73% of mid-market manufacturers are still testing AI, with zero at full deployment

Not one mid-market manufacturing company surveyed has reached full, company-wide AI deployment. That is the headline finding from Kaufman Rossin's State of AI in the Mid-Market report, which polled senior decision-makers across U.S. mid-market firms and was published by Automation World in July 2026. The data is striking in its specificity: 73% of manufacturing respondents remain in the testing phase, and the bottleneck is not ambition but infrastructure.

The data gap manufacturers can't work around

AI tools require clean, connected, accessible data. Mid-market manufacturers, by and large, do not have it. The Kaufman Rossin survey found that only 27% of manufacturing companies have a data warehouse or data lake in place. Across the broader mid-market, that figure is 60%. The gap is not marginal; it is structural.

Siloed data compounds the problem. According to the same research, 45% of manufacturers still operate with data spread across disconnected systems, and none of the manufacturers surveyed use machine learning platforms. Even zooming out to the full mid-market sample, only 16% have reached a fully governed and integrated data state.

Legacy ERP is the integration wall

Every manufacturer in the Kaufman Rossin research runs on ERP. Those systems are deeply embedded and do not connect easily to modern AI tooling. The report identifies legacy integration as the top barrier to AI adoption in manufacturing, cited by 55% of respondents. The broader mid-market average sits at 41%, meaning manufacturers face a meaningfully higher integration burden than their peers in other sectors.

The issue is not just technical. Industrial companies built competitive advantage on operational expertise and process mastery, not data-driven workflows. Shifting to AI-informed decision-making requires more than a software deployment; it asks leadership to reframe how operational knowledge is generated and used. That cultural layer sits underneath the technical one, and it is harder to address than any integration project.

Pilots are real wins, not finished journeys

The wins that manufacturers have captured are genuine but narrow. Process-level time savings, accounts payable automation, individual productivity gains; these outcomes matter, but they live inside workflows that still span disconnected systems. Vera Nieuwland, director of Kaufman Rossin's business consulting services practice, writing in Automation World, described the risk as mistaking a successful pilot for a completed transformation.

Investment appetite is not the constraint. The Kaufman Rossin data shows that 91% of manufacturers plan to increase their AI investment. That momentum is meaningful, but without addressing the underlying data infrastructure, additional spend is more likely to produce more pilots than operational scale.

Three priorities for moving from pilot to production

The Kaufman Rossin analysis, as reported by Automation World, points operations and IT leaders toward three sequenced moves. First, map and connect the data that powers the highest-value work; full enterprise overhauls are not a prerequisite, but targeted integration of the most-used systems is. Second, find the processes where data is already clean enough to prove enterprise-level value, and build outward from those, rather than forcing AI onto fragmented data. Third, treat AI readiness as an organizational shift rather than an IT project, with leadership actively repositioning data as a strategic asset rather than a back-office function.

The sequencing matters. Organizations that skip the data foundation step and move directly to broad AI deployment are the ones generating pilots without scale. Manufacturers that 91% plan to increase AI investment are exactly the organizations that need a clear decision framework before that next budget cycle closes.

What this means for your team

• Audit your data warehouse and integration status before committing to the next AI platform purchase. The Kaufman Rossin data suggests most mid-market manufacturers are spending ahead of their data readiness.
• Map your ERP integration gap specifically. With 55% of manufacturers citing legacy integration as their top AI barrier, the integration strategy deserves its own workstream, not just an IT ticket.
• Identify two or three existing processes where data is already clean and connected, and prioritize AI pilots there. Proving value on solid data is more useful than piloting on fragmented sources.
• Evaluate whether your AI program has executive sponsorship that treats data governance as a business priority. The cultural shift is the factor most correlated with moving past the testing phase, according to the Kaufman Rossin findings.