SkillReal

**Product and the concrete problem it solves.** SkillReal addresses a stubborn manufacturing failure mode: the digital design, the production line, and the finished part often drift apart, while conventional metrology and machine-vision systems inspect only a sample of features or require costly, fixed hardware. The result is that a missing stud, misplaced weld, wrong hole, or dimensional deviation may be discovered downstream, after labor, material, and additional assemblies have already been committed. SkillReal's Digital Twin Inspection and Digital Twin Alignment products are intended to move that check into the production cycle itself. The company describes a software platform that verifies presence, location, tolerance, and quality of manufacturing features against the CAD model, with factory dashboards and reports for immediate corrective action. Its current site emphasizes 100% of parts and features within cycle time, while the older Digital Twin Alignment product also supports virtual prototyping, manufacturing-cell layout validation, robot calibration, assembly verification, and as-built-versus-as-designed comparison. The commercial promise is therefore broader than defect detection: less scrap and rework, faster line commissioning, fewer manual checks, better engineering traceability, and earlier discovery of errors that are expensive or dangerous to correct later.

**Core technology and how it actually works.** The architecture combines ordinary production-floor equipment with proprietary calibration, geometric alignment, and AI. SkillReal says its camera-calibration algorithms correct lens and sensor distortion in real time, including wide-angle optics, so standard cameras can act as accurate measurement instruments rather than requiring a purpose-built metrology cell. Its CAD-based alignment stage uses AI detections and 3D calculations to register a physical part against its digital model and evaluate dimensions according to geometric dimensioning and tolerancing. The inspection layer then fuses image-processing data with accelerated, pre-trained large AI models. The company states that these models were trained on more than 35,000 real-world manufacturing variations and millions of augmentations, and that they can identify features without collecting hundreds of good and bad examples for every new part. The intended output is not a black-box pass/fail score alone: an open platform connects to PLC, HMI, MES, and PLM systems, exposes role-specific analytics, and preserves a digital record of what was measured. The product page reports up to 0.05 mm 3D precision and above 99.7% reliability for required features, while the Innovation Authority profile reports up to 0.07 mm precision; these are company and government-profiled claims, not independently audited benchmarks.

**Market, customers, and go-to-market.** SkillReal began with Automotive Body-in-White, where high-volume lines contain many weld points, studs, bolts, holes, and assemblies, and where a late quality escape can trigger expensive rework or a delayed vehicle launch. It has expanded its stated target industries to aerospace and defense, industrial machinery, heavy industry, batteries, and line builders. The go-to-market is a pragmatic B2B retrofit: use existing cameras, PCs, robots, and factory systems; avoid new floor space and dedicated jigs; integrate with the customer's CAD, PLM, MES, PLC, and HMI environment; and sell a measurable reduction in inspection labor, downtime, rework, and launch risk. The company's public materials say its solution is used by leading OEMs and Tier 1 suppliers, and the Start-Up Nation Central I4.0 guide lists BAE Systems, IAI, Rafael, and GE-Baker Hughes among high-level aerospace and defense customers. A SkillReal LinkedIn announcement also identifies Siemens/VW and Honda Motors among customers that purchased its software, but public sources do not disclose contract sizes, deployment counts, recurring revenue, or the exact scope of each customer relationship. The Siemens channel is strategically important because it places SkillReal inside an established industrial software and digital-twin workflow rather than requiring every sale to begin with a greenfield inspection project.

**Traction, funding, and third-party validation.** There are several meaningful validation signals, although the public evidence is not a substitute for customer diligence. The Israel Innovation Authority lists Skill Real Ltd. as established in 2019, with 30 employees, initial revenues, supplementary funding, and additional R&D-fund participation in 2024, 2025, and 2026. The BIRD Foundation's 2021 annual report lists SkillReal Ltd. in Ramat Gan together with Siemens DI in Plano, Texas, for the SkillReal Digital Twin Alignment project. SkillReal's own news archive identifies that project as a $1 million BIRD investment, and Siemens describes SkillReal as a portfolio partner since 2023 whose partnership helped secure BIRD funding. The company reports an OEM agreement with Siemens and successful installations by leading customers worldwide. In 2024, SkillReal publicly announced that it won the Global Automotive Mobility Innovation Challenge, a competition sponsored by the Society of Automotive Engineers, after competing among more than 1,500 startups; that claim appears in the company's LinkedIn announcement and is corroborated by a contemporaneous GAMIC winners post. Start-Up Nation Central selected SkillReal for its I4.0 solutions guide, including a dedicated aerospace-and-defense use case. Important gaps remain: financing amounts beyond the BIRD project, investor ownership, revenue, retention, gross margins, independent accuracy testing, granted patent numbers, and customer references are not publicly disclosed.

