Sumitomo Wiring Systems, a major manufacturer of automotive wire harnesses, has launched a full-scale initiative to build smart factories powered by physical AI technologies in collaboration with AI partners. Facing severe manufacturing challenges such as labor shortages and the loss of veteran expertise, the company is tackling these issues through next-generation facilities where multiple AI agents and robots work seamlessly alongside human operators.

The Roles of Multi-AI Agents Handling Auditing, Intervention, and Review

In August 2026, Sumitomo Wiring Systems Co., Ltd. and ExaWizards Inc. began joint development of a multi-AI agent system designed to operate on factory floors. Leveraging data collected from camera feeds and various shop-floor sensors, this initiative aims to construct a physical AI system capable of detecting anomalies while simultaneously transferring the tacit knowledge of experienced workers. Specifically, three specialized AI agents collaborate closely: first, the Audit AI analyzes real-time production line data to detect early signs of slowing output or equipment irregularities. Next, the Intervention AI uses those audit findings to immediately prompt line leaders and operators with targeted corrective actions. Finally, the Review AI interacts with veteran technicians and supervisors to learn and accumulate previously unarticulated tacit knowledge and advanced technical know-how. Please refer to the following system diagram.

By evolving step-by-step from a “see and notify” phase to a “learn and teach” phase, the system goes beyond conventional monitoring to create a virtuous cycle that automatically captures master skills and channels them back into workforce training and education.

Partnership with US and Japanese AI Firms and the Roadmap to Autonomous Factories

As part of its smart factory initiative, Sumitomo Wiring Systems partnered not only with ExaWizards for AI agent development, but also with Skild AI, a robotics startup originating from Carnegie Mellon University in the United States. By combining Skild AI’s advanced physical AI robotics with ExaWizards’ multi-AI agent technology, the company is advancing the establishment of production lines where AI-driven decision-making and autonomous robotic tasks operate as one. Sumitomo Wiring Systems has designated fiscal 2025 through fiscal 2026 as the foundation-building and field-testing phase, with plans to launch pilot testing on active production lines at overseas facilities during fiscal 2026. Furthermore, from fiscal 2027 through fiscal 2028, the company intends to construct model lines and drive full-scale global deployment, ultimately aiming to roll out autonomous factories worldwide by fiscal 2030 that seamlessly integrate fully automated lines with collaborative human-robot lines.

Structural Challenges and Strategic Context Facing the Auto Parts Giant

Shifting Away from Labor-Intensive Wiring Harness Production

Wire harnesses, often referred to as the circulatory and nervous systems of automobiles, feature intricate wiring configurations that vary by vehicle model. As a result, manufacturing has long relied on labor-intensive, manual assembly across global production sites. In recent years, however, the environment surrounding the manufacturing sector has shifted dramatically. Escalating geopolitical risks and rising worldwide labor and logistics costs have converged with severe managerial headwinds, including the loss of irreplaceable know-how as veteran technicians retire, chronic labor shortages, and skill disparities across different countries and facilities. To mitigate these multifaceted risks, implementing physical AI on the shop floor—where AI directly interacts with humans, machinery, and robots in real-world environments—has become essential, going well beyond standard back-office generative AI applications. Establishing an autonomous production framework that dramatically reduces human burden while ensuring consistent, high-quality manufacturing anywhere in the world is now an urgent imperative.

Mid-Term Management Plan 2028 and Integrated Cloud Strategy

This physical AI endeavor serves as a core initiative under the Sumitomo Wiring Systems Group 2030 Vision and Mid-Term Management Plan 2028 (28M). Under its corporate purpose, the company strives to deepen its wire harness technologies and remain an indispensable partner in the mobility society. To support smart factory transformation, Sumitomo Wiring Systems is simultaneously modernizing its IT infrastructure. In April 2023, the company initiated a comprehensive cloud migration project to move away from legacy on-premises data centers, adopting Azure IaaS and Azure VMware Solution (AVS). By establishing an advanced data foundation that integrates and analyzes vast volumes of shop-floor camera and sensor data in the cloud to instantly feed insights back into multi-AI agent decision-making, the company provides robust support for business continuity planning (BCP), cost reduction, and manufacturing digital transformation.

Competitive Landscape and the Future of Physical AI in Manufacturing

Precedents and Differentiators Among Domestic Japanese Manufacturers

The integration of AI into production sites and the codification of veteran technical knowledge are accelerating rapidly across Japan’s manufacturing sector. For instance, in July 2026, Hitachi, Ltd. announced the development of an optimal operation guidance system powered by physical AI that incorporates the decision criteria of master operators, with commercial launch targeted for chemical plants by the end of 2026. While Hitachi is prioritizing the commercial release of a specialized software package for plant-wide operations and predictive monitoring, Sumitomo Wiring Systems is taking an implementation-first approach directly geared toward its global manufacturing sites. By establishing a unique multi-agent architecture where distinct Audit, Intervention, and Review AI agents interact with physical robots, Sumitomo Wiring Systems achieves flexible, site-tailored autonomy tailored to the complex manual assembly characteristic of automotive component manufacturing.

Operational Challenges and Medium-to-Long-Term Outlook for Global Rollout

The pilot testing scheduled at overseas sites during fiscal 2026 represents a critical milestone in proving the practical viability of this initiative. Operational evaluations will focus on ensuring high detection accuracy for anomaly signs using equipment metrics and image analysis under real factory conditions, as well as testing how smoothly shop-floor operators and supervisors respond to intervention prompts from AI agents. Sumitomo Wiring Systems and ExaWizards aim to translate the insights gained from these trials into a standardized base for site-adaptable physical AI, expanding them across platforms such as exaBase Studio. For Sumitomo Wiring Systems—which operates 106 facilities across 32 countries and regions with approximately 250,000 employees worldwide—establishing a mechanism to instantly share master skills across borders will deliver a profound impact, driving local production for local consumption and reinforcing global supply chain resilience.

Reference Pages

• 【Sumitomo Wiring Systems News Release】 https://www.sws.co.jp/sws-news/2026/09171330.html
• 【ExaWizards Press Release】 https://prtimes.jp/main/html/rd/p/000000494.000030192.html
• 【Sumitomo Wiring Systems Group “2030 Vision & Mid-Term Management Plan 2028″】 https://www.sws.co.jp/corporation/strategy.html

【#SumitomoWiringSystems #ExaWizards #PhysicalAI #AIAgents #SmartFactory #ManufacturingDX #WireHarness】