The data readiness gap: Why manufacturers struggle to capture the value of AI | Velotic

Introducing the Industrial AI Value Cube – a practical framework for moving beyond pilots to measurable value.  Explore the framework

The data readiness gap: Why manufacturers struggle to capture the value of AI

Why do manufacturers struggle to get value from AI? Explore the data readiness challenges holding them back and how operational context supports better decisions.

Blog 10 min read

Expert perspective: Paul Adams

Paul Adams, Director of Partner Programs, Americas at Velotic, works closely with system integrators to understand the challenges manufacturers face. In this three-part series, he shares insights from conversations with four industry partners about data readiness, manufacturing modernization, and the evolving role of system integrators.

Manufacturers Don't Have an AI Problem. They Have a Data Readiness Problem.

Manufacturers don't have an AI problem. They have a data readiness problem.

As the leader of our System Integrator (SI) program, I regularly engage with partners across the manufacturing ecosystem. Over the past year, one topic has dominated nearly every industry conversation: artificial intelligence.

Manufacturers are exploring AI's potential to improve productivity, reduce downtime, optimize maintenance, enhance quality, and accelerate decision-making. Yet despite the excitement surrounding AI, many organizations are discovering a more fundamental challenge. Before manufacturers can successfully leverage AI, advanced analytics, or broader digital transformation initiatives, they need a strong operational data foundation.

To understand what manufacturers are experiencing, I spoke with four system integration partners: John Leppiaho of Leveraging Manufacturing Solutions (LMS), Andy Corsini of E Tech Group, Justin Vaught of RoviSys, and Tod Virden of Gray Matter Systems.

Why Manufacturers Struggle to Get Value From Their Data

Why manufacturers struggle to get value from their data

Over the past decade, manufacturers have invested heavily in modernization initiatives, yet many fail to deliver their expected outcomes. The reason is often not a lack of technology or ambition. It's a lack of data readiness.

Manufacturers are generating more data than ever before, but that data frequently remains fragmented across equipment, facilities, software platforms, and business functions. Tod Virden believes many manufacturers continue to struggle with a familiar obstacle. "Data silos are still one of the biggest challenges we encounter. Manufacturers often have many disconnected systems, each with its own database structure and way of organizing information."

As a result, organizations find themselves asking critical business questions:

  • What is causing downtime?
  • Which production lines are performing best?
  • Where are quality losses occurring?
  • How do we compare performance across facilities?
  • Which assets require attention before failure occurs?

The answers often exist somewhere within the organization. The challenge is that the data required to answer those questions is frequently disconnected, inconsistent, or inaccessible.

The New Challenge Isn't Collecting Data

The new challenge isn't collecting data

Collecting data is no longer the hard part. As John Leppiaho explains: "The biggest challenge isn't collecting operational data, it's turning that data into actionable information. Most manufacturers already have enormous amounts of data across PLCs, Historians, ERP, quality systems, and other applications."

Justin Vaught believes manufacturers often overlook another important consideration. "I think the biggest challenge is less about access to data and more about understanding what to do with it."

According to Vaught, many organizations focus on collecting more information or moving data into common platforms before clearly defining how that data will improve a business process. "Without that connection between the data and the business process, you end up with a lot of information but not necessarily better decisions."

Andy Corsini sees the same issue across many manufacturing environments. "The biggest operational data challenge is that there is a lot of data available, but it is typically spread out across multiple systems and owned by different groups. Turning that data into something people can trust and act upon has become the main goal."

The Missing Piece: Human Context

The missing piece: Human context

Without the right context, operational data can be difficult to interpret, trust, and use effectively. Leppiaho believes the next level of value comes from combining machine-generated data with human insight.

"The operator is also an important source of operational data. The challenge is making it easy for operators to quickly add the human context that automation and sensors can't provide. When machine data and operator knowledge come together, the data becomes much more valuable and actionable."

This is where Manufacturing Execution Systems (MES) often play a critical role, connecting operational events, workflows, quality processes, and human interactions into a unified operational narrative. By bringing machine data and operator knowledge together, manufacturers can gain a more complete understanding of their operations and make better-informed decisions.

Why AI Raises the Stakes

Why AI raises the stakes

In many ways, AI is forcing manufacturers to confront data challenges that have existed for years. Organizations can invest in advanced AI models and sophisticated applications, but if the underlying data infrastructure is inconsistent, incomplete, or unreliable, the quality of the output will suffer.

Vaught believes AI is simply the latest example of a longstanding challenge. "AI is obviously a major part of the conversation today, but I think we sometimes jump straight to AI without understanding the importance of proper data contextualization and process mapping."

He notes that previous technology waves often promised similar shortcuts. "Five years ago, IIoT and putting everything in the cloud were often presented as the answer. Today, a Unified Namespace is commonly touted as the solution."

While these technologies can deliver meaningful value, none eliminate the need to understand how information will be used to improve business processes. "You can build a very sophisticated technical architecture without creating much operational value."

Shifting the Focus From Technology to Outcomes

Shifting the focus from technology to outcomes

Virden believes recent AI initiatives are accelerating efforts to connect previously isolated systems. "More clients are connecting siloed systems, standardizing on reporting platforms, and collecting more information than before to solve a broader range of business challenges."

Corsini has seen a similar shift in customer priorities. "Over the last five years, the conversation has shifted from collecting and storing data to achieving desired outcomes that will help move the business forward."

The manufacturers that realize the greatest value from AI will be those that have already invested in creating connected, governed, and trusted operational data environments. Before organizations can scale AI successfully, they must first trust the information feeding it.

Continue Reading: Why Modernization Initiatives Stall

Continue reading: Why modernization initiatives stall

Data readiness is only part of the challenge. Even when manufacturers have the right technology and information, modernization initiatives can struggle to deliver lasting value. In the next article in this series, we explore why manufacturing modernization projects stall and how organizations can better align technology, processes, and people to achieve meaningful outcomes.

 

With thanks to our system integrator partners

Build the foundation for industrial AI

Data readiness is one of the keys to turning AI investment into operational value. Explore Velotic CEO Brian Shepherd's practical framework for assessing readiness, prioritizing opportunities, and planning for scale.

Explore industrial AI

Continue reading