The convergence of increased data availability and advancements in computing power make it the right time for supply chains to embrace machine learning and AI.
In our series of articles on machine learning we have gone on a journey reviewing key machine learning concepts in supply chain (Top 5 Machine Learning Use Cases in Supply Chain) and explored practical examples and use cases that organizations can deploy today (Three practical examples of machine learning transformations in supply chain). Now, let’s explore where AI/ML is headed to power the supply chain of tomorrow.
Gartner predicts that through 2024, 50% of supply chain organizations will invest in applications that support artificial intelligence and advanced analytics capabilities1. In study after study, there is evidence of the growing use of ML and AI across all industries and business functions including in supply chain. These new digital technologies help supply chains become more integrated, transparent and automated than traditional supply chains from the past and fuel the digital transformation of many organizations.
Download the white paper, Five Tips for Forecasting Uncertain Demand
Effective demand forecasting is a demanding activity based on a branch of mathematics - statistics - that’s nonintuitive to most laymen. Gartner’s supply chain benchmarking research shows that best-in-class companies can achieve forecast accuracy of eighty percent or higher. Given that no forecast made without a crystal ball is ever perfect, eighty percent sounds very good...
Digital transformations are also accelerated by the changing talent landscape. Millennials will comprise the majority of the workforce by 2025 followed closely by Generation Z (those born after 1997). These “digital natives” have always grown up connected with access to anything and everything at a moment’s notice. This rising generation of talent has new expectations of how they will rely on technology to augment and automate their supply chain planning decisions.
Digital is a strategic imperative for most supply chain organizations fueled by rapid technology advances, shifts in customer demands, business operations, ecosystems, and workforce. As organizations advance in their supply chain digital transformation, they will need to examine the opportunities of where the new machine learning-enabled supply chain of tomorrow can take them.
In the paper “Evaluate Four Key Areas for Machine Learning in Supply Chain Planning”, Gartner identified key areas where machine learning is changing the future of supply chains.
1. Better Predict an Event Happening:
Companies need a more accurate view to make their planning decisions. Machine learning algorithms focus on how to better predict an event will happen. Today, the most common use case is demand sensing.
2. Improve the Quality of the Supply Chain Model:
Many organizations spend too much time correcting and/or adjusting their plans because the data is inaccurate or static. Machine learning quickly improves the quality and accuracy of the supply chain model. Usage includes lead-time prediction, cleaning sales history to remove the impact of stock outs, cleansing master data, and more.
3. Improve and Advise the best “Course of Action” to Take:
Supply chain planning is about making decisions – some of these decisions can be very simple (for example, how to size an item’s safety stock level) and some are very complex (for example, how weather influences demand level at the SKU location level). When it comes to making the planning decision, there can be several approaches depending on circumstances – machine learning can help “advise” which is the best course of action to take given the situation. For example, determining product demand patterns and the degree of aggregation to find the optimal level at which to run the forecasting algorithms.
4. Compare the Decision Taken with the Eventual Outcome:
The primary goal of machine learning in supply chain planning is about making higher-quality decisions. This means comparing the decision taken with the eventual outcome and determining “how good was the decision versus actual performance” and “could we have made and taken a better decision?”. What-if scenarios have become popular to help planners identify the best outcome when comparing different constraints and trade-offs. Over time, machine learning algorithms will get better at recognizing which factors are most relevant and the next time a situation arises the system can utilize that ‘knowledge’ and help facilitate a better plan.
As advancements in machine learning continue to accelerate, supply chain leaders should evaluate the opportunities to apply machine learning today, and apply those learnings to a broader plan for the supply chain of the future.
Sources:
1: Andrew Stevens, et al., Predicts 2021: Supply Chain Technology, Gartner, December 2020
