Skip to main content
Skip to footer

Naveen Mehata Kondamudi

The tremendous increase in processing power and decrease in data storage costs in the last decade have fueled the growth of AI in real-world applications. A Gartner survey revealed that 66% of organizations increased or did not change AI investments since the onset of COVID-19, while a McKinsey report predicted that AI can deliver additional global economic activity of around $13 trillion by 2030.

AI, which forms the key technology backbone of manufacturing organizations, has evolved so much that it builds end-to-end intelligence in the value chains of firms. In product innovation, for example, AI supports model-based product development, digital twin, and seamless integration and collaboration of systems across the product value chain. AI not only provides manufacturing organizations the much required agility and resilience, it also enriches the user experience in product development and helps in supervised autonomous decision making. AI can be effectively deployed across multiple functions such as:

• Warranty claims analysis

• Sales hit ratio analysis

• Regulation impact analysis

• Cross-functional process optimization

• Organizational change sentiment analysis

• Defect root cause analysis (RCA)

But there are some prerequisites to reap the benefits of this technology.

Data forms the foundation of any AI initiative. The quality and quantity of the training data determines the effectiveness of the AI models. Today, organizations generate significant amounts of data, which is key to making manufacturing agile, responsive, and adaptable. But data is not AI-friendly, as it is disconnected, not harmonized, or validated. Typically, the input data for any AI model needs to flow through these necessary but time-consuming preparatory steps:

  1. Collection

  2. Cleaning

  3. Correction

  4. Conversion

  5. Contextualization

Apart from quality and quantity, clean data should precisely reflect the broad spectrum of possibilities across the entire value chain to make AI models robust.

What is required is a continuous flow of digital data across the ecosystem, transforming, evolving, and maturing over time. Such data will enable AI models to identify and learn patterns, which would help enterprises solve problems in a holistic manner. This interconnected digital data spread across the value chain providing inputs for analytics and AI, are what we would call digital thread for an organization. AI and data are also core to Neural ManufacturingTM, an approach for manufacturers and their ecosystems to become connected, collaborative, and cognitive. It drives agility and exponential growth, providing firms an intelligent edge.

Digital thread in manufacturing organizations

The concept of digital thread means connecting people, processes, assets, and their characteristics and dependent data throughout the value chain of the organization. Typically, in manufacturing, this information is managed in silos across multiple enterprise-wide systems such as product life cycle management (PLM), enterprise resource planning (ERP), supply chain management (SCM), and more. Digital thread forms the backbone of these systems, connecting and integrating them, which generate clean and AI-friendly data. But digital thread is not limited to systems integration. Listed below are the key building blocks of a digital thread and how they impact data quality:

  1. Model-based systems engineering: Data in common language for easy interoperability across all the systems involved in the product life cycle

  2. Closed-loop processes: Data improved with continuous feedback validation

  3. Co-innovation platform: Harmonized, governed, and readily available data which provides a holistic view to the research and development (R&D) department to improve products

  4. Analytics-based reporting: Easily consumable data with pre-mapped relationships

  5. End-to-end integration: Connected and continuous data flowing seamlessly across the PLM, ERP, and SCM systems

Overall, digital thread provides AI applications with a data lake with clean feed for ready consumption. In fact, depending on the application and industry, manufacturers can choose one critical block at a time using an agile approach to reap quick benefits. Many narrow AI tools can help firms in their digital transformation journeys, which would trigger a virtuous cycle - simple AI systems that help enterprise systems generate superior quality data for advanced AI systems that can provide insights, offering enormous value in return.

According to the Gartner Hype Cycle for Artificial Intelligence, 2020, AI industrialization and AI democratization are two megatrends that provide late AI adopters a level playing field by making AI tools more affordable and easily available. This has enabled many organizations to leverage their AI capabilities to build machine-first operating models, giving them a competitive edge and also enhancing the value of the key differentiator – data, now more valuable than oil or gold. Organizations that resist adopting data-centric digital transformation will lose their competitive edge, putting them on the path to obsolescence. So, a digital thread is critical for organizations, especially now, when digitalization is the new normal.

About the author

Naveen Mehata Kondamudi
Naveen Mehata Kondamudi is a consultant with the Innovation and Transformation group of TCS’ Manufacturing business unit. He has more than 16 years of experience working with global manufacturing customers in the areas of product life cycle management and supply chain management. Naveen holds a Six Sigma Green belt and has a keen interest in AI trends. He is a certified project management professional (PMP®) and SAFe® Agilist. He holds a Bachelor’s degree in Mechanical Engineering from the College of Engineering, Guindy, Anna University, Chennai, India.
Contact Contact