Large retail and distribution companies have long had specialist systems to run supply chains, manage warehouses, and take some of the effort out of planning and forecasting.
But often these systems do not reach to get a real-time view of inventory levels in stores, for example, or harmonize demand flows through the supply chain.
Another problem that was painfully obvious during the pandemic is that retail supply chains were designed for a simpler age, when demand was fairly predictable and stores did not have to suddenly close for extended periods.
It’s not just COVID-19 that is giving supply chain professionals sleepless nights. Consumers’ desires for near-instant gratification, greater personalization and more last-mile fulfillment options are having a profound impact on supply chains, adding more and more complexity.
These changes are obliging retail and distribution companies to adopt more sophisticated techniques that increasingly use AI to improve decision making in regards to the supply chain and other back-office functions.
For example, store network planning has become more difficult because of the constantly changing interplay of offline and online channels. Demand forecasting is also much more complex than it used to be. According to a report by analyst firm RSR Research, the unpredictability of demand is the main reason cited by retailers for not being able to do omnichannel fulfillment well.
Demand forecasting is the beginning point for a lot of back-office processes that make retailers function. So, if they can forecast likely demand with greater accuracy, it will be easier to forecast sales and costs. Accurate sales forecasts can be enhanced by the use of machine learning on downstream data, such as the records generated by point of sale terminals.
By adopting AI and ML-driven strategies for demand and supply management, retailers can get a more complete picture of what is driving demand and, equally importantly, their ability to fulfill that demand, and so make better decisions.
AI can take the hard work out of analyzing the myriad datasets that impact demand patterns, enabling retailers to better understand what consumers want to buy now, and so prioritize boosting stock levels for product lines experience a sudden uptick in demand – perhaps because a social media influencer has featured the product.
The other side of the coin is when retailers have overestimated demand and are forced to discount stock to clear inventories, with the consequent negative impact on the bottom line.
This was a problem at the height of the COVID-19 pandemic for fashion retailers whose stores were forced to close. When the stores reopened, unsold spring/summer inventory had to be cleared out to make room for the next season, and the stranded stock had to heavily discounted or written off.
AI is also being used to improve the operational aspects of retail supply chains, streamlining warehouse operations and optimizing routes for delivery drivers to ensure that deliveries arrive on time and with the lowest financial and environmental cost. The e-commerce boom has led to a huge increase in “last mile” deliveries and AI can mitigate the negative externalities.
One of the most interesting areas where retailers could benefit from using AI is in optimizing processes. Process mining is a promising application of AI that enables in-depth understanding of business processes. It gives a clear view of which events and actions - in areas related to customer service for example - positively and negatively influencing business outcomes, thus making it easier to adopt corrective measures for better performance and results.
Reducing operational risk is likely to be a key concern for retailers in the future, and while AI cannot easily predict the likely consequences of a “black swan” events like a viral pandemic, it can help retailers better manage their risk and take corrective action during disruptive events.
Even though the retail industry is passing through a particularly tumultuous period, it is clear that the winners will be those that leverage new technologies such as AI and machine learning to adapt most successfully to the short-term challenges imposed by pandemic and long-term trends like the inexorable growth of e-commerce and evolving customer expectations.