Why a Data Foundation is Essential for Success with Industrial AI
Guest Blogger: Jonathan Lang, Research Director, Worldwide IT/OT Convergence Strategies, IDC
Sponsored By: Sight Machine
Across all industries, AI is now the North Star for executives seeking to improve business performance. This technology promises to transform the fundamentals of nearly every process and even brings into question the role of human labor moving forward. Yet in some industries, such as manufacturing, the opportunity to effectively utilize AI to gain a competitive advantage is currently limited.
AI’s efficacy is dependent on its contextual awareness of the data it operates within. In industrial operations, data is difficult to source and even more difficult to contextualize from the raw data stores alone in their present form. AI requires transformation of the data including mapping and remapping of data within domain-specific schema to capitalize on its potential with the analytics of today.
For industrial enterprises to capitalize on AI, IT and line of business executives will need to get real about the organization’s state of data readiness and the steps needed to prepare for AI before an AI-based vision of the future can become reality. Fortunately, manufacturers and technology and service partners have worked for several years on developing and proving out the architectures and technology toolsets to develop a solid data foundation upon which all forms of analytics can achieve scale.
IDC has observed a significant market trend in which industrial organizations develop a data foundation or data fabric capability that enables data acquisition and ingestion from a diverse set of sources. They then add a data engineering capability for transforming and contextualizing this data and an industry-specific intelligence capability to store and apply analytics to it. The common elements of a data foundation can be observed in Figure 1.
Figure 1: The Building Blocks of a Data Foundation
Source: IDC Market Glance: Industrial Data Lifecycle Management, 1Q24, 2024
Today, everyone wants to talk about AI and its transformative impact on business now and in the future. For industrial enterprises seeking to capitalize on the opportunity new AI capabilities present, having a data foundation is essential for success. Industrial data is a mix of structured and unstructured data combining text-based and numerical inputs and requiring novel context unique to a given operation. To combine these elements requires both subject-matter expertise and purpose-built industrial data foundation offerings. If companies want to open a conversation about the use of AI in operations, they first need to open an honest dialogue with executives about the state of their data. IDC has witnessed many companies successfully launch and establish a data foundation based on executive interest and mandate for the use of AI. We therefore advise teams tasked with executing new technology initiatives in industrial operations to broach the conversation and secure executive buy-in for the data foundation in this way.
Benefits of a Data Foundation
A mix of new and existing use cases are enabled once industrial data is accessible, accurate, and actionable with a data foundation. First, long-time bread-and-butter use cases can be automated and enhanced through schema-based analytics that eliminate the need for business analysts to manually sort through different data sources. For example, companies can perform analytics to determine throughput bottlenecks and loss analysis. They can examine the root causes of quality issues and build greater flexibility in their processes without introducing production risks. They can improve sustainability and optimize their value chain despite external uncertainties from supply chain and external ecosystem partners.
But there are future capabilities that are becoming increasingly pressing. These will require robust data foundation capabilities as well. First, IDC believes these use cases to be integrated with one another and at broader systems or site level for analysis in the future. For example, on the horizon we see enterprise digital twins looming that extend beyond any individual asset or process and can conduct AI-based analysis at the system level. They will require all data to be accessible and contextualized by the AI.
Software-Defined Automation
Second, these systems-level AI capabilities will extend into asset and line optimization in real time through the next generation of closed-loop automation. IDC refers to this phenomenon as software-defined automation. As analytics become more accurate and real time, combining more inputs, organizations are examining ways that setpoint reconfigurations can be made autonomously as opposed to traditional methods of manual reprogramming.
If organizations want to succeed in building out data-driven operations capabilities, they need to start with their data. Industrial data carries such a unique set of considerations and challenges that a purpose-built industrial data fabric has proven itself to be the most effective way to cater to these requirements at scale. Whatever the next great innovation may be that will transform the way industrial companies do business, a data foundation is the key capability that will help organizations be positioned for a competitive advantage.
About the Author: Jonathan Lang is Research Director for IDC Industry Operations Insights responsible for the IT/OT Convergence Strategies practice. Mr. Lang’s research focuses on digital transformation strategies in environments where operations technologies are deployed including Manufacturing, Utilities, Oil & Gas and Healthcare Provider settings. As IT capabilities redefine and extend the core value drivers of operations technologies, Mr. Lang’s research examines strategies, roadmaps, and governance models to drive this convergence and manage the new data and processes it requires.
