There’s a quiet shift underway in manufacturing. You won’t hear it from the loudest voices or the splashiest product announcements. But it’s happening.

AI is no longer optional. According to KPMG’s recent global survey of 183 senior AI leaders in industrial manufacturing, 93% believe that companies that integrate AI at the core will pull ahead—permanently. The question is no longer if AI belongs in manufacturing. It’s how to do it in a way that actually matters.

What the report makes clear is this: most manufacturers are still early in that journey.

Hot Take #1: The industry is stuck in pilot mode

Despite almost every manufacturer experimenting with AI in some form, only 20% say AI is core across every department. Even fewer—just 26%—have embedded it deeply into the operating model.

And yet 96% report gains in efficiency. The benefits are real. What’s missing is scale. The problem isn’t AI—it’s that too many initiatives are stuck in isolation, without a plan to connect them.

You can feel the tension between what’s reported and what’s real. Teams say they’re ready to go end-to-end with AI (91%)—but on the ground, most aren’t there yet. Not even close.

Hot Take #2: Data integration—not AI models—is the real barrier

Everyone wants to talk about models. But what actually holds teams back is data infrastructure. More than half (56%) cite data challenges as a major obstacle. The issue isn’t access to AI—it’s that data is fragmented, siloed, and often unusable in real time.

Even among the 84% building AI in-house, the plumbing is a mess. Tools aren’t fixing the foundational problems. Until that’s addressed, companies will keep hitting the same ceiling: pilots that never scale.

Hot Take #3: Agentic AI is here—and quietly changing how decisions get made

There’s been plenty of hand-waving about autonomous agents. But the reality is more grounded—and more significant.

Sixty-seven percent of manufacturers already use agentic AI in production, and another 20% are planning to. These aren’t moonshot experiments. They’re quietly embedded in production schedules, procurement systems, and logistics flows.

What’s striking is that 91% of leaders say they’re comfortable letting AI run workflows end-to-end. That kind of trust doesn’t come from a slide deck. It comes from seeing real-world performance.

Hot Take #4: The workforce is adapting faster than expected

There’s a persistent narrative that AI will be met with resistance. The data paints a different picture: 89% of leaders say their teams are adapting well, and 80% have invested in AI training.

People aren’t the problem. They’re the path. The companies that invest in helping employees adapt—without patronizing them—are already seeing the payoff in productivity and morale.

And yet, 76% of companies still prefer to “wait and watch” before making big bets. That hesitation is understandable—but it may also be the gap that decides who leads and who follows in the next wave of manufacturing.

Hot Take #5: AI maturity is about trust, not just tooling

It’s easy to assume that technical capability equals readiness. But trust and governance matter just as much. Seventy-five percent of leaders now use AI to inform decisions across multiple business areas. Sixty-five percent have formal frameworks for risk management.

This is the frontier: not just automating tasks, but shaping how decisions are made. The companies that are winning aren’t just plugging in tools. They’re rethinking processes—and building trust around the outputs.

As one respondent put it: “Autonomy isn’t a moonshot anymore.”

Where we go from here

If there’s a single theme in this report, it’s that progress is uneven. There are pockets of excellence—but also a lot of friction. Most organizations are still figuring out how to make AI practical, responsible, and scalable.

In our work at Broadn, we see one of the most overlooked gaps: quoting. It’s not flashy, but it’s critical. Sales teams are still buried in spreadsheets, tribal knowledge, and outdated catalogs. And that lag kills deals.

Automating quoting—accurately, quickly, and intelligently—is one of the highest-leverage places to apply AI. Not because it’s trendy, but because it’s where value leaks out today.

The path forward isn’t about chasing the newest tool. It’s about building systems that are coherent, reliable, and tightly integrated. The companies that do that—quietly, patiently—are the ones who will define the next decade of industrial growth.

Not through noise. Through execution.