Future-proofing manufacturing with data

Future-proofing manufacturing with data
Traditional manufacturers are struggling to optimise productivity. How can data help companies transition to smart manufacturing?

Mediocre decisions in a manual input world 

Henry Ford’s moving assembly line single-handedly revolutionised the auto industry and led to a drop in the number of car manufacturers from 253 to 44 by 1929. Factories today are achieving less than 50 percent of their full production capacity due to the dependence on manual inputs in the areas of inventory control, pricing reports and production planning. Early adopters are already planning for next-generation manufacturing plans with real-time, data-driven trend analysis and decision-making. 

The Industrial Internet of Things uses the Internet of Things to harness sensor data, automate data inputs, and use smart machines to capture, analyse and communicate optimal solutions to manufacturing issues. The efficiency of smart manufacturing depends on the reduction of time taken to make supply decisions, as well as the increased accuracy and volume of historical data. Traditional manufacturers use human inputs via excel and other tools, to compare historical data against present-day performance, and then decide how to proceed. Data gaps, human error, and the slower decision-making process involved in traditional planning can actually impact the accuracy and robust planning capacity of a plant.

Different strokes for different folks

India’s major industries are still lagging in a data-driven approach to production as they lack guidance on technical infrastructure, capital and training resources required to bring their capacities up to standard. Baosteel has already been able to mitigate the supply chain impacts of COVID-19 quarantine protocols through an intelligent remote control system. The automated manufacturing process enhanced internal safety measures for employees as well as stabilised the supply of goods, without compromising on their quality and efficiency standards. Precision Farming uses sensor technology to track environmental and soil conditions, automation to optimise farming processes, and streamlines farming outputs through satellite data and software solutions. The textile and paper industries can utilise smart fabrics to increase the lifespan of products, reduce consumer waste, and provide a feedback loop to manufacturers based on environmental sensor responses by products already in the market. 

Benchmarking to ease SMEs into IR4

The fourth industrial revolution (IR4) has five key features: optimisation and customisation, adaptability and automation, a human-machine interaction platform, automated visualisation and communication, and value-added business and services. 

The industrial revolution is obviously difficult but not impossible. The process will require an incremental approach with clear guidelines, quality checks and a visible, achievable benchmark. Through the training of the human resource to optimise data inputs, acquisition of technologies to analyse data, and the supervision of automated decision-making, smart manufacturing can help stabilise economies and make them resilient to unforeseen disruptions. 

The Indian government may not be able to consolidate every aspect of this change under general manufacturing reform, so it is up to industry leaders within sectors to set benchmarks and internal guidelines. They alone would have the initial rush of technical and financial capital, along with the ability to streamline the change management process, to develop the right solutions for their own businesses. Once proven to be successful, sector-wise guidelines can be prepared by domestic ministries or certifying bodies like the International Organization for Standardization (ISO) for small and mid-size enterprises (SMEs) to develop smart technology at their own scale and pace, with minimal operational disruptions. 

Are we giving up human intuition in exchange for autonomous talking appliances?

Machine learning is an algorithmic framework of data analysis and prediction, which is usually recommended to be used with an 80-20 approach. However, with artificial intelligence (AI) driving backend outputs and slowly replacing manual decision-making, the risk of losing that “human touch” grows. For example in the mortgage sector, a 100% AI-driven loan approval software could negatively impact first time buyers who have community/family support in their repayment plan, which is not accounted for in the platform algorithm. A local financier, on the other hand, would be able to gauge the community's impact through personal experience and understanding.


  • If large scale corporations monopolise sectors through smart manufacturing, late adopters may never get the chance to catch up. Governments must encourage industry leaders to light the way and set benchmarks for smart manufacturing in their respective industries. This can then be scaled down for SMEs through regulatory entities run by the government or third-party certifying bodies. 

  • The lack of depth in the transparency of decision-making by both AI and the human brain makes it difficult to propose a complete removal of human intervention between analysis and action.  Human involvement in decision-making should still be a significant factor in supplementing nuance and human intuition to the logical solution provided by a data-driven output. 

  • Prior to the development and funding of smart technologies, companies must try to optimise their historical data inputs. By studying industry leaders and global supply chains, they can curate and control data entry points to maximise the accuracy, depth and breadth of data collection. Eventually, when technological frameworks replace existing manual processes, the accuracy and scope of the historical data collected until that point will determine the efficiency of the smart manufacturer.