ML Journal

Industrial DataOps enable manufacturers to be more agile, improve continuously, and move toward smart manufacturing. 

 

TAKEAWAYS:
● Industrial DataOps solutions enable real-time data flows across hardware and software.
● Manufacturers can use data connectivity to respond quickly to emerging risks and constraints across their facilities and the supply chain.
● Manufacturers need to focus on turning raw data into useful information to drive decision-making.   

 

 

Most manufacturers are on a technology journey, increasingly adopting automation, analytics, and active sensors. In the next two years, 41 percent of manufacturers plan to prioritize investments in factory automation hardware, and 40 percent plan to invest in analytics, according to Deloitte’s 2025 Smart Manufacturing and Operations survey.

The common denominator for these and other technology investments is data. As manufacturers advance, there is a fast-growing imperative to establish foundational capabilities in data quality, connectivity, and access. One effective approach to achieving this is through Industrial DataOps.

Many enterprises are currently working to leverage a unified namespace (UNS), an architectural strategy that centralizes data sources into a single source of truth. A UNS provides a real-time view of the business, and Industrial DataOps facilitates this process at scale. As data silos diminish through a UNS, Industrial DataOps solutions enable real-time data flows across hardware and software, whether at the edge, on-premise, or in the cloud. This robust data foundation paves the way for enhanced agility in the face of constraints, fosters new insights for continuous improvement, and supports the transformation to smart manufacturing.

In adopting an Industrial DataOps strategy, manufacturers should consider five essential questions:

1. What criteria should shape our data product priorities for long-term impact?
2. How can we foster local innovation without sacrificing global data standards for creating silos?
3. How does talent and structure limit our ability to turn data into actionable information, and how can we bridge those gaps?
4. What cultural and operational shifts are needed to embed and scale DataOps sustainably?
5. How can we design our DataOps strategy to support future AI needs while managing data quality and governance?

1.   Shaping Data Priorities for Long-Term Impact

With DataOps, manufacturers convert centralized information into actionable insights. Manufacturers grapple with large, complex data flows from IT and operational technology (OT) infrastructure. With numerous opportunities available, it can be challenging to determine where to begin.

To achieve widespread DataOps use, a data product strategy can be implemented incrementally, capturing value along the way. Initially, manufacturers might focus on the factory floor, developing data products related to bills of materials, parts and inventory, and work instructions. These early solutions demonstrate value, foster relationships and repeatability, and lay the groundwork for more ambitious DataOps applications.

“To achieve widespread DataOps use, a data product strategy can be implemented incrementally, capturing value along the way.”

 

Once progress is made at the factory level, manufacturers can broaden their focus to data products that span facilities. These might include supply chain data products to clarify incoming supply and demand, products for product lifecycle management (PLM) as it relates to engineering, products for orders and customer demand, and downstream products for quality control and aftermarket service.

Ultimately, manufacturers should start by identifying specific business problems that data products can resolve while also envisioning how to maximize value delivery at scale.

2.  Foster Local Innovation without Sacrificing Standards

The emergence of generative AI (GenAI) has sparked a wave of experimentation across various levels, from the factory to the enterprise. New proofs of concept (POCs) have emerged in many areas, raising challenges around AI governance and the coordination of experiments within a broader, strategic framework. Similarly, DataOps presents the challenge of balancing innovation with control. Organizations must grant teams a degree of freedom to experiment while also ensuring adherence to standards critical for maintaining interoperability, scalability, and data quality.

Discovering where DataOps can unlock visibility may be best approached with both top-down and bottom-up strategies for identifying opportunities. For example, building data products for PLM is a global initiative, as products should be consistent across the manufacturing footprint.

Conversely, when it comes to OT data within the factory walls (e.g., from SCADA systems), manufacturers can manage data products at the facility level while leveraging tools and frameworks defined at the enterprise level. This approach allows for innovation within facilities that is permitted, governed, and strategically aligned.

