Correct me if I’m wrong but are two camps basically forming about industrial AI? It’s sometimes hard to see through the hype of this technology—driven a lot by software companies trying mightily to vend new products. But AI does seem to be on a fast track to permanence. It has such momentum—and billions behind it—that it can’t be held back.

Smart Industry posts a lot of surveys about manufacturing technology trends. We do this on purpose. This is such an age of change in industrial advancement that how else are we to know which way the wind is blowing? And we solicit a fair number of opinions, too, from some leading voices. This is true in spades with industrial AI.

Podcast: Takeaways from IFS Unleashed and how industrial AI is advancing

One such survey came out this week from Rootstock, the Silicon Valley vendor of a cloud-native ERP platform. They build manufacturing, distribution, and supply chain management solutions natively on the popular Salesforce platform.

Rootstock’s research found that three-quarters-plus of manufacturers—77%— plan to increase budgets or resources for AI-powered tools over the next 12 to 18 months, so AI’s march continues onward, though that percentage is down slightly from 82% in 2024.

Rootstock’s 2026 State of AI in Manufacturing Survey also reveals that AI is moving further into everyday plant operations. It’s one thing to pilot, it’s another to integrate for everyday operations.

Nearly two-thirds of manufacturers (64%) have progressed beyond exploration, according to Rootstock, with AI initiatives in pilots or already operating, integrated, or broadly supporting business processes.

See also: ServiceNow research warns of ungoverned agentic AI risks

The Rootstock survey also shows manufacturing professionals balancing optimism about AI with concern about its possible impact on jobs.

To wit: 41% of respondents expect AI to make them more productive, while 61% are at least somewhat concerned the technology could limit career opportunities. The results suggest that as AI becomes more embedded in work processes, companies need to consider worker attitudes alongside AI investment, according to Rootstock.

CEO Rick Berger expanded on AI’s potential to impact jobs on the factory floor.

“AI’s potential isn’t only about what the technology can do—it’s also about how well employees are set up to adopt it,” Berger said in a Rootstock release.

“With AI, manufacturers have the potential to boost productivity and help employees deliver more value in their work, but realizing AI’s full value requires vision, training, and a commitment to keeping human judgment at the center of key processes.”

Just this week, Smart Industry posted a piece by Rachel Zepernick of Thomas on how Xometry research was showing that AI and automation operators are now the hardest roles for manufacturers to fill, beating out quality engineers, maintenance technicians, machinists and CNC programmers.

AI is here, folks. The workforce is going to have to adapt.

See also: New IFS study reveals gap between AI strategy in the C-suite and actual factory-floor usage

Back to the Rootstock survey’s positive outlook for AI itself.

Among the types of AI being adopted by manufacturers, agentic AI is gaining traction: 31% of respondents deploying AI say they are already using agentic, which can independently plan and execute tasks with minimal human intervention. Automation remains the most widely deployed form of AI at 52%, followed by generative (49%) and predictive (40%).

Added Rootstock VP of Product Ohad Idan” “Barriers around AI are evolving. In 2024, skills, integration, and cost were key hurdles. Today, all three have eased, and the top concern is whether AI’s output can be trusted.”

Speaking of trust, a couple more tidbits from Rootstock’s research and some more from the vendor of a remote computer access application, TeamViewer, before we plug our in-house skeptic: Manufacturers want AI with human guardrails, according to Rootstock. In its survey, 73% prefer either human-directed AI or AI that can operate autonomously within defined limits and escalate when needed, while only 13% prefer fully autonomous AI.

See also: As technician demand in manufacturing explodes, AI and adjacent industries could help

Governance is still developing—a point that arose from another recent study, this one by global enterprise software vendor and AI company IFS that Smart Industry intercepted at IFS Unleashed in Orlando a few weeks ago. Employee training, according to Rootstock’s data, is the most common responsible AI practice, 40% of respondents said.

TeamViewer found that 46% of respondents to its research want human involvement when AI actions could affect security, data or operations.

U.S. manufacturers in TeamViewer’s study said AI promises to help resolve IT issues before they disrupt production, but demonstrating AI’s business value remains a challenge. Its findings show that 68% of manufacturing IT decision-makers use these approaches, compared with 56% in tech.

At the same time, nearly a third (32%) cite uncertainty about AI’s value as a barrier to adoption. Security, privacy, and risk concerns are another significant hurdle, cited by 47% of respondents in the TeamViewer data.

November is 'AI skepticism month'

We’re borrowing ARC Advisory Group’s Colin Masson this month to “bring the burn” on AI on purpose, to inject a healthy amount of skepticism about this tech in manufacturing and to cut through some of the vendor euphoria about its role in digital transformation.

See also: The death of conversational chatbots in industrial organizations

A new part of Colin’s five-part series will debut every Monday the rest of this month, on Nov. 9, Nov. 16, Nov. 23, and Nov. 30. Check out Part 1 out already: Why billion-dollar AI model claims fall apart on the factory floor.

Colin does get technical (a bonus for our audience), but his viewpoints are clear-eyed and cautionary about AI. He’s going to strip away the hyperbole, which we see as healthy.