The AI conversation has a gap nobody's talking about. The discussions center on chatbots and content tools, on industries where the feedback is instant and the output fits on a screen. What we're seeing with mid-market manufacturers is a different story—and a more complicated one.

The pressure isn't just on the production line. It's hitting finance, sales, HR and every function in between.

Kaufman Rossin's research report, The State of AI in the Mid-Market, which surveyed senior decision-makers across U.S. mid-market companies, makes the divide clear: industrial and mid-market manufacturers are experimenting plenty, yet scaled, company-wide AI deployment remains rare. The wave of disruption has reached them at last, but the foundation underneath isn't ready to hold it. 

Disruption travels through the value chain 

Digital disruption over the past decade didn't hit every industry at once. It moved like a wave, starting at the customer-facing edge of the economy. Retailers, banks, and consumer brands felt the pressure first, so they transformed first, pouring investment into data infrastructure, digital platforms, and new ways of working. 

That pressure is now rolling backward through the value chain. Manufacturers, distributors, and industrial suppliers are being asked by their customers and partners to digitize, integrate and automate. These companies aren't slow;  they were simply last in line. The difference is that their runway is shorter, and the expectations arriving at their door are already fully formed. 

Is AI creating a data problem, or just exposing it? 

Here's the uncomfortable truth: AI doesn't work without clean, connected, accessible data, and industrial companies have historically not invested in that foundation.

The numbers make the gap hard to ignore. 

Only 27% of manufacturing companies in the survey have a data warehouse or data lake, compared with 60% across the broader mid-market. Some 45% still operate with siloed data, and none use machine learning platforms. Look across the entire mid-market and the picture isn't much brighter: just 16% have reached a fully governed and integrated data state. 

Then there's the legacy-system challenge. Every manufacturer in the research runs on ERP, and those deeply embedded systems don't connect easily to modern AI tools. It's no surprise that legacy integration ranks as the top barrier in manufacturing at 55%, well above the 41% market average. 

There's a cultural layer underneath the technical one. Industrial companies built their competitive advantage on operational expertise, process mastery, and deep domain knowledge, not data-driven decision-making. Gut instinct earned through decades on the floor has served these businesses well. AI asks them to operate on a different assumption, and that shift is harder than installing any tool. 

What "stuck in testing" really means 

The data is striking. Some 73% of manufacturing companies are still in the testing phase, and not one in the research has become a full operator, meaning that AI is simply how the business runs. Zoom out to the whole mid-market and 73% of companies remain at early or foundational readiness. Only 7% are ready to scale company-wide. 

Today's wins are real, but they're narrow. Time savings here...accounts payable automation there. Individual productivity gains that help one person move faster within a process that still spans disconnected systems. Those wins are worth celebrating, but they aren't transformation. 

The risk is mistaking a successful pilot for a finished journey. Experimentation feels like progress, and it is, until it stalls. Organizational readiness bridges the gap between a promising pilot and operational scale. 

Foundation first, then scale 

The good news is that the will is there. Every manufacturer surveyed agrees AI saves time, and 91% plan to increase their investment. That appetite is exactly what makes this moment so important, because investment without a foundation produces more pilots, not more scale. 

Three priorities can change the trajectory. 

Get your data connected. Start by knowing what data you have, where it lives, and how clean it is. Break the silos that matter most and invest in one or two integration platforms that link your most-used systems. You don't need a full enterprise overhaul to begin. You need targeted progress on the data that powers your highest-value work. 

Start with data-ready use cases. Resist the urge to force AI onto broken or fragmented data. Find the processes where your data is already clean enough to prove value at an enterprise level, then build outward from those wins. 

Treat this as a culture shift, not an IT project. This is the hardest and most important move. Leadership has to reframe data from a back-office function into a strategic asset, and model that mindset across the organization. Tools don't transform companies; people do. 

The wave is here 

The same disruption that reshaped retail and finance has arrived on the shop floor. That isn't cause for alarm, but it is cause for urgency. The companies that pull ahead won't be the ones that bought the most tools. They'll be the ones that built the foundation, connected their data, and treated AI as the organizational transformation it can be. 

The technology is ready. The real question is whether your operating model is ready to put it to work.