DeepHow Partners with Yazaki North America to Bring AI-Powered Time and Motion Analysis to Automotive Manufacturing
DeepHow, a leading provider of Physical AI solutions for industrial operations, has announced a strategic agreement with Yazaki North America, one of the world’s largest automotive parts manufacturers, to deploy its DeepHow Time and Motion AI platform across Yazaki’s production lines. The collaboration marks a significant step toward transforming traditional manufacturing process analysis through artificial intelligence, computer vision, and advanced automation technologies.
The implementation was made possible through the ongoing collaboration between DeepHow and NVIDIA, leveraging NVIDIA’s AI infrastructure and advanced computing technologies to deliver next-generation manufacturing intelligence. By integrating AI-powered video analytics, vision-language models, and real-time production monitoring, the initiative aims to modernize how manufacturing operations are analyzed, optimized, and continuously improved.
The partnership reflects the growing adoption of Physical AI across industrial environments, where manufacturers are increasingly turning to intelligent automation to improve operational efficiency, reduce waste, increase throughput, and support data-driven decision-making. Rather than relying on periodic manual observations, manufacturers can now continuously analyze production performance and identify opportunities for improvement in near real time.
Reinventing Traditional Time and Motion Studies
For decades, time and motion studies have served as one of the most important tools for improving manufacturing productivity. Industrial engineers traditionally conducted these studies by observing operators on production lines, recording activities with stopwatches, clipboards, and handwritten notes before manually calculating cycle times and identifying inefficiencies.
While this approach helped manufacturers optimize operations, it was also time-consuming, labor-intensive, and highly dependent on the observations of individual engineers. Different analysts could produce different results, making it difficult to maintain consistency and objectivity across facilities.
Modern manufacturing environments have become significantly more complex, with greater product variation, faster production cycles, and increasing pressure to improve efficiency. These changes have created demand for digital alternatives capable of capturing operational data more accurately and continuously.
DeepHow’s AI-powered solution addresses these challenges by replacing manual observation with automated video analysis powered by artificial intelligence.
AI Brings Real-Time Visibility to Manufacturing Operations
At the heart of the deployment is DeepHow’s Time and Motion AI platform, which combines computer vision, video analytics, and advanced vision-language models to automatically observe production workflows as they naturally occur.
Instead of requiring engineers to manually monitor production, the system continuously analyzes manufacturing activities in real time.
The AI platform automatically recognizes production tasks, measures cycle times, tracks workflow progression, and classifies individual work segments throughout each production cycle.
This real-time visibility allows manufacturers to better understand how work is actually performed while eliminating much of the variability associated with traditional manual observations.
By continuously monitoring production processes, manufacturers gain a more objective and comprehensive understanding of operational performance under normal working conditions.
Leveraging NVIDIA AI Technologies
The deployment integrates DeepHow’s AI platform with several NVIDIA technologies that significantly enhance manufacturing intelligence.
Among these is the NVIDIA Metropolis Blueprint for Video Search and Summarization (VSS), which enables intelligent analysis of production video streams.
The solution also utilizes NVIDIA RTX PRO Servers, providing the computing performance required to process high volumes of visual data and execute sophisticated AI models directly within industrial environments.
Together, these technologies enable manufacturing workflows to become visible in ways that were previously impossible through traditional industrial software.
Rather than simply recording video footage, the system interprets production activities, understands workflow sequences, and generates actionable operational insights.
Vision-Language Models Expand Manufacturing Intelligence
One of the most advanced components of the solution involves the use of vision-language models.
Unlike conventional computer vision systems that simply recognize objects or movements, vision-language models understand relationships between visual information and manufacturing activities.
This enables the AI platform to identify individual process steps, classify production operations, recognize workflow variations, and interpret complex assembly activities.
The technology allows manufacturers to analyze sophisticated manual operations involving human workers, tools, components, and equipment with a much higher level of contextual understanding.
As manufacturing becomes increasingly flexible and customized, this capability becomes essential for accurately evaluating production performance.
Automatically Measuring Cycle Times
Cycle time measurement remains one of the most valuable performance indicators in manufacturing.
Traditional measurement methods require engineers to manually observe production and record start and stop times for every activity.
DeepHow automates this entire process.
Using AI-powered video analysis, the platform continuously measures production cycle times without interrupting normal operations.
The system captures detailed timing information across thousands of production cycles while automatically identifying variations, delays, and performance trends.
This allows manufacturers to collect significantly larger datasets than would ever be practical using manual studies.
The result is more accurate operational analysis supported by statistically meaningful production data.
Identifying Waste and Value-Added Activities
Beyond measuring cycle times, the AI platform evaluates how work is performed throughout each production cycle.
