Artificial Intelligence in Supply Chain Management
A February 2019 U.S. executive order outlines U.S. priorities in Artificial Intelligence (see DisCo’s post on the American AI Initiative here). In May 2019, the Organisation for Economic Cooperation and Development (OECD) approved key principles to guide a healthy adoption of AI (see DisCo’s coverage here). Amidst this focus on regulating AI, private businesses are eager to incorporate these technologies to improve performance. This post examines an area where AI application has been pervasive and robust across industries: supply chain management (SCM).
What We Mean by AI
AI refers to a category comprised of a whole host of technologies that mimic cognitive functions, traditionally ascribed to the human mind, including neutral networks, natural language processing, robotics, expert systems, and intelligent systems. In the commercial context today, the term is most often used to refer to speech and vision recognition systems, machine learning, and deep learning. Machine learning is a branch of AI where systems can “learn” from data, identify patterns, and make decisions with minimal human assistance. As Adeel Najmi, Senior Vice President, Products at One Network Enterprises puts it, “learning occurs when a machine takes the output, observes the accuracy of the output, and updates its own model so that better outputs will occur.” Deep learning, a specialized form of machine learning, uses many layers of neural networks to classify images without extracting features from images.
Consumers do not need to look far to see the applications of these technologies. Virtual assistants like Apple’s Siri and Amazon’s Alexa use machine learning and natural language processing to parse and classify voice commands. Interactive “chat bots” on many retail or banking websites use natural language generation to respond to customer’s questions. From autonomous vehicles, to facial recognition devices and personalized medicine, the use cases of AI extend across industries.
Getting Goods to the End Consumer
Behind the scenes, companies have been applying AI in supply chain management in ways less obvious to the end consumer. From the automotive industry, to the pharmaceutical and consumer electronics industries, businesses are dealing with increasingly complex and globalized supply chains. In modern multi-tier supply chains, hundreds or thousands of suppliers may contribute to a single product. The process of procuring raw materials, managing trading partners, and sequentially planning and executing tasks with huge volumes of data has become a much larger task, requiring heavy-duty data analytics. Unnecessary delays, caused by mismanagement, could result in supply-demand mismatch, shortages, overstocking, and poor customer experience.
Demand Forecasting. Machine learning has been used to forecast demand using historical shipping data since the early 2000s. Procter & Gamble Company (P&G), for instance, has used sophisticated modeling to reconcile demand signals from point-of-sale data, retailer warehouse and outlet inventory, and retailer forecasts for over a decade. In 2018, P&G announced that it will globally adopt the demand planning tool by E2Open, an AI software provider for supply chains. P&G, along with Amazon, UPS, Walgreens Boots Alliance, and other Fortune 500 companies, are using advanced machine learning algorithms to optimize demand plans for product launches, adjust stocking strategies, and/or find optimal delivery routes. Some retailers are now incorporating competitor pricing data, store traffic, and weather data to improve demand forecasts.
Process Engineering. Supply chain managers are also using robotic process automation (RPA)—software bots that mimic human actions—to automate repetitive, rule-based operational tasks. While RPA has traditionally involved less AI, this is changing. Businesses are coupling automation with predictive analytics to better manage operational processes. Some specialized supply chain-focused AI software providers, like AspenTech, help with plant production schedules and manage assets at chemical companies, wastewater treatment facilities, and metals and mining companies in particular, where shutting down machinery could have environmental, health, and safety implications.
Risk Management. The AI-enabled rapid decision-making has also helped mitigate unexpected disruptions to a supply chain—whether it be a cyber attack, bankruptcy or failure by a supplier to meet regulatory standards. KPMG, for example, was commissioned to build a tool to review thousands of documents to facilitate an acquisition deal. The client had sold billions of dollars worth of assets to the acquiring firm and needed to send to the acquirer the documents relevant to the sold assets, while withholding sensitive internal documents. KPMG used machine learning and cognitive automation technology to analyze 2 million documents and classify those that would go to the seller versus the acquirer. As this case exemplifies, machine-learning software can significantly reduce transaction costs for the seller and acquirer and lessen the impact of the merger or acquisition to supply chains.
In sum, businesses are applying AI to optimize supply chain management across industries. In addition to having more personalized products, the benefit to consumers also includes the more cost-effective means in which businesses are delivering those products.
