Key Pain Points
Steps 4-5 are confirmed bottlenecks. David stated: "Most time is spent reviewing and refining measurement data and accurately translating it into SolidWorks models." At ~2.5 designs/month, they cover 70% of the product market. The remaining 30% represents unrealized revenue, and OEMs keep changing specs - making it a perpetual race.
Phased Implementation
Phase 1 (4-6 wks): Drawing automation - Steps 6-7. Quickest win, lowest risk.
Phase 2 (6-10 wks): Parametric templates - Steps 4-5 for known filter families.
Phase 3 (8-14 wks): AI scan-to-CAD - Steps 2-5 with machine learning feature recognition.
| # | Step | Current Process | AI-Powered Process | Time Impact | Phase |
|---|---|---|---|---|---|
| 1 | Receive OEM Part | Physical sample arrives, cataloged manually | No change - physical handling required | - | N/A |
| 2 | Faro CMM Scan | Engineer manually operates arm, exports data | Scripted scan patterns for known families, guided setup for new | ~20% reduction | Phase 3 |
| 3 | Import to SolidWorks | Manual file import, manual data cleanup | Auto-import pipeline: noise removal, mesh cleanup, alignment | ~80% reduction | Phase 3 |
| 4 | Review & Refine | BOTTLENECK - Engineer manually reviews all measurements, identifies features, maps connections | AI identifies features automatically. Engineer validates and corrects flagged items only | ~50-70% reduction | Phase 2-3 |
| 5 | Create 3D Model | BOTTLENECK - Engineer builds model from scratch in SolidWorks | Parametric template populated from dimensions. Engineer validates and adjusts | ~50-70% reduction | Phase 2 |
| 6 | Sales Drawing | Engineer manually creates drawing views, adds dimensions and specs | Auto-generated from 3D model via SolidWorks API. Engineer reviews | ~90% reduction | Phase 1 |
| 7 | Mfg Drawings | Engineer manually creates mfg views, weld callouts, BOM | Template-driven from 3D model. Engineer adds specialty callouts | ~70-80% reduction | Phase 1 |
Bottom Line
If current throughput is ~2.5 designs/month with 1.75 engineers, and Steps 4-7 represent the majority of engineering time, a conservative 50% time reduction on those steps could push output to 4-5 designs/month - nearly doubling capacity without adding headcount. The exact impact depends on time data we'll gather in the deep dive.
What We Still Need to Validate
All estimates marked with [TBD] or ranges require actual time data from David's engineering team. The single most valuable data point is: how many hours does each of the 7 steps take for a typical design? Without this, ROI projections remain illustrative rather than precise.