The modern factory floor is a mosaic of automation gear, each piece often speaking a different language when it comes to programmable logic controllers. Engineers spend countless hours rewriting ladder logic or structured text whenever a new robot, conveyor, or sensor is added, because Siemens, Mitsubishi, Rockwell, and LS Electric each maintain distinct programming environments and instruction sets. This fragmentation inflates engineering costs, extends project timelines, and creates a verification burden that grows with every additional device. As humanoid robots and advanced collaborative arms begin to populate production lines, the need for a common control layer becomes not just convenient but essential for safety, scalability, and speed to market.
Enter JLT, a South Korean smart‑factory automation specialist founded in 2019, which has secured roughly 6.7 billion KRW through the Ministry of SMEs and Startups’ Global TIPS investment‑finance linked program. Partnering with the Korea Institute of Industrial Technology, JLT is extending its existing PLC code‑generation platform into a full‑featured Factory Control Operating System dubbed SYNEX+. The initiative seeks to abstract away vendor‑specific PLC semantics, allowing a single control script to run across disparate hardware without manual re‑coding. This move aligns with broader industry trends toward software‑defined manufacturing, where hardware heterogeneity is masked by a unified middleware layer.
At the heart of SYNEX+ lies the PackML standard, an OMAC/ISA‑derived framework originally created to harmonize packaging machine states. PackML defines a common set of equipment modes—such as Start, Stop, Pause, Error, and Recovery—and a consistent data‑exchange model that simplifies integration. By mapping each vendor’s PLC state machine onto PackML’s canonical model, JLT can translate a control routine written for, say, a Siemens S7‑1500 into equivalent instructions for a Mitsubishi FX5U or a Rockwell CompactLogix. The approach promises to reduce the semantic gap that has historically forced engineers to relearn vendor‑specific nuances for every new device.
Beyond state mapping, SYNEX+ couples PackML normalization with an AI‑driven code generator. Users input high‑level process requirements—desired sequence, timing, and I/O configuration—through a natural‑language interface or a simple form. The underlying AI model, trained on a vast repository of PLC programs and process descriptions, synthesizes the appropriate ladder logic or structured text automatically. JLT describes this as “engineers speaking control logic instead of typing it,” aiming to eliminate repetitive coding and lower the skill barrier for technicians who may lack deep PLC expertise but understand the production workflow.
Before any generated code touches real hardware, it undergoes rigorous validation inside a digital twin. Utilizing NVIDIA Isaac Sim for physics‑based robotics simulation and Unity for factory‑floor visualization, SYNEX+ creates a virtual replica of the target cell. Engineers can observe robot trajectories, conveyor timing, sensor triggers, and safety interlocks in a risk‑free environment. Discrepancies between simulated behavior and expected outcomes trigger automatic feedback loops that refine the AI model, ensuring that the final code is not only syntactically correct but also dynamically sound.
The verification strategy unfolds in four progressive stages. First, an internal testbed exercises basic functionality and edge cases. Second, the digital twin simulation expands to complex multi‑device interactions and fault injection scenarios. Third, domestic manufacturing partners conduct technology proof‑of‑concept (PoC) runs on actual production lines, providing real‑world data on cycle times, yield, and safety incidents. Finally, a U.S.–based PoC—anchored in Georgia—tests the system under North American regulatory and operational conditions, including differing power standards and safety norms. This staged approach mitigates risk while generating comparable data sets for performance benchmarking.
North American commercialization is being accelerated through a partnership with Millennium New Horizons (MNH), a deep‑tech venture firm with stakes in companies like CoreWeave, Mistral AI, and Joby Aviation. MNH’s portfolio brings expertise in high‑performance computing, AI infrastructure, and autonomous systems—capabilities that can enhance SYNEX+’s scalability and reliability. While JLT plans to exchange technical insights with MNH‑backed firms, the exact scope of joint development remains undisclosed, leaving room for future collaborations in areas such as edge AI acceleration or cloud‑based monitoring dashboards.
To staff the impending U.S. rollout, JLT is collaborating with the University of Georgia to design a localized engineer training program. The curriculum will cover SYNEX+ architecture, PackML concepts, AI‑assisted code generation, and digital‑twin validation procedures. Graduates are slated to support PoC installations at strategic sites, including Hyundai Motor’s Georgia plant and LG Magna’s facilities. By embedding skilled personnel directly in customer environments, JLT hopes to accelerate adoption, gather immediate feedback, and build a services‑based revenue stream around system commissioning and ongoing optimization.
JLT has set ambitious performance targets for the SYNEX+ project: a reduction in development time exceeding 60%, an integrated control success rate above 90%, and a real‑time control response latency under 20 ms. Internal benchmarks from predecessor projects suggest a 58% cut in average development cycle and a 60% decline in PLC engineer man‑hours after SYNEX+ deployment. Although the specific project sample size and equipment mix used for these estimates are not public, the numbers signal a potentially transformative impact on automation project economics, especially for mid‑size manufacturers that lack large dedicated controls teams.
Safety and correctness remain paramount when AI‑generated PLC code directly drives physical actuators. An incorrect sequence or mistimed signal can cause equipment shutdowns, defective products, or even endanger workers. Consequently, SYNEX+ enforces mandatory manual approval steps, version control, and automated rollback procedures alongside its code‑generation speed. The system also logs every simulation‑to‑reality deviation, enabling continuous model improvement and providing auditable traces for compliance with standards such as ISO 13849 or IEC 62061.
The applicability of PackML beyond its packaging origins is another critical consideration. Extending the standard to welding cells, autonomous guided vehicles, humanoid robots, and intricate logistics conveyors requires defining additional states and safety conditions that reflect the unique hazards of those domains. JLT must work with standards bodies and industry groups to ensure that PackML extensions remain interoperable and do not introduce ambiguities that could undermine the very uniformity the system seeks to achieve.
If SYNEX+ succeeds, its ripple effects could reshape how factories approach equipment procurement and line reconfiguration. Manufacturers might gravitate toward best‑of‑breed hardware knowing that control logic portability will mitigate vendor lock‑in. System integrators could offer faster turn‑key solutions, charging less for engineering hours and more for performance optimization. Moreover, the ability to reuse control logic across sites accelerates the scaling of pilot innovations—such as a new humanoid‑assisted assembly cell—from a single line to global networks with minimal re‑engineering.
For stakeholders looking to capitalize on this shift, several actions are advisable. Automation engineers should familiarize themselves with PackML fundamentals and explore natural‑language‑to‑code tools to stay ahead of the skill curve. Plant managers ought to pilot a SYNEX+‑style middleware on a non‑critical line to measure actual reductions in engineering hours and downtime risk. Investors should track not only the funding milestones of projects like JLT’s but also the adoption rates of AI‑driven code generation and digital‑twin validation across the broader industrial IoT landscape. By aligning early with these emerging standards, firms can position themselves to reap the efficiency gains and resilience benefits that a truly vendor‑agnostic control layer promises.