAR creation through root cause investigation, containment action, permanent corrective action, and effectiveness verification. Automated escalation ensured overdue CARs could not be ignored: 48-hour reminders went to the assigned owner, 7-day overdue alerts went to the quality manager, and 14-day overdue alerts went to the plant manager.
Root Cause Analysis Dashboards Dashboards provided real-time visibility into defect trends by type, operation, machine, and time period. The quality manager could identify emerging patterns before they became systemic, and present data-driven root cause analysis to customers instead of the reflexive "operator error" response.
## Quantifiable Impact
- Defect rate reduced from 4.2% to 0.8% within eight months
- Customer returns reduced by 91% (from an average of 11 per month to fewer than 1)
- Scrap costs reduced by 63% through in-process inspection catching defects earlier
- Overdue corrective actions reduced from 30-40% to under 5%
- Average CAR closure time reduced from 47 days to 11 days
- ISO 9001 surveillance audit passed with zero non-conformances for the first time in three years
- Both automotive customers removed the company from "quality watch" status
- Frontend: Next.js
- Backend: .NET API
- Platform: TMX (quality data layer, NCR management, role-based access)
- Cloud: Microsoft Azure
- Database: Azure SQL
- AI & Automation: Azure OpenAI (defect pattern analysis, root cause correlation, quality trend prediction)
- Mobile: Power Apps (shop floor inspection forms on ruggedized tablets)
- Workflow Orchestration: N8N (corrective action workflows, non-conformance escalation, in-process hold triggers)
- Communication: Twilio (SMS alerts for critical quality holds), SendGrid (customer CAR status updates)
Ready to Modernize Your Business?
Let's talk about where technology can move the needle first.