It’s Time to Start Using AI for Supply Chain Risk Management
Microsoft and the US Department of Energy grabbed headlines by asserting a partnership to develop AI-powered functions to assist first responders reacting to pure disasters. One of the primary AI prototypes in growth will make use of pc imaginative and prescient to detect and predict the frontiers of energetic wildfires and floods. The second software makes use of an AI software that must be in each danger supervisor’s toolbox: simulation. This simulator will assist groups in working mock eventualities to higher plan and put together for the following massive pure catastrophe.
Many corporations had been caught flat-footed by COVID-19, maybe the biggest disruption to international commerce in 100 years. While they scramble to realign their provide chains to meet the fact of 2020 and past, now could be exactly the time for these organizations to think about AI-based danger administration instruments, from cutting-edge predictive analytics methods to the tried-and-true strategies of simulation and optimization.
Mathematical simulation and optimization kind a robust combo that helps create lean, cost-efficient provide chains which can be additionally resilient to disruption. The first step in integrating these applied sciences into your group’s operations includes digitizing the availability chain, typically referred to as a digital twin. This digital twin of your bodily provide chain ought to element contingency choices of all types, resembling alternate suppliers, lead occasions, stock ranges, site-specific redundancies, payments of supplies, and even enterprise continuity plan paperwork (BCPs). This step might be taken for granted and must be collected earlier than, not throughout a disaster.
Next, simulate full or partial outages: buyer areas being flooded, fires disabling manufacturing strains, and many others. You can ascribe chances to these eventualities by acquiring related exterior knowledge resembling hundred-year flood maps and county-level pure catastrophe data.
After utilizing these eventualities to illuminate any vulnerabilities within the provide chain, the following step is to guarantee cheap protections in opposition to the worst weaknesses. In one examine, Ford discovered that it was not essentially essentially the most expensive automobile elements that introduced the best danger, however slightly a small, missed group of crucial elements like o-rings and valves. Strengthening these areas of weak spot may contain onboarding a brand new provider or carrying extra security inventory stock.
As COVID-19 demonstrates, we can not assume that disruptions are short-lived or that any ensuing adjustments will probably be short-term, even for provide chains which were rigorously designed and optimized to stability price and danger. When one thing inevitably goes awry, having a ready-made contingency playbook generally is a game-changer. And if the disruption takes a extra everlasting kind, it’s a good suggestion to re-optimize the availability chain to be sure that your operations are cost-effective, whether or not that disruption has resulted in undersupply or oversupply.
There are many dangers on this world which can be really unpredictable, particularly in advanced international provide chains. Nonetheless, danger and provide chain managers would do effectively to intelligently incorporate methods from the ever-advancing area of predictive analytics into their toolboxes. For instance, half failure might be predicted on a producing line, in a truck engine, or in tools at a buyer website. Ryder Logistics and different fleet operators have deployed predictive upkeep software program to scale back asset downtime. What different dangers might be predicted? The unfold of pure disasters resembling hurricanes and flooding, poor crop yields, late supply, capability utilization, and even monetary efficiency of a provider or buyer. Furthermore, higher predictive fashions make simulation and optimization fashions much more helpful. If a fleet operator improves a probabilistic predictive elements failure mannequin, she or he can optimize operations with extra certainty or run simulations with extra reasonable eventualities.
As we modify to the brand new regular introduced on by COVID-19, it’s price taking a contemporary have a look at how AI can enhance provide chain resiliency, whether or not that’s making a digital twin, constructing predictive fashions, or including simulation instruments to put together for no matter could come your means.
About the Authors
Nate DeJong is a Senior Consultant with LLamasoft. He has spent seven years making use of algorithms to clear up a wide range of provide chain issues from forecasting to provider danger. He spends important time throughout enterprise management, data science and engineering groups with the objective of delivering analytically sound options throughout organizations.
Colleen Eland is an Engagement Manager with LLamasoft. She has 18 years of numerous international expertise in fast-paced shopper items corporations. Her roles have concerned serving to to enhance scale in provide chain, procurement, buyer advertising and sourcing methods for many fortune 500 corporations.
Sign up for the free insideBIGDATA e-newsletter.