NTT DATA and Hyster-Yale Materials Handling, Inc. (HYMH) are bringing “physical AI” out of the lab and onto the factory floor, embedding intelligence directly into manufacturing workflows at HYMH’s Berea, Kentucky plant.
The new application uses sensor data so machines and systems can perceive, understand and act in real time inside actual production environments, rather than only in simulations or test cells.
In practice, the project introduces AI driven quality assurance into HYMH’s assembly operations, making it a first of its kind example of physical AI applied to industrial assembly by wiring intelligence into the production steps themselves.
The system is designed to help ensure lift trucks are built to consistently high standards, catching issues earlier and reducing the chance of defects slipping through downstream checks.
NTT DATA engineered the solution on site in Berea, integrating vision sensors, edge AI that processes data locally and advanced analytics into a critical assembly workflow. Running at the edge means the models and processing stay inside the factory, which supports faster rollout, lower latency and quicker time to value compared with traditional cloud only analytics.
Working with partner Archetype AI, NTT DATA and HYMH adapted a physical AI model that continuously compares assembly activity against expected production steps, verifying that parts are installed and stages completed before a unit moves forward.
When the system detects a deviation, it flags it immediately so operators can intervene, turning quality validation into an ongoing process rather than a final gate at end of line.
Early results suggest the approach is a step change in how AI can be used in manufacturing, with edge based physical AI cutting deployment timelines from months to weeks when measured against legacy techniques.
Shorter deployment cycles make it easier to iterate across different lines and sites, accelerating how quickly manufacturers can standardise best practices and scale quality gains.
As manufacturers accelerate automation, demand is rising for physical AI that can operate safely in complex environments while improving efficiency, quality and resilience.
NTT DATA says it is well placed to deliver these capabilities at scale by combining industry expertise with end to end services that integrate AI across IT and operational technology, enabling more intelligent, data driven operations.
Today’s deployment builds on a long running collaboration between NTT DATA and HYMH, as the partners work toward more adaptive manufacturing and explore how physical AI can be scaled to deliver repeatable, high quality production outcomes.
Leadership Comment
“Our confidence in physical AI continues to grow, and we’re starting to see the countless benefits that AI can bring to our global manufacturing operations,” said Barbara Binda, Director of Global Manufacturing Innovation at HYMH, highlighting the impact on frontline teams and product reliability. Shahid Ahmed, Global Head of Edge Services at NTT DATA, said the deployment shows physical AI working in real production environments with tangible impact on the factory floor, combining real production data with edge models to support workers and improve product quality.
