COVID-19 has exposed the inherent vulnerabilities in supply chains. With COVID-19, major parts of the world continue to impose social distancing and unpredictable lockdowns, leaving many organizations unable to decide on a reasonable course of action. Reason: todays supply chains are extremely complex. Over the last few years, supply chains have been tailored to meet faster lead times, changing delivery expectations, a broader mix of products and services, new markets/locations, and a combination of suppliers with differing relationships. In the past, supply chains have managed to respond to all types of extreme events like volcano eruptions and tsunamis. But the COVID-19 pandemic has caught us most unprepared because the impact has been global, simultaneous, rapid, and unrelenting. But inaction can be expensive. It takes a toll on bottom lines, customer confidence, and organizational reputation.
Organizations are trying to respond to the challenge placed before them by COVID-19. They now want to map supply networks at a more granular level, drive visibility and acquire real-time information across suppliers, sites, parts, inventory, and products that can be at risk. But the leaders in pharma, manufacturing, retail, consumer packaged goods, etc., are showing the way by taking the long-term view. They are deploying Artificial Intelligence (AI) and Machine Learning (ML) to manage even the most extreme exceptions and build resilience into their supply chains.
One reliable analyst forecast places the use of AI in the form of digital twinsor the digital representation of physical supply chains along with simulation modelsas among the top 8 supply chain trends for 2020.[i] A recent market study released in June places the growth of AI in the supply chain market at a CAGR of 45.3% from 2019 to reach $21.8 billion by 2027.[ii]
The gap in AI-powered supply chain management
At the moment, the upside for organizations using AI and ML to strengthen their supply chains is considerable. Barely 25% of the manufacturing industry has adopted AI.[iii] Across industries, adoption levels could be as low as 12%.[iv] This means quick adoption of AI and ML has the potential to give organizations a significant competitive advantage.
The path to accessing the advantage is to create digital twins of their supply chains. A digital twin uses a data-driven, AI-based ML approach to create effective virtual operational models of supply chains. The digital twin uses historical supply chain data combined with industry-specific event data (provided by a new breed of service providers[v]). The result is a system that can be used to simulate a variety of conditions, events, and variances. At the heart of the system is a real-time data repository, a library of algorithms, AI-based decision models, and ML engines that improve the system through continuous learning using external and internal sources of information. The digital twin can then be used to simulate events and variance in environmental, geopolitical, economic, and technological dependencies to predict the impact of demand, inventory, and production. The simulation model is a practical necessity for supply chain managers as it allows them to take targeted action and initiate corrective measures.
Fast and reliable path to AI-based systems
The fastestand most reliableway to build digital twins of supply chains is to opt for a technology provider like ITC Infotech. ITC Infotech has the domain and technological expertise required for such implementations because of the experience gained from implementing and running complex supply chain applications for its parent conglomerate, ITC Limited, which has diversified investments in Manufacturing, CPG, Hospitality, Retailing, Education, and Agriculture. ITC Infotech combines its domain experience with commercially available tools and platforms and internally developed IP to create supply chain models. The outcome of these models is a quantifiable assessment of risk and its impact on the supply chain.
As an example, ITC Infotech created a digital twin for a pharma customer to address the problem of a product it manufactures that has a short shelf life and a highly variable demand. The goal was to ensure that customer service was maintained at a given threshold while wastage was minimized. With complete unpredictability in demand, only a simulation model could help the pharma customer arrive at reasonable decisions.
In another case, ITC Infotech worked with a bullion metals organization to predict production and meet delivery commitments. Bullion is unpredictable with uncertainties over production time and volume of output. In such instances which are not deterministic, only simulation can provide support in decision-making. Put another way, wherever there is uncertainty, AI and ML can be deployed effectively.
Responding to COVID-19 type events
How does this relate to the specificand extremely unusualchallenges are thrown up by COVID-19? To begin with, ITC Infotech has the ability, to not only use data to simulate supply chains, but also the analytical models required to consider social distancing as a factor that impacts supply chains. Through the work done for airports and railway stations to predict crowd patterns, ITC Infotech can fast forward the ability to predict the space and time required with social distancing as a parameter and how it will impact queuing and wait times.
The key to building digital twins/simulation models is to identify the top pain points the simulation needs to address and ask the question, Can simulation solve this problem? The effectiveness of the model, in turn, depends on the richness and volume of historical data available with an organization. Now combine this with real-time external industry-specific data to improve the fidelity of the insights and outcomes. Organizations can improve this even further by leveraging technology partners like ITC Infotech for best practices that are cross-pollinated from a variety of industries.
Organizations that want to have the level of preparedness required for an event like COVID-19 will use AI and ML to proactively manage risk. Those that dont will continue to find that disruptions can be very expensive. In 2019, the impact of the inability to forecast disruptions and take action was a loss of ¬ 100 M for 1 in 20 organizations, with the average annual cost of disruption being ¬ 10.5 M.[vi] AI and ML can insure organizations against this loss.
President- Manufacturing & CPG
Founder and Lead Analyst, EIIR Trend and Pareekh Consulting