Manufacturers hit data bottleneck in AI scale-up push

Rockwell Automation has published research on how food and beverage manufacturers are trying to scale artificial intelligence beyond pilot projects. The findings point to data use as a central barrier to wider deployment.

Drawing on its 2026 State of Smart Manufacturing Report, Rockwell found that almost 90% of food and beverage manufacturers already use AI and machine learning or plan to do so soon. Yet they effectively use only about 45% of the data they collect, limiting efforts to turn early AI projects into broader operational change.

The research suggests the sector has moved past asking whether AI has practical value and is now focused on how to extend successful trials across manufacturing networks. That shift places greater weight on data quality, system integration and workforce preparation than on launching individual tools.

Manufacturers making more progress with AI tend to place it within wider digital transformation programmes. In practice, that means linking operational and business systems, improving access to data, and tying AI work to specific targets such as product quality, productivity, cost control and risk reduction.

For many companies, however, AI efforts remain isolated. When data sits in separate systems or processes are not standardised, manufacturers can struggle to repeat results from one site or function across a larger organisation.

Data bottleneck

The report places particular emphasis on connected, contextualised data. AI systems depend on reliable information from across production and business operations, and fragmented datasets can limit the quality of the insights they produce.

That issue becomes more pressing as manufacturers increase their use of AI in day-to-day processes. Industry respondents expect AI-supported manufacturing processes to rise from roughly 37% at present to more than 60% by 2030.

The findings also underline the role of IT and operational technology integration. Bringing those systems together can improve visibility across plants and supply chains, reduce information silos, and give companies a clearer basis for faster decision-making.

For food and beverage groups, this has direct operational relevance because the sector faces constant pressure over product consistency, throughput, cost management and safety controls. AI projects not connected to those priorities may be harder to justify beyond an early trial phase.

Rockwell's analysis suggests companies scale AI more successfully when they create repeatable processes rather than rely on standalone use cases. Stronger foundations in data management and system integration make it easier to apply AI across multiple sites or functions without rebuilding projects from scratch each time.

Workforce impact

The study also highlights labour and skills as a major part of the equation. It found that 93% of manufacturers expect smart manufacturing initiatives to reshape their workforce, while 40% said they had reskilled employees during the past year.

That points to a broader industrial pattern in which adoption of digital tools depends not only on software and equipment spending, but also on whether staff can work with new systems. Reskilling can include training employees to interpret AI-generated insights, manage integrated data environments, or oversee more automated production processes.

Greg Schmidt, Vice President, CPG Global Industry, Rockwell Automation, said the market conversation had moved on from proving that AI can work.

"Many manufacturers have moved beyond asking whether AI can deliver value," Schmidt said.

"The focus now is on how to scale successful pilots across the enterprise. That requires more than new technology. It requires connected data, integrated operations, and a clear link between AI initiatives and business outcomes."

The figures reflect a wider issue facing manufacturers as they try to convert digital investment into measurable returns. Early enthusiasm for AI has produced a large number of pilot schemes across industry, but turning those tests into sustained operational gains has proved more difficult where data is incomplete, systems remain separate, and workforce planning lags behind technical change.

In Rockwell's assessment, the manufacturers most likely to achieve broader results are those that treat AI as part of a larger operating model rather than as a single technology project. These organisations are building stronger data foundations, integrating IT and OT systems, and preparing workers for a different production environment as AI-supported processes move towards more than 60% by 2030.