Image of supply chain logistics staff member in warehouse with AI overlayed

Using Artificial Intelligence to Mitigate Supply Chain Risks

3 July 2023
From our ‘Thinking outside the box’ series
Professor Ben Fahimnia looks at how AI-driven analytics can help our essential supply chains to build resilience capabilities through systematic detection of mitigation strategies.

From COVID-19 to the war in Ukraine, the world has scarcely felt more chaotic – and the disruption to our fragile supply chains has been universal. But what if there was a way to predict and prepare for such random disruptive events? This science fiction could become science fact through harnessing the power of artificial intelligence (AI).

There is no real end in sight to the disrupted global supply chains. Continuing geopolitical instability, labour shortages, severe weather, and lingering inflation threaten to keep supply chains unsettled for the foreseeable future. Our essential industries must find effective ways to live with disruptions by systematic detection, evaluation, and implementation of mitigation strategies.

Australia is particularly vulnerable to trade disruptions in the global marketplace because most of our supply chains heavily rely on overseas supply with limited tolerance for disruption. Risk mitigation is even more crucial for the supply of essential goods and services (i.e., food, water, health, and shelter) as their continued supply during disruptions plays a critical role in Australia’s economic functioning, the wellbeing of its people, and our national security.

“Predictive analytics” use probability theories to determine what is likely to happen based on patterns and trends revealed from analysing large historical and real-time data. Such tools are able to predict the impact of future disruptions on our essential supply chains.

Predictive analytics have been around for decades, but only recently they started to become mainstream thanks to AI methods capable of analysing large amount of unstructured data. For example, machine learning methods can now use historical disruption data as well as real-time operational data to provide an early warning of future supply chain failures. Many companies in the semiconductor industry have learned since Covid-19 how to utilise such AI tools to predict potential failures across the supply chains.

“Prescriptive analytics” use the results provided by predictive analytics to take a step further and determine the best action plans to reach a desired outcome. Such tools use advanced optimisation models and decision logic rules to find out the best mitigation strategies for the essential industries to capitalise on.

 Predictive and prescriptive analytics can also assist with federal and state policy decisions on supporting risk management initiatives. For example, the use of predictive analytics will reveal whether resilience-building initiatives by certain industries are hampered by regulations. Prescriptive analytics can help policymakers take informed actions to provide special services to companies involved in the supply of critical products, or to take direct ownership of the risk management of certain products.

As impressive as AI tools are, their implementation is easier said than done. The most important challenge in Australia is data restrictions. AI tools require large amounts of precise digital data in order to train algorithms and produce reliable results (ChatGPT was trained on a corpus of over 570 GB of text data). In the past few years, most organisations have generated more data than ever before. However, effective data management systems need to be established by these organisations to deal with data clustering, availability, and security constraints. The second challenge is the initial capital investment for design and deployment of such AI models and acquisition of the AI-specific hardware that the models need to work with cloud-based systems.

We know that the frequency and magnitude of disruptive events will continue to rise, so will their significant impact on our supply chains. AI innovations can make such disruptions a thing of the past. AI-driven analytics can help our essential supply chains to build resilience capabilities through systematic detection of mitigation strategies to capitalise on. Australian industries and research organisations must urgently increase research and development in AI-driven analytics to empower our essential industries to build future-ready supply chains.

Related articles