The outside of a laptop takes the most wear and sets the first impression. Graded by eye, a scuff on the lid can pass at one station and fail at the next, and each wrong call sends a device to the wrong grade or back into rework. OptiX automates exterior cosmetic grading and evaluates every surface against your exact specification.
OptiX: Automated grading for laptop exteriors
OptiX grades every laptop exterior against your exact specification. AI evaluates the actual condition of each surface and makes the same call at every station, on every shift.
-
30%+
Expected efficiency gain
Across OptiX and OptiY systems in production
-
5
Systems in production
Complete OptiX and OptiY systems operational
Exterior scuffs and dents look different depending on who is inspecting and how the bench is lit. Over a week of production, that adds up to:
- Different grades for the same device, depending on the operator
- Decisions that drift from shift to shift and site to site
- Over- and under-grading that costs yield and value
OptiX evaluates the actual condition of every exterior surface and makes the same call every time.
- Paired with OptiY for the whole laptop OptiX grades the exterior while OptiY grades the internal surfaces, so every laptop is covered inside and out.
- AI trained on real returns and synthetic defects The Optiline grading model is trained on 20,000+ samples, extended with synthetic defect images generated in NVIDIA Omniverse.
- Grading logic you configure Standards adjust by customer, region, or program without changes to the core AI model.
- Every decision on record Each grade is recorded with retrievable images, and results feed the ERP for routing and parts decisions.
- Serial-level records Every cosmetic decision is stored with retrievable images.
- ERP integration Automated workflows and real-time tracking.
- Version-controlled standards Specifications mapped to customer requirements, with change management.
- Analytics Dashboards help QA teams validate warranty claims and tune specs.
OptiX video
How OptiX works
Load
The operator places the laptop into OptiX, positioned for inspection.
Inspect
Cameras capture every exterior surface of the laptop under controlled lighting.
Grade
AI checks the findings against your specification and assigns the grade.
Route
Graded devices move on to OptiY or the next process step.
Common questions
Why does OptiX need a controlled environment?
Consistent grading depends on calibrated lighting, precise camera alignment, and maintained hardware. A controlled facility removes the variables that cause grading drift in distributed setups.
Can grading specs differ by market or device age?
Yes. Standards can be adjusted by region, customer tier, or product maturity.
How are cosmetic operators guided?
ERP receives the grading results and flags which components need replacement, based on zone and defect-level analysis.
How long does it take to add a new laptop model?
About 20 working days, covering spec mapping, AI training, validation, and SOP updates.
The Optiline family
Every Optiline machine runs on the same grading platform, with AI trained on real returns and synthetic defects from NVIDIA Omniverse, grading logic you configure to your own standard, and a record of every decision. Each machine is built for one kind of device.
Other Tech Solutions
Pegasus: AI-Driven Optical Receiving
Optical receiving that identifies products, reads part numbers, captures condition, and creates an image-verified record for every unit.
Atheris: Automated Label Removal
AI-guided label detection and removal that replaces manual scraping with fast, consistent automation built to protect device cosmetics.
Hydra: Automated In-Line Cleaning
Multi-stage cleaning with continuous flow, delivering uniform results inside line-style recovery environments.
OptiM: Mobile Cosmetic Grading
Automated 360-degree inspection and AI grading for smartphones and tablets, with controlled lighting built for glass and metal and the same call on every device.
OptiY: Laptop Interior Grading
Automated cosmetic grading for the inside of the laptop, with display testing built into the same pass.
OptiZ: Complex Device Grading
Robotic cosmetic inspection for headphones and other complex devices, with every defect marked on a 3D model before rework starts.
Proteus: Enterprise Drive Erasure
Our gold standard for HDD and SSD testing and erasure, with ADISA-tested erase functionality, up to 64 test cells, and full audit trails.
Rackwipe: In-Rack Data Erasure
End-to-end, fully automated data erasure for data centers, clearing drives, ports, switches, and sensitive configurations with detailed digital reporting.
Phoenix: Bench-Scale Memory Diagnostics
A compact FPGA-based tester that runs up to ten DDR3, DDR4, and DDR5 modules at once for deep, module-level diagnostics.
Maya: Fast NAND Testing
High-speed testing for NAND flash memory in BGA and TSOP, with advanced ECC verification and full protocol support from SLC to 3D NAND.
Bujang: Production-Line DRAM Screening
Proprietary high-throughput DRAM screening for DDR3 and DDR4, built to catch defects early across full production volumes.
Nest: Mobile Device Initialization
Automated initialization, charging, and OS preparation that feeds Revolution a steady supply of ready-to-test units.
Revolution: Full-Functional Device Testing
Automated full-functional testing that picks up each device in the state Nest prepared, with no manual handling between stages.
Titan: Automated CPE Testing
A scalable test platform for modems, gateways, and set-top boxes, with up to 24 test slots and robotic loading.
ICT Tester: Circuit-Board Testing
In-circuit and functional testing for small control boards, with rapid changeover through swappable upper and lower probes.
The PSUT: Power Supply Testing
Customizable power supply testing using advanced electronic loads and analytics to prove reliability and compliance at scale.
RCU Tester: Remote Control Testing
Automated remote control testing across IR, Bluetooth, buttons, LEDs, and audio, replacing manual cycles with fast, repeatable runs.
Ready for consistent exterior grading?