As manufacturers continue to invest in AI, a new report from Hitachi has found that deployment isn't equal across uses as organizations struggle to move from adoption to full-scale factory deployment.

See also: How AI agents will transform MES and manufacturing  

See also: Smart Industry's annual State of Initiative Survey is open!

The report, Beyond the Pilot: Industry's Physical AI Progression, draws on answers from 236 completed responses from manufacturing, transportation and logistics, energy and other industries across six global markets: the U.S. and Canada, Germany, the U.K., Japan and India.

According to Hitachi, AI-powered inspection is leading factory deployments as 43% of manufacturing respondents reported production deployments of computer vision for inspection, followed by automated quality inspection (38%) and predictive maintenance (34%).

However, advanced automation falls behind, as only 19% of respondents reported production deployments of collaborative robots and 15% of smart CNC machines, while 45% are still testing AI-driven process optimization.   

According to Hitachi, this indicates the challenge with industrial AI stems from moving from adoption to scale. 

Across the industries surveyed, 97% reported using industrial AI in some capacity, yet only 8% have reached continuous optimization and innovation. 

The report found that the largest obstacles are in skills and legacy systems. Over 40% of respondents across industries reported that limit AI expertise was a leading barrier to adoption, while 31% reported integration challenges as their main barrier.  

See also: How data, governance and organizational change define AI success   

This echoes other recent studies that have found AI implementation barriers in the manufacturing industry. 

For example, Futurum found that the top three agentic AI use cases for manufacturers so far are materials planning, customer orders, and inventory as organizations struggle with the capacity to invest in agentic AI, according to market research commissioned by industrial AI software provider IFS.     

However, over 90% of manufacturing organizations are using some form of AI as adoption expands beyond automation and into predictive, generative, and connected applications despite supply chain uncertainties and implementation challenges or outright failures, according to another survey from Silicon Valley software and digital transformation stakeholder Rootstock.  

See also: The death of conversational chatbots in industrial organizations 

Other key Hitachi findings include: 

• 87% of respondents associate delaying adoption with material risk, and over two-thirds view physical AI as very or extremely important to company goals. 

• 75% of respondents agree that leadership actively supports AI initiatives, although only 38% believe they are adequately resourced for adoption. 

• Only an average of 16% of respondents reported significant AI usage by frontline workers and maintenance and repair teams via collaborative robots, smart PPE, and other human-centric applications.   

• Most AI solutions in production are used in inspection and predictive maintenance applications. However, investment is shifting to behavioral analysis, smart computer numerical control machines, and computer vision for safety.