**Founders and team background.** SkillReal identifies Shai Newman as founder and CEO. Its public leadership pages list Sagi Rorlich as CTO, Aviel First as CRO, Pini Brosh as chief delivery officer, Pete Grabowski as USA COO, and a board that includes Eric Bernheim, a senior partner emeritus at McKinsey, Koby Lif, and Rick Wagoner, the former CEO of General Motors. The older company page also names research, algorithms, product, sales, and R&D contributors, including Yuvalal Liron, Sagiv Rosenstein, Yaron Isovich, Meir Lebel, and Yehuda Sapir. This is a useful team shape for an industrial deployment company: algorithm and computer-vision capability must be paired with factory integration, delivery, product management, and access to automotive and aerospace buyers. The public record is stronger on named roles than on individual biographies. It does not provide a complete account of the founders' prior companies, military service, academic credentials, or patent inventorship, so those facts should not be inferred from the company's Israeli origin or defense customer list. The Innovation Authority records Shai Newman as CEO and founder and lists the company at a Ramat Gan address; the official contact page currently gives a Ramat Gan office and a North American office in Livonia, Michigan. Headcount is reported as 30 by the Authority, while a 2024 media article described approximately 15 Israel-based and four US-based employees, suggesting growth or differing counting dates.

**Competitive dynamics.** SkillReal competes in a layered market rather than against one identical product. Traditional metrology vendors such as Hexagon Manufacturing Intelligence and Creaform offer accurate 3D measurement through scanners, coordinate-measuring machines, and dedicated inspection workflows, but can require capital equipment, specialized programming, and production interruption. Industrial-vision incumbents such as ISRA VISION and Perceptron compete with camera-based inspection cells that are proven in automotive plants, yet often depend on fixed regions of interest, dedicated hardware, and re-teaching when models change. Siemens, Dassault Systèmes, and PTC compete at the digital-twin, PLM, and manufacturing-software layer; in SkillReal's case Siemens is simultaneously a channel and ecosystem partner, which creates leverage but also platform dependence. Emerging AI-vision providers can offer defect classification, but a model that recognizes a visual anomaly is not automatically a CAD-aware dimensional verification system. SkillReal's strongest claimed differentiators are (1) CAD as the inspection specification rather than a manually curated image library, (2) pre-trained models intended to reduce part-specific data collection, (3) standard cameras and zero-footprint retrofit, (4) inline operation at production cycle time, and (5) bi-directional integration with industrial engineering systems. The central diligence question is whether those advantages survive the messy calibration, lighting, vibration, model-change, and certification conditions of real factories and remain defensible as larger automation vendors add AI.

**Defense, security, and resilience dual-use relevance.** SkillReal has a direct dual-use case because the same core inspection technology serves commercial manufacturing and aerospace-and-defense assembly. The Start-Up Nation Central aerospace-and-defense profile describes real-time work instructions, assembly validation, reports, remote assistance, plug-and-play PLM/MES/ERP integrations, and automatic validation of each assembly step; it specifically names BAE Systems, Israel Aerospace Industries, Rafael, and GE-Baker Hughes among high-level customers. For defense production, quality escapes are not merely a margin problem: an incorrect fastener, weld, hole position, cable route, or structural assembly can compromise flight safety, mission availability, maintenance burden, or the reliability of a low-volume platform. A camera-and-CAD layer that provides traceability at the workstation can help address surge manufacturing, distributed repair, depot modernization, and the challenge of maintaining consistent quality when experienced inspectors are scarce. The resilience angle is similarly concrete: faster line commissioning and less dependence on bespoke integrators can shorten the time required to change a production line for vehicles, batteries, heavy equipment, or critical components. Calibration is still necessary. Public evidence establishes defense-sector targeting and named high-level customers, but not classified programs, military contracts, security certifications, export-control posture, or field performance under austere conditions. SkillReal should therefore be treated as a defense-manufacturing and industrial-resilience enabler, not as an operational battlefield system.

**Growth stage, trajectory, and key diligence risks.** SkillReal is classified as mid-stage: it has a released product, initial revenues in the Innovation Authority record, an OEM agreement with Siemens, reported installations, a BIRD-backed US-Israel development project, and public aerospace-and-defense references, but it remains a private company with limited disclosed financial and operating data. The trajectory is attractive if the company can turn one successful integration into a repeatable template across factories and industries. CAD-driven inspection could become a control point in the feedback loop from engineering change to shop-floor verification, especially as battery, aerospace, and defense manufacturers need more traceability without proportionally expanding inspection labor. Key diligence points are (1) verify the named customer relationships and determine whether they are pilots, paid production deployments, or channel references; (2) reconcile the company's reported 0.05-0.07 mm precision and 99.7% reliability with independent tests across surfaces, lighting, vibration, and line speeds; (3) understand patent scope, freedom to operate, and whether Siemens owns or licenses any jointly developed IP; (4) measure software gross margin, implementation hours, deployment repeatability, and the rate of model-change updates; (5) assess dependence on Siemens for distribution and access to Teamcenter or Process Simulate environments; (6) review cybersecurity, on-premises deployment, data retention, and defense export controls; and (7) test whether established metrology vendors can copy the software layer or bundle it into larger contracts. The company has a credible industrial base and strong strategic relevance, but customer concentration, proof of scale, and independently verified performance remain the decisive unknowns.