3.  Bridging Gaps and Turning Data into Action

Implementing DataOps requires specific technical skillsets. Manufacturers will need data engineering talent to plan and implement a UNS, as well as data architects to manage data structure. Cloud computing skills are essential for setting up platforms and infrastructure to capture the right data (rather than all data).

“Some manufacturers may pursue building services and solutions based on their data products, making the development of DataOps a core competency.”

 

However, merely collecting data is not sufficient. Manufacturers must also focus on transforming raw data into meaningful information that can drive decision-making and insights. This requires data owners within the business who are accountable for managing and maintaining data quality, context, and relevance. In the long term, as the data foundation supports increasingly ambitious automation, there will likely be a heightened demand for skilled talent in data science, machine learning, and AI modeling.

To attract top candidates, manufacturers may need to increase pay rates and ensure their global workforce strategy is guided by leaders with robust skills and experience in data and technology.

4.  Cultural and Operational Shifts for DataOps

Even with the right skilled labor, adopting a UNS and DataOps can be challenging. Manufacturers will require an ecosystem of partners and providers with the necessary tools and capabilities to support their journey. Some manufacturers may pursue building services and solutions based on their data products, making the development of DataOps a core competency.

For most manufacturers, however, this domain is often deprioritized because, while it is foundational, it typically does not serve as an immediate performance differentiator. Achieving success requires a balance between daily operations, limited talent and funding, and the complexities of data product development and management.

Manufacturers must make important decisions regarding where to deploy, when to scale, how much control to retain or delegate, and how to coordinate within the strategic ecosystem. Engaging partners to manage DataOps on behalf of the manufacturer may offer an expedient, cost-effective way to access the right talent and competencies.

5. DataOps Strategy Design for Future AI Needs

AI relies on high-quality inputs, and a UNS with well-designed data products transforms raw data into assets ready for AI training and deployment. The ability to manage data in a repeatable, cost-effective way is essential for realizing AI’s full potential value. For instance, if a manufacturer seeks to deploy a sensing system on the factory floor to detect and correct human error, such a deployment requires the right data products to ensure repeatability.

“Investing in DataOps enables manufacturers to build predictable data sets where scale is known, and data are governed and owned.”

 

Some manufacturers have explored limited AI deployments that are relatively straightforward. For example, a GenAI-enabled chatbot navigating work instructions is a deployment that requires minimal DataOps. However, an ambitious smart manufacturing vision does require DataOps, and as Deloitte research shows, nearly all (92 percent) manufacturers view smart manufacturing as the primary driver of competitiveness over the next three years. From this perspective, DataOps is a competitive imperative.

Investing in DataOps enables manufacturers to build predictable data sets where scale is known, and data are governed and owned. This level of data maturity is essential for the future of smart manufacturing. Longer term, DataOps serves as the gateway to developing fully autonomous agentic systems and unlocking the greatest rewards from software defined manufacturing.

Compounding Outcomes with Industrial DataOps

Industrial DataOps will become increasingly important as more devices and factories come online. Connectivity and real-time data access allow manufacturers to extract more value from their data today, with even greater longer-term potential.

Data connectivity and access help manufacturers understand and quickly respond to emerging risks and constraints across facilities, the supply chain, and the marketplace. For continuous improvement, DataOps highlights areas where plans and operations can be enhanced or adjusted. Additionally, data quality and a foundation for AI use cases can help mitigate skilled labor shortages and attract new talent to the industry.

Data are a strategic asset, and like any enterprise asset, it must be leveraged to achieve business goals. Industrial DataOps serves as a vital bridge between the rapidly growing data flows of today and ambitious smart manufacturing visions of tomorrow. M

About the authors:

 

Patricia Henderson is a principal in AI & Data with Deloitte’s Industrial Products & Construction Practice.

 

 

 

Rohini Prasad is a manager in Deloitte’s Strategy and Operations Group with Supply Chain & Network Operations.