October 23, 2025 Laying the Foundations for the AI Factory of the Future

Sharath Prasad AVP, IoT & Engineering at Cognizant, explains how manufacturers can prepare their operations for a future built around AI.    

Artificial intelligence (AI) is fast becoming a foundational technology within smart factories across the United States. With a vast array of potential use cases, it offers opportunities to boost productivity, efficiency, and bottom-line performance. 

The potential is such that more than half of manufacturers (51%) are already using it within their operations—and 80% say it will become essential to growth by the end of the decade. 

However, while interest is clearly high, there are numerous blockers preventing rapid adoption. These include regulatory issues and skills shortages. But, chief among the challenges, as highlighted in a recent report by the National Association of Manufacturers (NAM), is data quality and accessibility. 

The NAM report revealed that almost two thirds of manufacturers (65%) lack the necessary data to enable AI applications. A key reason for this is the continued presence of legacy assets that produce little or no usable information. On average, around a third of factory assets are still reliant on legacy equipment. 

Quality and contextualized data is the cornerstone of AI, so, if manufacturers want to ensure their factories are future-proofed, they will need to invest—or risk falling behind in an increasingly competitive landscape. To make data available to AI solutions, they will either be required to replace legacy assets or retrofit existing equipment with sensors—and then build out a reliable way to access contextualized data for AI applications.

Accessibility and usability

As a result of data now coming from various sources—including both new and legacy machinery—manufacturers will need to create an infrastructure that’s capable of pulling this information together. This presents a significant challenge for manufacturers, as 62% cite unstructured or poorly formatted data as a major barrier to adopting AI. 

While there is no standard way of making machine data AI ready, manufacturers are increasingly opting to tackle this issue by adopting the unified namespace (UNS) framework to provide a single source truth. UNS is an approach more synonymous with IT projects, but manufacturers are adapting it for operational technology. 

This effectively adds an extra layer of technology to standardize data coming from various devices and industrial control systems, such as supervisory control and data acquisition (SCADA) systems and programmable logic controllers (PLCs). 

It also helps to take high-speed and high-volume data and convert it into a seamless  flow of information for manufacturing execution systems (MES), which typically run at a lower speed and at lower volumes. UNS also provides contextualization, which enables AI solutions to identify the source of operational issues. 

The ability to take a holistic view of factory operations opens up endless possibilities. For instance, it enables a digital twin of a manufacturing operation to be created. This would allow AI to simulate different scenarios, test a response to any possible disruption within the factory or its supply chains, and optimize processes as a result. Businesses could also compare machinery performance across production lines, plants, and supply chains.

Investment possibilities

These possibilities do highlight another barrier to AI deployments, however—where should companies be focusing their investment? 

With the potential of AI so great, there can be competing interests within organizations. That’s why it’s so essential that the business goals are clear and that AI deployments align with them, as this can significantly impact investment decisions 

For example, if a company’s goal is to improve working conditions, they may want to deploy a health and safety solution. This could require additional investment in edge computing to lower latency and enable AI to step in instantly if imminent danger to workers is detected. 

Investment in edge solutions may also be needed if, for example, large amounts of data needs to be processed—and reliance on external telecommunication networks needs to be reduced, to improve reliability and keep costs down. 

Additional considerations

There are also security risks to consider. By making data more accessible, manufacturers will expand their connected devices, and increase the potential attack vectors. 

Cybersecurity should be a key priority for any business using AI. If successful, an attack has the potential to force the closure of a production line, at huge financial and reputational costs—especially if orders are missed.  

It’s vital, therefore, before any AI solutions are deployed, that manufacturers invest in good security practices. They will need to identify all connected assets, initiate live threat detection, and enable a holistic view of all security events—so investigation and remediation becomes easier. Manufacturers will also want to ensure they are isolating networks, using zones and conduits, to prevent any attacks from spreading. 

Finally, businesses also need to consider the impact on factory workers. This is likely to require an organizational change management program, with careful communication and training being provided.  

All these factors can slow down adoption, so it shouldn’t be so surprising that AI deployments in factory settings haven’t been as rapid as many may have expected. But this should not deter any organization from putting plans in place. 

If manufacturers invest in the fundamentals—with data accessibility, usability, reliability and security all addressed—they will put themselves in a strong position to seize the potential of AI. They will also have the best possible foundations in place to ensure deployments achieve their business goals faster.  

For deeper insights and actionable strategies, read Cognizant’s latest whitepaper, Engineering AI into the factory of the future.

About the Author:
Sharath Prasad leads Cognizant’s Smart Manufacturing initiatives across the Americas, helping industrial clients harness AI, IoT, and automation to drive operational resilience and digital transformation. With deep expertise in application software development, intelligent automation, and connected product ecosystems, he partners with global enterprises to improve uptime, accelerate time-to-market, and maximize ROI from technology investments. Sharath is based in Chicago, IL.

Subscribe to Industry Today