Applying Neural Concepts and Networks in Manufacturing

By Ralph Rio

Technology Trends

Neural networks with multi-node connectivity are gathering interest and adoption in several aspects of manufacturing, particularly for supply chains.  Neural networking enables a collaborative, resilient and adaptive ecosystem that can help manufacturing enterprises succeed in today’s dynamic business environment.  This was the crux of the discussion between ARC’s Ralph Rio, Vice President, Enterprise Solutions, and Sreenivasa Chakravarti, Vice President, Manufacturing Business Group, at Tata Consulting Services (TCS).  Speaking about his role at TCS, Sreenivasa explained that his consulting practice works at the intersection of manufacturing industries and their ecosystems; the evolution of value chains; and the impact of emerging technologies in improving the operations.  You can listen to the entire podcast here.

Applying Neural Concepts in Manufacturing

Responding to a query on the neural concept in manufacturing, Sreenivasa said that neural networks have been in existence for a long time, and adoption has accelerated with more powerful artificial intelligence (AI) and the need to manage complex extended supply chains.  Giving the example of a global manufacturing company with a very long value chain, he explained how it connects tiered suppliers, intermediaries, and the end customer across geographical locations.  Sometimes, there are delays in the response times/signals, and these can be identified and overcome by neural networks.  Inspired by the human biological neural brain, TCS drew parallels.  Both require the ability to sense, perceive and act.  The manufacturing ecosystem must have the ability to sense changes in the market/supply side and respond with agility.  Other capabilities in the biological system, such as distributed decision making and a central command structure, can be applied in the manufacturing system too.  Further, he said that the neural concept “will have a very large role to play in how organizations and the value chains are shaped in the future.”  TCS defines neural manufacturing as an intensely networked set of partners, which is aligned to one common purpose where the value chains are very responsive, adaptive, and even personalized, and the intelligence is built throughout the network and into the edge.

Ralph opined that applying neural concepts to manufacturing was very innovative, but how did it benefit the manufacturers?

How Applying Neural Concepts Benefit Manufacturers

Sreenivasa said that the benefit to manufacturers was the prime consideration.  Today, the ecosystem has changed with smarter, connected, and software-embedded products.  There’s also a significant electronic component in products like automotive, airplane or smart farm equipment that is already “smart”, and the future demands that they become smarter.

Ralph said that he was thinking about the supply chain in a linear fashion and this idea that the components in the supply chain are intelligent, smart and can add data to facilitate decision making, is thought-provoking.  “Could you tell me about the role neural is taking in shaping a manufacturer's ecosystem of partners?”

Partner Ecosystem

The manufacturer’s partner ecosystem is very important as customers are moving beyond just buying a product.  They are often looking at buying an asset-as-a-service with supplier support throughout its lifecycle.  For example, when you buy a car, besides insurance you might expect charging facilities (for an EV) or continuous health monitoring.  These services are not only provided by the car manufacturer.  They include an ecosystem of partners who need to share information to provide a great customer experience.  “And that really becomes the differentiator,” explained Sreenivasa.  All the data points (customer behavior, location, habits, and expectations) and the product (design, knowledge, and information) must be interfaced in a much more agile, self-driven model where these nodes connect with each other and help drive the end objective.  Sreenivasa said that the healthcare business is not just about supplying medicines; it is about caring for the end customers by proactively ensuring that they maintain good health by creating a wellness ecosystem.  Here, the neural element plays a very critical part.

Next, Ralph asked how neural concepts help manufacturers become more resilient and adaptable.

Neural Concepts Help Build Resilience and Adaptability

TCS believes that emerging enterprises need to have three distinct capabilities:

  • Resilience – the connected, cognitive, and collaborative value chain and its ability to spring back after a disruption
  • Adaptability – shuffle the product mix, product functions or capabilities for an agile and flexible response to market demand
  • Purpose driven - a partner ecosystem that delivers a complete product including services to the end customer

Adopting Neural Concepts for Sustainable Competitiveness

“How and where would a manufacturer start adopting neural concepts?” asked Ralph.  It’s a strategic  business decision like any other, explained Sreenivasa.  The starting point would be the organization’s focus area, like using data to differentiate as the business operating model.  A maturity assessment of the enterprise and partner ecosystem provides a start to gauge various aspects, such as if the products are smart enough, suppliers are prepared for demand variations etc.  Besides the technology aspects, the risk propensity of the enterprise also needs to be considered.  He said, “Because there's a lot intelligence on the edge, you should be prepared to delegate decision making to these partners and keep monitoring it in a way that you continue to head in the direction of your competitive advantage.” He gave examples of neural capabilities, chosen based on specific focus areas for sustainable competitiveness, such as cost or product differentiation strategy.  These were substantiated by interesting case studies.  

In conclusion he said that neural networks will be game changers since they enhance efficiency and ensure faster market response, which provides a competitive edge.

Case Studies

Premium Product

Sreenivasa gave the example of a premium market product for which customers continuously seek more advanced features, and given the hyper competitiveness in the marketplace, these features keep changing.  But in the traditional model of the supply chain, the suppliers are building up capacity based on a plan and making investments.  However, due to the rapidly changing demand, suppliers have no clue as to what has changed on the front end that could be disastrous in terms of investments made and the end product.  If supply chains are neural networked to the planning and supplier systems, then the supplier is enabled to sense and respond to the change.  This is the cognitive and connected value chain.

Business Operating Model

This client had challenges in trying to meet and track the entire order patterns.  There were order failures because of the long value chain.  After application of some neural capabilities, they were able to sense if a particular order would go through or not and the reasons for those decisions.  The client addressed these problems and saw significant business benefits.

A podcast of the interview is available at: 

Engage with ARC Advisory Group