AI Workflow · Predictive Maintenance
Hengtuo Precision Machinery · CNC Spindle Health Monitoring
Pain pointCNC spindles run under high load; veterans judged them by sound, and sudden failures idled delivery for 1-2 days.
SolutionIngest spindle temperature, current and alarm codes; build a health-score dashboard tracking anomalies and maintenance priority.
Unplanned downtime down 38%; shifted from post-failure firefighting to proactive parts & shift planning
"The dashboard flags high-risk machines early — we schedule maintenance with much more confidence." — Production Manager Zhou
Duration6 weeks (1 req + 3 dev + 2 pilot)
Team1 PM + 2 industrial engineers + 1 AI engineer
Key challengeDigitizing veteran experience: turning sound into quantifiable vibration + temperature data
StackPython + InfluxDB + Grafana
Custom Build · AI Visual QA
Weicheng Electronics · Connector Pin Defect Inspection
Pain pointManual inspection standards were inconsistent across shifts; miss rate ~4.8%, complaints centered on minor defects.
SolutionDeployed a single-station AI visual QA POC with industrial camera + stable lighting, linked to scanner for lot traceability.
Defect accuracy 98%+; miss rate cut from 4.8% to under 0.3%
"AI fixed the standard — shifts no longer argue over what counts as a defect, and we have records for customer audits." — QA Manager Zhao
Duration8 weeks (3 data + 3 model + 2 integration)
Sample data5000+ labeled images (12 defect classes)
Key challengeSmall-object detection: 0.3mm pins, traditional vision <80% accuracy
StackYOLOv8 + industrial camera SDK + MES integration
AI Partnership · Smart Scheduling
Desen Equipment · High-Mix Low-Volume Scheduling
Pain pointHigh-mix low-volume orders scheduled on Excel + gut feel; frequent insert orders forced constant rework and inaccurate due dates.
SolutionMapped orders, BOM, routings and capacity; built a lightweight scheduling tool with Gantt views and insert-order simulation.
Scheduling cut from half a day to 30 min; insert-order impact assessment in 5 min
"Insert orders used to be guesswork — now we see which orders are affected, far less internal friction." — Planning Lead Liu
Duration10 weeks (4 data governance + 4 build + 2 training)
Order volume50+ orders/day, 200+ product models
Key challengeMulti-constraint solving: due date + capacity + material readiness + routing
Stack.NET + PostgreSQL + Google OR-Tools
AI Workflow · Knowledge Assistant
Woneng Equipment O&M · Industrial LLM Knowledge Base
Pain pointManuals, SOPs and repair records scattered across drives and chats; field engineers had to phone experts on every alarm.
SolutionCleaned manuals, SOPs and work orders; built an O&M assistant answering natural-language queries by model / alarm code / symptom.
60%+ self-service rate for common issues; new-hire training cut from 2 weeks to 1
"Ask 'how to handle alarm E23 on model X' on-site — the answer and past cases come right up." — O&M Director Ma
Duration6 weeks (3 data cleaning + 2 RAG + 1 integration)
Knowledge base1200+ docs (800 manuals + 300 SOPs + 100 work orders)
Key challengePDF table parsing + industrial terminology normalization + multi-model classification
StackDify + DeepSeek + PGVector + Feishu bot
Custom Build · Energy Optimization
Dingxin Die Casting · Multi-Device Energy Management
Pain pointOnly saw total power bill at month-end — no idea which line or shift cost most; air compressors idling wasted heavily.
SolutionConnected main meter, die-cast machines, furnaces and air stations; built 3-tier energy dashboard flagging anomalies by shift/device.
Monthly power usage down 7.5%; shifted from lump-sum to per-device, per-shift management
"We used to only know the bill was high — now we know where. Granular data gives us a lever." — GM Assistant Xu
Duration7 weeks (3 data + 3 dashboard + 1 tuning)
Metering points30+ devices (8 die-cast + 4 furnaces + 3 air + 15 other)
Key challengeMulti-vendor meter protocols + real-time latency + anomaly detection
StackModbus + InfluxDB + Grafana + Python anomaly detection
Custom Build · AI Visual QA
Ruijing Auto Parts · BIW Weld Seam Visual Inspection
Pain pointSpot welds on body-in-white were sampled manually; cold/missed welds caused downstream rework costing multiples of inspection.
SolutionIndustrial camera + edge box at the weld station; real-time defect detection with position tagging, diverting via PLC.
Weld miss rate cut from 3.2% to 0.5%; QA throughput up 4x
"Sampling used to be luck — now every weld goes through AI. Downstream rework orders visibly dropped." — QA Lead Chen
Duration9 weeks (3 data + 3 model + 3 line integration)
Sample data8000+ labeled weld images (8 classes: cold/missed/burn-through etc.)
Key challengeHigh-glare imaging + inference within 30s cycle time + PLC integration
StackYOLOv8 + Jetson edge box + industrial camera + PLC
AI Workflow · EHS Safety
Anda Chemical · Unsafe Behavior Recognition
Pain pointSite patrols were manual; no helmets, smoking, zone intrusion hard to watch full-time — incidents were only found afterward.
SolutionReused existing CCTV for edge-based real-time recognition; auto screenshots + Feishu/WeChat alerts to safety officers.
Shifted from post-incident review to second-level alerts; monthly violations down 60%
"We had cameras no one watched — now AI watches and safety officers only handle alerts." — EHS Lead Wu
Duration5 weeks (1 camera + 2 model + 2 rules)
Coverage12 cameras (8 process + 4 perimeter), 24/7 recognition
Key challengeLow-light night + occlusion + reusing existing cameras to cut cost
StackYOLOv8 + edge inference box + alert platform + Feishu/WeCom
AI Partnership · Smart Scheduling
Xianyuan Food · Flexible Line Changeover Optimization
Pain pointHigh-mix low-volume SKUs with long cleaning changeovers; gut-feel scheduling caused needless changeovers; peak-season capacity strained.
SolutionMapped changeover matrix (flavor / packaging / allergen constraints); built an optimizer that merges similar SKUs to cut cleaning.
Daily changeovers down 35%; capacity utilization up 12%
"Scheduling is no longer a veteran's secret art — new hires produce solid plans, and output is actually higher." — Production Director Lin
Duration8 weeks (3 process + 3 build + 2 training)
SKU scale60+ SKUs across 5 lines and 3 categories
Key challengeChangeover-time matrix + allergen/cleaning constraints + shelf-life back-planning
Stack.NET + PostgreSQL + Google OR-Tools