Nasscom Community

Supply Chain Resilience – Building a Risk-Optimized Sourcing Strategy

3 Mins read

COVID-19 has rattled established global supply chains. While impacted companies grasped the shock and realized the need for an alternative sourcing strategy, they also realized that global supply chains were not truly “global.” Offshored supply chains leveraging cost arbitrage became a trend when China ramped up its manufacturing scale, diversity, and speed in an unprecedented manner, starting 2001. As many companies shifted the cost base, they also ended up shifting the ability to control risks.

The pandemic has reinforced the essence of one famous saying – “Keep your friends close, and your enemies closer (so the enemies can be observed)”. There is no bigger enemy than an unknown and uncontrollable risk.

As global enterprises now look towards a more well-informed supply strategy, the focus has shifted from pure cost or pure risk diversification to a flexible approach that requires more frequent assessment of the traditional risk factors, and a constant vigil on the new emerging ones, vis-à-vis cost. There is also a realization that whether a reactive strategy is more applicable, or a proactive one, a robust data and analytics framework is the first and the most important step in containing supply chain risks.


Data-driven supply chain risk management (SCRM) strategies entail using a combination of Big Data Analytics and AI across the 3-stage risk management process:

  1. Sensing Risk – There are multiple frameworks that can be applied to assess risks. PESTLE – political, economic, social, technological, legal, and environmental – is one such comprehensive framework that offers multi-dimensional analysis that can be applied to segment the risk factors and apply suitable data collection and assessment to each segment.

With AI, machine learning techniques can be applied to build complex decision models. Further, with advancements in deep learning, the framework for prioritizing the risk markers can itself evolve in real-time, based on constantly changing scenarios.

Several other AI techniques, such as Petri Nets can be used to model concurrent distribution flow, while Multi Agent Systems are used to further assess whether the nodes in a chain are coordinated or conflicting, revealing bottlenecks.

Further, with advancements in deep neural networks, it will become possible to predict the severity, likelihood, and detectability of a risk to understand the “unknown unknowns” and plan the supply chain response much in advance.  


  1. Quantifying Risk – A critical aspect of risk assessment is the impact quantification, which then determines the scale and speed of response, entailing certain costs. More accurate assessments mean optimized cost of risk management, which in case of global supply chains can run into millions of USD.

Cost of SCRM can be a tricky affair – CFOs favor a lean and mean machine, however, as they say even in medical practice that some fat can be lifesaving. Same applies to supply chains, as some global JIT supply chains (Toyota, for instance) learnt during the Fukushima nuclear disaster and ensuing tsunami. Back then, Toyota had instituted an intricate system based on analytics and AI to assess its vast supplier network and the thousands of parts it sourced to come up with a dynamic suppliers-to-parts map that could help respond with flexibility in case of any such future crisis. Result – Toyota emerged quickly out of the semiconductor supply crisis due to the pandemic in 2021, as it quickly shored up inventory (essential fat). Further, Toyota used this AI-generated prescription to initiate conversations between auto component makers, chip makers, and other parts makers that relied on the chips.

Big data analytics and AI can be significantly helpful in reducing the time lost, and cost impact of a delayed risk response. Big data analytics has evolved significantly to deliver highly accurate pattern matching capabilities when fed with good quality historical data. AI and machine learning have further offered the ability to combine big data with fast data dynamic fast data from social media streams that provides valuable information about ground reality of locations, assets, business sentiments, and weather that can significantly add to the accuracy of pattern-matching and build prescriptive decision-guiding models.


  1. Responding to the Risk – The two above-mentioned steps, when managed by feeding sound AI algorithms with good quality data, can maximize the ability to prevent or circumvent a risk. In the event of black swans, such as this pandemic, nimble and real-time mitigation response is key. In such as scenario, risk response time has been demonstrably lower with the efficient use of big data analytics and AI.


While the steps to arrest supply chain sourcing risks are no different with this unknown crisis, the critical elements of context, time, and cost that amplify the dynamic complexity of any supply chain risk, can be better managed with a well-thought enterprise-wide and supply chain wide Big Data and AI strategy. McKinsey estimated in a 2018 study that AI could generate as much as $1.3 trillion in value for global supply chains by 2040, the second highest after marketing and sales.