In the age of e-commerce, consumers are demanding shorter delivery times, multiple buying channels, and a seamless shopping experience. Meeting these expectations requires a variety of distribution channels and store locations closer to the end consumer. However, such strategies often lead to additional challenges, such as fragmentation of supply and increased cost pressures.
AI can be defined as the ability of a machine to perform cognitive functions associated with the human brain. Prediction, the ability to translate existing information into new information, is at the heart of AI. Algorithmic advances, data proliferation, and huge increases in computer power and storage have helped to make AI-based prediction algorithms both abundant and inexpensive. As data availability expands, the ability to predict outcomes becomes more valuable and prevalent in a wider variety of supply-chain applications.
As predictive power continues to increase, the era of forecasting the behavior of mass populations based on aggregated historical information is coming to an end. A new era, relying on vast volumes of data derived from online commerce, has begun. The availability of this data enables companies to understand consumer purchasing habits and preferences with much greater precision. E-commerce companies are already capable of predicting what a particular consumer will buy in the next few weeks.
Another capability that AI is delivering is enhanced supply-chain visibility. AI, and particularly machine learning, provide the methods and algorithms required to digest huge data sets, while presenting only the information that’s relevant to the decision maker.
In order to achieve enhanced predictability and visibility by embracing AI, companies should focus on the following challenges.
Top-down mandate. It’s important to have a clear strategy around AI. Artificial intelligence, like any other technology, is a means to achieving a goal, not the end goal itself. Top management should identify business problems that can be addressed through enhanced prediction and visibility, and prioritize these problems based on the company’s needs. Senior management should also conceive new business models that draw on the greater predictive power and visibility made possible through the application of AI.
Data as an asset. AI requires lots of data. Machine learning algorithms aren’t magical tools; data is a critical component in the application of AI for gaining enhanced predictive capability and visibility. Companies should start thinking about data as a valuable asset. They need to devote resources to harmonizing, storing, and analyzing data, as well as to developing effective data governance and stewardship policies, in order to build a single source of truth.
Talent. Talent is critical to putting AI to work in supply chains. Data engineers and data scientists are among the technically skilled individuals whom companies are currently recruiting. Companies will also require translators who can serve as a bridge between quants and business associates.
Change management. AI requires companies to adopt new ways of doing things. Consequently, many processes and procedures will change, including employees’ roles and tasks. Enterprises should be aware of this challenge, and define strategies to make the transition as smooth as possible.
Demand planning will continue to drive most successful applications of AI in supply chains over the next few years. Similarly, AI applications in warehousing and transportation, where information such as product identification and the truck’s estimated arrival time is generated in real time, will gain more attraction. Finally, AI applications where humans and machines interact will become more common in supply-chain facilities.
Sergio Caballero is a research scientist at the MIT Center for Transportation and Logistics.
Timely, incisive articles delivered directly to your inbox.