TMCnet News
DeepHow and Yazaki North America Drive Physical AI to Boost Factory Production EfficiencyDeepHow today announced an agreement with Yazaki North America, one of the world's largest automotive parts manufacturers, to deploy DeepHow Time and Motion AI across its production lines. This engagement occurred through DeepHow's and NVIDIA's collaboration. While traditional time and motion studies historically relied on manual observations with stopwatches and clipboards, modern studies are shifting toward automated, digital, and AI-powered alternatives, which dramatically reduce the time to perform the studies and eliminate the variability inherent in manual observation allowing for a more accurate, objective assessment of process performance under natural conditions. Integrating with NVIDIA Metropolis Blueprint for video search and summarization (VSS) and NVIDIA RTX PRO Servers, DeepHow's agentic solution makes process execution visible in real-time as workflows naturally unfold. Powered by advanced vision-language models, the system automatically captures cycle times and classifies workflow segments across production cycles, identifying opportunities for value-add and waste reduction. Next, DeepHow is launching automatic report generation and Factory Flow Intelligence powered by NVIDIA Cosmos and built with VSS agent skills. The resulting intelligence supports Operational Analysis Improvement, operator and station comparisons, and standard-work analysis, turning manual time studies that once took engineers weeks into insights that can be generated in minutes. Beginning with operations in Mexico, Yazaki North America expects to reduce line-analysis time substantially, cutting its current process from weeks to days and unlocking millions of dollars in annual savings. When scaled globally, the same efficiency gains could represent tens of millions in annual value. For Yazaki North America, the impact extends beyond faster analysis. Instead of relying on prescheduled studies a couple of times per year, teams can analyze any line on demand, identify bottlenecks sooner, and rebalance work before inefficiencies become embedded in the operation. By using AI to surface lost capacity earlier, Yazaki North America expects to improve throughput with better line balancing, reduced waste, and faster continuous improvement cycles. The larger opportunity is the dataset this creates. Factories generate enormous amounts of machine data, but almost none about the nuanced dynamics of manual assembly workflows: the sequences, variations, and operational decision points that impact quality and throughput outcomes. By capturing that work in a structured format, Yazaki North America begins to build a continuous record of how its products are made. Combined with engineering expertise and vision-language models, that record can be analyzed to improve methods, increase capacity, and identify optimization opportunities across the network. "Operational Analysis Improvement is the starting point, not the destination," said Joseph McCorry, Head of Commercial & VP YNCA Business Units at Yazaki North America. "Combining video, vision-language models, DeepHow's AI platform, and the ingenuity of our manufacturing teams opens the possibility of designing new production environments, lines, and factory concepts that are safer and more productive than anything in operation today." About DeepHow DeepHow is a physical AI company for operational excellence, enabling manufacturers to capture knowledge, verify work, and optimize frontline execution at scale. Used by more than 100 customers across 1,500 locations in 28 countries, DeepHow serves manufacturers across industrial manufacturing, electronics, pharma, utilities, and food and beverage. Learn more at deephow.com.
View source version on businesswire.com: https://www.businesswire.com/news/home/20260716662554/en/ |
