The COVID-19 pandemic caused a chain reaction in the global supply of industrial products. A severe shortage of semiconductor chips resulted in automotive plants shutting down and scaling back production output. The auto industry is estimated to have suffered a revenue loss of USD 200 billion in 2021.
At the same time, climate change has put sustainability at the heart of the manufacturing agenda. During the lockdown in 2020, restrictions on the movement of people and goods reduced carbon emissions from road transport by ~10% and aviation by 40%, according to researchers at the University of East Anglia, University of Exeter, and the Global Carbon Project.
The pandemic has thereby shone a light on two business imperatives: sustainability and supply chain resilience. The manufacturing industry needs to maintain the reduction in emissions by altering traditional ways of working as well as minimising the carbon-intensity of the value chain. Business needs to rethink and set time-bound targets for reducing carbon emissions. Emerging technologies offer smart solutions to reboot manufacturing in the post-Covid era. Tiny AI, the smaller version Artificial Intelligence (AI), and predictive analytics are niche technologies that drive net-zero journeys by designing a collaborative ecosystem for manufacturing operations and supply chain management.
Driving sustainable operations
Manufacturers are prioritising sustainability, as evident from the pledges of business leaders at the COP26 Conference in Glasgow. Automobile, aircraft, and industrial manufacturers are adopting measures to reduce energy and resource-intensity of processes. AI empowers enterprises to optimally design and operate a green ecosystem, decarbonise operations, and eventually manufacture eco-friendly products. However, the humongous computational resources required for training and operating AI models increase the carbon footprint of an enterprise. To put it in perspective, the carbon dioxide equivalent (CO2e) emissions for training an AI model can exceed the average lifetime carbon emissions of a car.
Tiny AI enables manufacturers to achieve the goal of sustainable development by rationalising the carbon cost of AI, while using AI solutions for informed decision making. They do so by undertaking data processing at the extreme edge, which minimises energy requirement while enhancing the performance of applications. This enables manufacturers to boost productivity and safety by deploying a bouquet of Industry 4.0 applications. These smaller algorithms process petabytes of data to realize higher levels of accuracy and performance, but with significantly less computational power. Hence, Tiny AI solutions not only enhance energy and resource efficiency of manufacturing processes, but also minimise the carbon footprint of AI itself.
Decentralized data processing and analysis by Tiny AI drives sustainability by streamlining functions – from product design to quality control and maintenance and repair. Tiny AI algorithms combine real-time data and historical patterns to share actionable insights, be it to minimise resource consumption, eliminate defects, improve recycling of goods, or maximize the lifespan of assets. For instance, in aviation, AI-driven predictive maintenance plans prevent unplanned downtime based on visibility into the condition of diverse components of an aircraft.
Ensuring supply chain resilience
Supply chain constraints hamper industrial growth. Technologies such as Tiny AI, machine learning and predictive analytics capitalise on data to connect diverse parts of the supply chain. This integration helps decision making and boosts efficiency. An efficient supply chain drives synchronised manufacturing where sales and distribution functions are aligned with demand-supply dynamics. Synchronisation of production, procurement, warehousing, inventory management, and logistics enables manufacturers to become more responsive to customers / dealers, fulfil commitments on time, and deepen customer relationships.
Predictive analytics facilitates advanced functional alignment through adaptive processing of multi-dimensional data from distributed streams. Moreover, predictive analytics allows manufacturers to evaluate, benchmark, and redesign the supply chain. It minimizes risks in sourcing and delivery of parts and products. In addition, it helps procurement, logistics and inventory managers take data-driven decisions to meet production targets while rationalizing supply chain costs. Also, real-time supply chain visibility ensures smooth forward and reverses material flows, which supports product replacement, recycling, and dynamic pricing.
Analytics is the tool that mines real-time supply chain insights from field data, to identify and address constraints by bridging systemic gaps. Further, it predicts events that could affect production. In such cases, predictive analytics also foresees trends, such as the failure of a product, demand for a new product or the sale of spares. The ability to predict key supply chain metrics enhances customer service via timely order fulfilment and resolution of issues. For instance, accurate demand forecasting helps an appliance manufacturer align its production plan, inventory at dealer location, and logistics to address dynamic changes in consumer behaviour. On the inventory and logistics management side, analytical tools ensure adequate stock of fast-moving spare parts and accessories at distribution centres to serve customers on time every time.
I envision frugal and power-efficient computing resources shaping a new era in manufacturing. Green IT solutions combined with machine intelligence will reboot every aspect of the manufacturing value chain.
(Jasmeet Singh is the Executive Vice President and Global Head of Manufacturing at Infosys)