Key takeaways:
The workforce is the weakest link in smart manufacturing, by the industry’s own account. In Deloitte’s 2025 survey of 600 manufacturing executives, human capital ranked as the least mature of every smart manufacturing category, and only 48% had a training and adoption standard in place.
Skill expectations are outrunning the training to meet them. Across industries, 72% of workers say AI has already shifted the skills their jobs require, but only 36% feel they’ve been adequately upskilled, and only a third say leadership communicates about AI clearly.
The best AI tool stalls if the people expected to use it were never trained or brought along. The solution is contextualized, hands-on training and visible leadership support.
Manufacturing executives rate their own workforce as the least ready part of their smart factory efforts. In Deloitte’s 2025 Smart Manufacturing and Operations Survey of 600 executives at large U.S. manufacturers, human capital came in at the lowest maturity level of any category measured, below technology, operations, and quality. More than a third named adapting workers to the factory of the future as a top concern, and fewer than half had any training and adoption standard in place.
AI on a plant floor is only worthwhile when the people around it can use it. A predictive maintenance system that operators don’t trust, or a quality vision tool a line lead doesn’t understand, ends up being just expensive shelfware. The return depends on the workforce, and the workforce is where manufacturers are investing least.
The training doesn’t reach the ones who need it most
When manufacturers do invest in AI skills, the training tends to reach the people who need it least. Deloitte found the most common workforce development tactics are in-house leadership training (53%) and external training programs (43%), structured programming that more readily reaches managers and salaried staff than the operators on shift.
This happens outside of manufacturing, too. Boston Consulting Group’s 2026 AI at Work study, based on a survey of nearly 12,000 workers, found that 72% say AI has already changed the skills their roles demand, but only 36% feel they’ve received adequate upskilling. Just a third of frontline employees say leadership’s communication about AI is clear, and only 28% see a strong link between what leaders say about AI and what their organization actually does. (Note that BCG’s frontline sample skews toward white collar roles rather than the plant floor.) In short, the people closest to the work are furthest from the training.
A tool nobody uses has no return
It’s not a true adoption if you skip the plant floor. BCG describes frontline AI use hitting a ceiling, with leaders and managers getting ahead while frontline usage stalls. What moves frontline adoption most is visible leadership backing. BCG found that when employees sense strong leadership support for AI, the share who feel positive about it climbs from 15% to 55%, but only about a quarter of frontline employees report getting that level of support.
Without that support, an operator who doesn’t trust a vision system’s reject call will override it. A line lead who was never shown how a scheduling agent reaches its recommendations will fall back on the spreadsheet. A maintenance tech who sees AI as a threat will work around it. Every one of those is a training and trust failure rather than a technology failure, and it wastes the capital spent on the tool.
The skills problem adds to the challenge. A 2024 Deloitte and Manufacturing Institute talent study found demand for specialized capabilities like simulation and simulation software skills jumped about 75% over five years, and the same study projected up to 1.9 million manufacturing jobs could go unfilled by 2033. The bar on a frontline role keeps rising, but the pipeline to clear it stays thin.
What floor-level AI training looks like when it works
Reaching the floor takes a different approach than booking managers into a workshop. The training has to meet operators in the work, in their language, on their schedule. A few principles separate programs that change behavior from those that check a box:
Train in the workflow, not the classroom. Short, hands-on instruction tied to the actual task and machine sticks better than a generic AI seminar. Operators learn the tool by using it on the line.
Use the floor’s own experts. Designate respected operators as AI champions and train them first. Peers adopt faster when the person showing them the tool is someone who does their job rather than a consultant.
Make leadership support visible. The data shows that frontline attitudes swing on whether leaders back the technology in a way workers can see. Show up on the floor, explain the why, and connect the tool to the operator’s day rather than the company’s dashboard.
Close the communication gap. With only a third of frontline workers saying leadership’s AI message is clear, plain and consistent communication about what a tool does, and what it doesn’t do to someone’s job, is itself a form of training.
One encouraging point from Deloitte’s survey is that manufacturers are targeting human capital for their largest maturity improvement, ahead of every other category. They know what needs to happen; now the work is turning it into floor-level capability, which will determine whether the AI earns its keep.
The plant floor is where food manufacturers expect AI to deliver, whether that’s in maintenance, quality, scheduling, or yield. It’s also where the training is thinnest and the trust is lowest. Aligning the investment in people with the investment in technology is what turns a purchased tool into a working one.