The system automatically distinguishes between value-added activities that contribute directly to product creation and non-value-added activities that may represent waste or inefficiency.
These insights help manufacturers identify unnecessary movements, waiting periods, process interruptions, material handling inefficiencies, and other operational bottlenecks.
By continuously identifying opportunities for waste reduction, manufacturers can improve productivity while supporting lean manufacturing initiatives.
Next Generation Factory Intelligence
DeepHow also announced plans to introduce additional capabilities that expand beyond traditional time and motion analysis.
Upcoming features include automatic report generation and Factory Flow Intelligence, powered by NVIDIA Cosmos and built using Video Search and Summarization agent skills.
These new capabilities will automatically generate detailed operational reports without requiring engineers to manually compile observations.
Factory Flow Intelligence will provide broader visibility into manufacturing operations by analyzing interactions across entire production lines rather than individual workstations alone.
This expanded intelligence supports more comprehensive decision-making across manufacturing facilities.
Turning Weeks of Analysis into Minutes
One of the most significant advantages of the platform is the dramatic reduction in engineering effort required to complete operational studies.
Historically, industrial engineers often spent several weeks observing production lines, collecting data, organizing measurements, and preparing reports.
DeepHow automates much of this work.
The AI system generates insights within minutes after analyzing production data.
This allows engineers to spend less time collecting information and more time implementing operational improvements.
The technology also enables frequent analysis rather than occasional studies, supporting continuous improvement initiatives across manufacturing operations.
Initial Deployment Begins in Mexico
The first phase of implementation will begin within Yazaki North America’s manufacturing operations in Mexico.
The company expects the AI-powered system to substantially reduce the time required for production line analysis.
Processes that currently require weeks of engineering effort are expected to be completed within just a few days.
This dramatic improvement in efficiency is projected to generate millions of dollars in annual savings through increased productivity, improved operational decision-making, and reduced engineering labor.
As the deployment expands across Yazaki’s global manufacturing network, management believes similar efficiency gains could translate into tens of millions of dollars in annual value.
Continuous Operational Improvement
The benefits extend well beyond faster engineering studies.
Traditional time and motion analyses are often conducted only a few times each year because of the extensive resources required.
With AI continuously monitoring production, manufacturing teams can analyze any production line whenever needed.
This enables engineers to detect bottlenecks earlier, respond more quickly to operational changes, and rebalance production lines before inefficiencies become permanent.
Continuous access to operational intelligence supports faster improvement cycles while increasing manufacturing flexibility.
Building a New Manufacturing Dataset
One of the most valuable long-term outcomes of the partnership is the creation of a completely new category of manufacturing data.
Modern factories already generate enormous amounts of machine information through industrial automation systems.
However, relatively little structured data exists regarding manual assembly work performed by human operators.
DeepHow’s platform captures detailed information about worker movements, assembly sequences, workflow variations, decision-making processes, and task execution.
Over time, this creates a continuously expanding digital record of how products are manufactured across the organization.
Combined with engineering expertise and advanced AI models, these datasets become powerful resources for improving production methods, increasing capacity, standardizing best practices, and identifying optimization opportunities across multiple factories.
Leadership Perspective
Joseph McCorry, Head of Commercial and Vice President of YNCA Business Units at Yazaki North America, described Operational Analysis Improvement as only the beginning of what the partnership can achieve.
According to McCorry, integrating video technologies, vision-language models, DeepHow’s AI platform, and the expertise of manufacturing teams creates opportunities to rethink factory design itself.
Rather than simply improving existing production lines, AI-powered manufacturing intelligence could help companies develop entirely new production environments that are safer, more efficient, and significantly more productive than current manufacturing systems.
About DeepHow
DeepHow is a Physical AI company focused on helping manufacturers achieve operational excellence by capturing frontline knowledge, verifying work quality, and optimizing production execution through artificial intelligence. Its platform is used by more than 100 customers across 1,500 manufacturing locations in 28 countries, serving industries including industrial manufacturing, automotive, electronics, pharmaceuticals, utilities, and food and beverage.
Looking Ahead
The partnership between DeepHow and Yazaki North America highlights the growing role of Physical AI in reshaping modern manufacturing. By replacing manual time and motion studies with AI-driven video analytics, vision-language models, and real-time operational intelligence, manufacturers can gain unprecedented visibility into production performance while dramatically accelerating continuous improvement initiatives.
Supported by NVIDIA’s advanced AI technologies, the deployment enables Yazaki North America to transform engineering workflows, optimize manufacturing operations, and build valuable datasets that will support future innovation. As manufacturers worldwide seek smarter, more adaptive production systems, collaborations like this demonstrate how artificial intelligence is evolving from an analytical tool into a core driver of operational excellence, productivity, and next-generation factory design.
Source link:https://www.businesswire.com/
