DEEPHOW + YAZAKI NORTH AMERICA + NVIDIA

Yazaki North America Advances Factory Operations with Physical AI and Visual Agents

How one of the world’s largest privately held automotive suppliers is transforming their operations to drive productivity and workforce efficiency.

Boosting Productivity By Automating Time Studies

Powered by advanced vision-language models, DeepHow Time and Motion AI automatically captures cycle times and classifies operator actions across production cycles, including value-added, non-value-added, and waste.
INDUSTRY
Automotive Manufacturing
CAPABILITIES
Time and Motion AI

CHALLENGE

Automotive suppliers operate in complex, manual-assembly environments. Small differences in motion, sequence, station balance and cycle time can compound across thousands of workers and production cycles. Yet the traditional process for identifying those losses remains manual, resource-intensive and too infrequent to keep pace with changing production conditions.

SOLUTION

DeepHow Time and Motion AI automates the manual observation, timing, and classification traditionally required for production studies. Yazaki North American gained an on-demand view across operators, stations, and shifts, allowing teams to find bottlenecks, rebalance lines, update standard work, and improve throughput faster, while freeing industrial engineers to focus on process improvement instead of data collection.

THE RESULTS

Days
Automation
Through agents and VLMs, reducing line-analysis time from weeks to days.
Minutes
Vision AI Agents
Removing bias from manual observation and generating insights from video in minutes.
Millions
Workforce Efficiency
Projecting to save tens of millions annually by reducing the time and costs of generating line analysis.
As Needed
On-Demand Process
Conducting line analysis more frequently, as needed, instead of limited pre-scheduled times.
A COLLABORATIVE PARTNERSHIP

Delivering Yazaki North America, NVIDIA, and DeepHow Objectives

DeepHow, Yazaki North America, and NVIDIA are moving beyond incremental improvements and creating a new model for how manufacturing knowledge is captured, analyzed, and applied.
DeepHow Time and Motion AI is a starting point, but greater potential comes from creating a new source of manufacturing data. While factories capture extensive information from machines, they have far less visibility into the detailed movements, sequences, variations, and decisions that shape manual assembly.
By converting those workflows into structured, analyzable data, Yazaki North America can develop an evolving record of how work is performed across its operations. Combined with engineering expertise and vision-language models, this information can reveal better methods, recover capacity, and surface opportunities to improve quality, throughput, and productivity.
THE OPPORTUNITY

Remove Bottlenecks, Reduce Waste, and Increase Productivity in Less Time

Starting with its operations in Mexico, Yazaki North America expects to compress line analysis from weeks to days, generating millions of dollars in annual value. Applied across its global manufacturing network, those gains grow to tens of millions of dollars each year.

Make Lost Capacity Visible

Manual time studies capture only a limited snapshot, making it difficult to see where cycle variation, imbalance, and waste reduce throughput.

Analysis in Minutes

DeepHow automatically captures cycle times and classifies value-added work, necessary non-value-added work, and waste across production cycles.

Optimize Lines On Demand

Structured workflow data allows teams to compare operators, stations, and lines, improve methods, and rebalance production across the manufacturing network. 

"Manual observation and end-of-line inspection identify defects after they've already moved downstream — rarely revealing the execution issues that caused them."

"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."
Read More
Joseph McCorry
Head of Commercial & VP YNCA Business Units at Yazaki North America
DEEPHOW TIME AND MOTION AI

Generating Time Studies 10x Faster

DeepHow Time and Motion AI uses physical AI and visual language models to automatically classify cycle time and operators' work on the factory floor.

Physical AI
Understands operator actions, work sequence, cycle variation, and production flow.
Powered with NVIDIA
DeepHow uses Cosmos Reason and NVIDIA Metropolis Blueprint for video search and summarization (VSS).
Automatic Cycle Detection
Identifies cycle start and end points across repeated production activity.
Micro-Action Classification
Classifies each workflow segment as value-added, necessary non-value-added, or waste.
Line Balance Visibility
Compares operators, stations, and cycles to reveal bottlenecks, imbalance, and lost capacity.
Faster Continuous Improvement
Turns time studies that take weeks into actionable analysis in days or minutes.

See Live Time and Motion AI in Action

See how DeepHow and NVIDIA are helping manufacturers like Yazaki North America automate their time and motion studies.

Request a Demo

Generate
AI Summary

Request a DemoHelp Center
Use Cases
Automate Time and Motion StudiesVerify Standard Work in Real TimePreserve Industrial KnowledgeDigitize Work Instructions
Platform
Knowledge Capture
On-demand Visual InspectionLive SOP VerificationTime and Motion AIAI CapabilitiesPharmaCloud
Industries
Customer StoriesManufacturingPharma & Medical DevicesElectronics Assembly
Resources
BlogProduct Login

Generate
AI Summary

Copyright © 2026 DeepHow
PrivacySecurityTerms and ConditionsAccount & Data Deletion