As manufacturers across the U.S. increasingly digitize their operations, data management and standardization have become a growing concern. 

More than 60% have a strategy or guidelines for data management, but only 15% follow the plan in its entirety, according to a recent survey by the National Association of Manufacturers (NAM). 44% of manufacturing leaders say the amount of data they collect has doubled in the last two years alone, and it’s anticipated to triple by the end of the decade. 

86% of the respondents believe that the effective use of manufacturing data will be essential, yet only 1 in 4 have high confidence that the right data is being collected.

AI models and agentic systems depend on accurate, real-time data for every decision, so when this data is wrong at the source, it can have ripple effects on model output and cost millions, said Prateek Kathpal, president of SymphonyAI’s Industrial Division. The company makes IRIS Foundry, specifically designed to integrate data from IT, OT, and engineering to deliver AI-ready analytics. 

“Machines in plants generate massive volumes of data that are often trapped in isolated systems, limiting visibility and slowing down everything from root-cause analysis to predictive maintenance,” he said. 

AI ambition vs. AI readiness

While most companies are excited about the potential AI presents, few are actually prepared to implement it. 

There’s a gap between AI ambition and AI readiness, said Jasmeet Singh, executive vice president and global head of manufacturing at Infosys. “In the same way that physical infrastructure determines how efficiently a factory runs, data infrastructure now determines how intelligently an enterprise can operate.”

Specright, a data management platform, partnered with UserEvidence to survey 45 manufacturers across different industries and found 60% were managing specification data in spreadsheets with tools like Excel or Google Sheets. Nearly half relied on shared drives like SharePoint or Google Drive, and 38% were pulling from ERP systems that weren’t designed for spec-level complexity. Nearly 7 in 10 admitted to struggling with maintaining up-to-date data, and the average team loses 468 hours a year to manual spec tasks. x

Many teams have worked around broken systems for years, said Mike Boese, CEO of Specright. “Changing those habits – and getting buy-in across R&D, packaging, quality, operations, and IT – requires both a cultural shift and executive sponsorship that not every company has mobilized yet.”

“When that’s the foundation you’re building AI on, you’re not going to get intelligent outcomes,” he added.

“You’re going to get “garbage in, garbage out” – just faster.”

This problem might be arising from the push for immediate results. “Boards want AI returns on a data foundation that doesn’t exist yet,” said Brian Zakrajsek, specialist leader, smart manufacturing and operations at Deloitte. The pressure to show AI value makes organizations fund the use case and under-fund the foundation, he added. 

In-house vs third-party solutions

Even when companies decide to work on data management and standardization, a major roadblock lies in whether to develop the technology in-house or work with third-party solutions. 

“The very large manufacturers almost always start with an in-house attempt,” said Maor Farid, founder and CEO of Leo AI, a company building AI for mechanical engineering. “They have data science teams, a budget, and the understandable instinct that engineering knowledge is too sensitive and too specific to hand to an outside vendor, especially defense and medical-device customers, who have the strictest requirements.”

Manufacturers understand their products, processes, plants, and customers better than anyone, and that domain expertise is essential, Singh said. 

“However, scaling AI and data modernization across a large manufacturing enterprise also requires specialized capabilities in cloud, platform engineering, cybersecurity, data governance, AI model operations, system integration, and change management,” he said.

This is making more manufacturers turn to external solutions, especially as trying to build systems from scratch can lead to lengthy, costly rollouts, Kathpal said. Increasing competition is also a key factor, as “your competitors could be deploying projects at scale and already seeing results,” he added. 

Use cases with the best results

Some of these results include solving the biggest problems plaguing the industry for decades. 

One of them is “preserving and transferring knowledge that lives in the heads of senior people who are retiring, and in design decisions buried in old projects that no one ever documented,” Farid said. 

When that knowledge is captured in an accessible, searchable form, onboarding new talent becomes more efficient, senior workers can focus on important tasks instead of training workers, and the company doesn’t lose crucial data when someone leaves, he said. 

“When an engineer can ask a plain-language question and instantly get an answer drawn from their company's own data and the relevant industry standards, the time savings are dramatic,” he added. “For regulated industries, this doubles as a compliance benefit, because traceable, sourced knowledge is exactly what an FDA or ISO audit wants to see.” 

“Sustainability reporting requirements are expanding rapidly – Extended Producer Responsibility (EPR)

enforcement alone can expose companies to tens of thousands of dollars in penalties per day, per state,” Boese said. EPR is a policy that holds businesses responsible for the end-of-life recycling and disposal of their products and packaging.

“As the regulatory environment continues to tighten, companies with clean, connected spec data are going to have a significant competitive advantage,” Boese added.

Another use case is supply chain optimization. “With better visibility across suppliers, inventory, logistics, and demand, manufacturers can use AI to improve forecasting, optimize flows, and respond faster to disruption,” Singh said. 

Accessible data is also helping with more efficient predictive maintenance. Manufacturers can enable real-time alerts on equipment anomalies and reduce unplanned downtime, Kathpal said. Doing this prevented one of their clients from losing $2.5 million annually due to operational inefficiencies, and the figure climbs substantially higher for mega corporations. 

“The manufacturers who win the next decade will not be the ones with the fanciest AI,” Farid said. “They will be the ones treating data as the foundation rather than an afterthought.”