Breaking Down Silos: AI and Digital Twins in Modern Manufacturing
M. Ahmad S.
CTO | Chief Architect | CDO | Data and AI Strategist | Technology Leader | D.Eng Candidate in AI/ML
Manufacturing, like most legacy operations, is generally organized in a departmental functional groups like engineering, manufacturing, R&D, sales, distribution, supply chain, and after-market support. While this structure allows organizations to concentrate expertise and best practices, it often results in silos, creating boundaries and barriers to data sharing and end-to-end operational analytics. In linear operations, decision-making is frequently compromised by what could be called LIOH syndrome—Lack of Information on Hand. This becomes especially evident in R&D and New Product Introduction (NPI) operations, where teams strive to incorporate new features or develop entirely new product lines but often lack comprehensive enterprise data.
During NPI gates, engineers assess the impact of their design changes, but they do so with limited access to enterprise-wide information. In manufacturing, the true impact of a design on production often only becomes fully apparent at the engineering-to-release stage, and the effects on after-market performance may not be understood until the product has been in the field for some time. This delay in information flow across the product lifecycle hampers decision-making and stifles innovation.
The Promise of Digital Twins and AI
The concept of a Digital Twin—a computerized replica of a physical asset—offers a promising solution to these challenges. If integrated with comprehensive data and analytics, including AI-powered predictive models and simulations, Digital Twins could provide a real-time reflection of the physical asset, offering significant advantages in terms of prototyping, simulation, and design impact analysis. Generative AI (Gen AI) could be used to simulate and generate possible design outcomes, improving the prototyping phase and enabling more effective decision-making. However, for Digital Twins to be effective, a seamless integration of data across various systems is critical. Without access to key data from procurement, manufacturing, parts replenishment, and social media, the full impact of design changes cannot be fully understood.
Manufacturers face the challenge of working with disparate legacy systems—Product Lifecycle Management (PLM), Enterprise Resource Planning (ERP), and Customer Relationship Management (CRM) systems are often siloed and not integrated. Many companies operate multiple instances of these systems, making the vision of a true Digital Twin complex and difficult to realize. To unlock the potential of Digital Twins, organizations must overcome these integration hurdles, leveraging AI to intelligently bridge the gaps between these disparate systems.
The Road to Data Integration, Governance, and AI Adoption
The solution to this problem is straightforward but requires significant effort: organizations need a top-level commitment to data integration, governance, and the adoption of AI technologies. Only with strong leadership and the right infrastructure can manufacturers begin to effectively leverage the power of Digital Twins, AI, and analytics across the product lifecycle.
The transformation requires companies to move away from the traditional, fragmented approach to data management and adopt an enterprise-wide strategy that prioritizes data quality, governance, integration, and the deployment of AI-driven insights. AI can automate the integration process, uncover patterns in large datasets, and predict design or production issues before they arise. This includes implementing secure, cloud-based solutions that can handle the sporadic nature of NPI processes while embedding AI-driven analytics as a core component of R&D and engineering workflows.
Learning from Consumer Electronics: The Case of Wearables
To illustrate the importance of ecosystem readiness for new product adoption, let’s look at the consumer electronics market. When new in-home connectivity devices were launched, many were skeptical about their success. Similar doubts surrounded early smart wearable devices, which seemed innovative but ultimately failed to gain mass market traction. Despite the promise of merging the physical and virtual worlds through augmented reality (AR), these devices failed to capture widespread consumer interest. Analysts have pointed to reasons such as product flaws, privacy concerns, and health risks, but the real issue was a lack of clear purpose and ecosystem readiness. Was the device intended as a fashion accessory, or as a tool for specific tasks, such as maintenance and repairs?
In contrast, successful AR products like Pokémon Go had a clear purpose and an easily adoptable platform. Their success came from meeting consumer needs and aligning with the market ecosystem. Similarly, early smart devices like the first smartphones faced similar challenges, as the market wasn’t ready for them, despite the innovative technology they offered.
New Product Introduction: Cognitive and AI-Driven Approach
Launching a new product is never straightforward. Companies often find themselves in a race to create the next best thing, pushing innovative features to stay ahead in a rapidly changing market. A successful new product introduction (NPI) hinges not just on consumer excitement, but also on careful analysis of factors like aftermarket support, beta launches, and environmental/social impacts. AI can play a critical role in analyzing these factors, forecasting product performance, and providing data-driven insights to optimize the NPI process.
The reality is that NPI is no longer just an engineering function—it requires deep integration with data analytics and AI. Many companies overlook the importance of embedding AI into NPI processes. This is where cognitive cloud-based solutions—platforms with embedded analytics, AI, and generative capabilities—can make a significant difference. These solutions are well-suited to handle the irregular nature of NPI processes and offer companies the flexibility they need to manage data, generate insights, and visualize information effectively.
Key Features for an AI-Powered NPI Solution:
Embedding AI into NPI is a challenging but necessary task. Many manufacturers struggle to find the right balance of investing in systems and infrastructure to support this transformation. While cloud adoption may be slow in some industries, leveraging secure, AI-enabled cloud platforms with strong data governance and security frameworks can significantly ease the transition. These platforms offer robust data governance, security, and AI capabilities, providing manufacturers with the tools they need to integrate data across all stages of the product lifecycle.
Conclusion
The path to effectively using Digital Twins, AI, and generative analytics in NPI processes is not easy, but it is necessary for manufacturing companies to remain competitive. As industries face growing pressure to innovate faster and more efficiently, the integration of data, AI-driven insights, and governance frameworks will be key to success. Companies that embrace this shift will not only enhance their NPI processes but will also unlock the full potential of their product development and lifecycle management efforts.
data architect at self employeed
1moM. Ahmad S., Is merely adopting AI and generative models enough, or do manufacturers need a foundational overhaul of data infrastructure to fully realize these benefits? Can companies achieve operational agility without rethinking their entire approach to data and organizational silos?